r =*■ Qfatnell IniuerBitg Slihrara BOUG-HT WITH THE INCOME OF THE SAGE ENDOWMENT FUND THE GIFT OF HENRY W. SAGE 1891 Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924012487157 THE MECHANISM AND GRAPHIC REGISTRATION OF THE HEART BEAT. THE MECHANISM AND GRAPHIC REGISTRATION OF THE HEART BEAT BY THOMAS LEWIS, M.D., F.R.S., F.R.C.R, D.Sc. Honorary~Consulting Physician, Ministry of Pensions; Late Consulting Physician in Diseases of the Heart [Eastern Command) ; Physician of the Staff of the Royal Medical Research Committee; Physician and Lecturer in Cardiac Pathology, University College Hospital, London; Fellow of University College, London 1 -sVi ft 1 sis* I FAUL D-HOEDB.B^ NEW YORK PAUL B. HOEBER 1921 By the same A^uthor: Clinical Disorders of the Heart Beat Fifth Edition, $3.00 net Clinical Electrocardiography Second Edition, $j.oo net The Soldier's Heart and the Effort Syndrome $2.50 net Lectures on the Heart $2.50 net .■;' , The Mechanism and Graphic Registration of the Heart Beat With Especial Reference to its Clinical Pathology $16.00 net Heart A Journal for the Study of the Circulation Edited by Thomas Lewis. Per volume, $7.50 net PAUL B. HOEBER, Publisher 67'69 East 59TH Street, New York PREFACE. IT is now some eight years since I published, under the title of " The Mechanism of the Heart Beat," a book in which I attempted to review a collection of numerous, and at that time recent, researches upon normal and disordered action of the heart beat. The subject-matter at my disposal had been gathered by precise methods of research ; it illustrated the apphcation of such methods to clinical studies ; it emphasised the value of reproducing in animal experiment changes witnessed during the progress of human maladies. For the last-named reason a number of clinical and experimental observations were placed side by side in the book so that they might be compared closely with each other. Mj^ general purpose was to show the solid progress which may be achieved by exacting methods and to preach an assiduous gathering together of all evidence in support of any given proposition. For it was in my mind, and still is, that clinical pathology, by which I mean the study of disease in the living man, suffers both from inexact observation and from hasty conclusion. Many methods of clinical observation, being subjective, lack accuracy. Some subjective methods are commonly believqd to be effective only in the hands of a chosen few ; such methods, while always open to scepticism, never can be productive. In no science, in the true sense of the word, is there the almost open vaunting of and reliance upon personal skill, in no science is there the veneration of supposed manipulative dexterity which we discover in our realm of medicine. Creditable methods of research are those which are capable of such vi Preface. description or demonstration that the intelligent and diligent may employ them to the end results which their originators claim for them. It is by this test primarily that methods are weighed in science ; it is by this test also that clinical methods will be judged eventually. A method which depends for its success upon a supposedly unusual sensitiveness of touch or of hearing, is a method whose clinical scope is admittedly limited, whose scope in science is almost negligible. Observations undertaken by any method which is not generally approved and which has not passed a rigorous censorship, are observations which add little to the sum of general knowledge. When we ponder upon the bedside tests, general and special, now in vogue, and ask ourselves how many of these have filtered through the sieve of rigidly controlled experiment, when we ask how many when filtering comes to them as come it must, will show a casting in so fine a mould that nets, though strictly meshed, will not entangle them ; then do we probe the foundation upon which an art, aspiring to the place of science, rests. It is urgency^ — the immediate and daily promptings of the sick — ^which bolsters and preserves many of the- expedients and opportunisms of the bedside worker. But these same expedients and opportunisms litter and obstruct the path of progressive knowledge. A physical sign or method, for lack of better, may serve the purpose of the moment, it serves no purpose in the pursuit of science till the measure of its fallibility is taken. Clinical observers, for the most part, do not yet recognise that many methods which they regularly employ, useful as they may be in the practice of an art, are inadmissible to the science. To advocate the general use of laboratory methods involving costly devices and time-robbing technique is not my desire ; it is to emphasise the vital . importance of methods of precision Preface. vii in progressive studies of disease. While it is to be freely acknowledged that simpler methods are essential to the practice of medicine, it is clearly right to insist that in compiling reports contributory to scientific medicine precise methods are desirable. It is from this standpoint that the example of electrocardiography is to be stressed. Inexact method of observation, as T believe, is one flaw in clinical pathology to-day. Prematurity of .conclusion is another, and in part follows from the first ; but in chief part an unusual craving and veneration for hypothesis, which besets the minds of most medical men, is responsible. Except in those sciences which deal with the intangible or with events of long past ages, no treatises are to be found in which hypothesis figures as it does in medical writings. The purity of a science is to be judged by the paucity of its recorded hypothesis. Hypothesis has its right place, it forms a working basis ; but it is an acknowledged makeshift, and, as a final expression of opinion, an open confession of failure, or, at the best, of purpose unaccomplished. Hypothesis is the heart which no man with right purpose wears willingly upon his sleeve. He who vaunts his lady love, ere yet she is won, is apt to display himself as frivolous or his lady a wanton. For this reason oftentimes I was particular to set forth evidence upon evidence for a given conclusion in the book, feeling that no doctrine is sufficiently supported if yet another serious argument may be urged in its favour. For the same reason the leading theses chosen were those founded upon abundant evidence. My readers will find no reference to the still controversial question of the " neurogenic " or " myogenic " origin and transmission of the heart beat. The walls of the heart are composed of a syncytium of muscle fibres, closely interwoven with nerve fibrils and ganglia ; in speaking of viii Preface. the musculature of the heart, I have done so upon the distinct understanding that this term refers to muscle in full functional connection with the nerve elements surrounding it. For, dealt with in this manner, it is immaterial to the subjects considered and to the conclusions reached, whether one or other view is held. It may not be inappropriate to refer here to a general terminology employed by writers upon disorders of the heart and especially to certain terms which are used to denote physiological properties of cardiac muscle. The strict separation of five cardiac functions, rhythmicity, excitability, contractility, conductivity, and tonicity, is one which, as I am well aware, is jealously guarded by some writers, more especially by the disciples of Engelmann. This doctrine may or may not be justified. Be that as it may, the classification of cardiac disorders upon this basis is neither exact nor serviceable. Therefore, in ijsing the convenient terms rhythmicity and conductivity, and similar derivatives, they are employed only in so far as they indicate the observed facts, namely, the origin of heart beats in a limited area and in rhythmic fashion, the transmission of waves of contraction from one chamber to another, or from one portion of the musculature to another. Physiological hypothesis does not concern pathologists ; physiological observations do ; pathological observations are still more relevant, and from these we may form an idea to guide us in our work. If the basis of our working hypothesis is still speculative, if physiological imagination governs it, as pathologists we work upon a foundation guaranteeing no reasonable security. The nature of my subject necessitates abundant reference to graphic records, of which I have endeavoured to give examples in which the analysis will be unquestioned by those familiar with them. Preface. ix Of the immediate value of graphic methods to practical medicine, it is my desire to speak but briefly. These records have placed the entire question of irregular or disordered mechanism of the human heart upon a rational basis, so giving to the worker the confidence of knowledge ; they have influenced prognosis and have rendered it more exact ; they have potentially abolished the promiscuous administration of cardiac poisons, and have clearly shown the hhes which therapy should follow. The new clinical observations have stimulated and directed a host of laboratory researches, anatomical, physiological, pathological and pharma- cological of a valuable nature. The records in themselves constitute the most exact signs of cardiac affections which we possess. Imprinted by the disease itself, thej^ form permanent and unquestionable testimony of events which have occurred, and may be placed in the balance, without disquietude, while experiences of a more subjective kind fill the opposing scale. Judged from this standpoint they possess also a great didactic value ; they demand and impress accuracy of observation and, while nicely delimiting facts and fancies in their own sphere, sharpen the perceptions of similar boundaries in other spheres. The book was written originally in the hope that it might stimulate the study of heart affections by precise methods. Its subject awakened a more widespread interest than I had anticipated. The exhaustion of the printed copies of what was intended as a monograph, four years ago, has forced me to consider the publication of a second edition. My objects remain unchanged, but eight years have brought many additions to our knowledge, it has hardened former conclusions, it has filled in many gaps. It has become possible and desirable to expand the scope of the book ; but the X Preface. expansion, and some dissatisfaction with the original title, have persuaded me to change its name ; while I have maintained the general plan of the " Mechanism o£ the Heart Beat " the text has been rewritten almost wholly and a new series of figures is employed, which I trust will be found clearer to decipher. A large proportion of the text is written as an abstract of my own observations ; for this emphasis of my own work, I offer no apology, in that it seems to me desirable that an author should write mainly of that which has passed within the range of his own experience. Where I discuss the observations of others I have done so after fully studying their writings ; when I have been in doubt as to the validity of experiments or the interpretation of them, I have, wherever it has been possible, repeated the experiments. Consequently, few phenomena are spoken of in this book with which T am not personally familiar. To my old friend, Dr. T. Wardrop Griffith, of Leeds, I express my deep indebtedness for his great care in reading the proofs, and for the many improvements which he has suggested and of which I have fully availed myself. THOMAS LEWIS, 8eptem,ber, 1919. CONTENTS. Preface I-X CHAPTER I.— SPECIAL ANATOMY The sino-auricular node The auriculo-ventricular connection Morphology of the special tissues ANAIOMICAL NOTES 1 1 2 12 14 II.— MECHANICAL RECORDS OF THE HEART BEAT, WITH ESPECIAL REFERENCE TO INTRA- AURICULAR PRESSURE CURVES AND VENOUS RECORDS Methods of investigation Curves of muscle shortening Curves of heart volume . . Sound records Curves of pressure Qualities of the recorder . . The venous pulse in animals and man. Oesophageal curves Cardiographic curves Carotid and radial curves The records and their meaning The intraventricular and aortic pressure The carotid pressure curve The radial pulse curve . . Intra-auricular pressure curve The venous curve 15 15 15 16 17 . 17 18 10 21 21 23 23 curves 25 25 26 , . 26 . 27 XII Contents. .— GALVANOMETRIC METHOD pAaE 30 The string galvanometer 30 Obtaining standardised electrocardiograms . Testing the properties of the string Suitable contact electrodes 33 34 37 Photographing Time-markers 39 40 Simultaneous records 41 Measurement 42 Registration of heart sounds 42 IV.— THE BROAD FEATURES AND TIME-RELATIONS OF THE NORMAL ELECTROCARDIOGRAM AND CERTAIN PRINCIPLES OF INTERPRE- TATION 44 The broad features of the normal electrocardiogram 44 The leads adopted . . . . . . . . 44 The physiological type of human electrocardiogram 44 Variations in the physical type in the several leads 48 The chief time-relations of the electrocardiogram, 49 Certain principles of interpretation . . . . . . 53 Leading directly from the muscle . . . . 53 Indirect leads . . . . . . . . . . 55 v.— THE NORMAL PACEMAKER OF THE MAMMALIAN HEART Forcing an excitation wave to follow the natural path . The point of primary negativity estimated by the direction of deflections Extrinsic and intrinsic deflections The point of primary negativity estimated hy timing the excitation icave . . Warming and cooling . Other observations Destruction of the sino-auricular node . OBSERVATIONS ON THK DYING HEART, 13TC 58 59 60 63 65 66 67 67 70 Contents. XI 11 CHAPTER VI.— FUNCTION OF THE A- V BUNDLE AND CURVES RESULTING FROM LESIONS OF ITS TWO DIVISIONS 71 The evidence that the auriculo-ventricular bundle transmits the impulses from auricle to ventricle in the mammalian heart . . . . . . . . 71 Curves resulting from lesions of the chief divisions of the bundle in the dog . . . . . . . . 75 VIL— SPREAD OF THE EXCITATION AURICLE AND VENTRICLE In the auricle . . In the ventricle . . The Purkinje pathway Conduction rates . . Further arguments NOTES WAVE IN 81 81 86 89 91 93 94 VIII.— THE MEANING OF CERTAIN VENTRICULAR DEFLECTIONS 95 Q, R, 8, THE INITIAL DEFLECTIONS . . . . . . 95 The duality of the normal electrocardiogram . . . . 95 Method of calculating the electrical axis . . . . 98 Trigonometrical method ( Einthoven's formulce) 99 Simple geometric examples . . . . . . 100 notation of the electrical axis. . . . . . . 102 The end-deflection " t " . . . . . . . . 108 IX.— ABERRANT CONTRACTIONS AND THE ELECTRO- CARDIOGRAMS OF HYPERTROPHY, ETC. .. Curves held to represent defects of the right division (the human levocardiogram) Curves held to represent defects of the left {the human dextro-cardiogram) Comparison with experimental curves Curves associated with preponderance of other ventricle . . Minor forms of aberration Displacement of the hearts axis division one or 117 118 119 120 122 125 127 XIV Contents. CHAPTER PAQE X.— THE ANALYSIS OF DISORDERED MECHANISM ARTEBIAL PULSE CURVES 131 Pulse intermissions . . .. .. .. •■ 132 Signs of a dominant rhythm . . . . . . ■ • 133 Signs of an undisturbed dominant rhythm . . . . 135 A regular sequence of beats originates from a single source . . . . . . . . . . • - • • 136 Phasic variation of the dominant rhythm . . . ■ 138 The rate of the dominant rhythm as an index of its source . . . . . . . . . . . ■ ■ ■ 138 The length of the cycle following a premature beat . . 138 Complete irregularity of the pulse . . . ■ . 139 XI.— THE ANALYSIS OF DISORDERED MECHANISM VENOUS CURVES 140 Identifying the " c " wave . . . . . . . 140 Identifying the " a " wave . . . . . . . . 141 The " As-Vs " interval . 142 The value of the summit " v " . . . . . . 142 The intekpeetation oe curves in which there are disturbances oe sequence . . . . . . 144 Prominent summits . . . . . . . . . 144 Supernumerary waves . . . . . . .. . 148 Records of premature beats . . . . . . . 149 Absence or apparent absence of " a " waves . . . . 149 XIL— THE ANALYSIS OF DISORDERED MECHANISM ELECTROCARDIOGRAPHIC CURVES .. 153 Auricular and ventricular contractions of physiological types . . . . . . . . . . . . . . 154 Ventricular contractions of supraventricular origin 1 54 Beats arising in the vicinity of the pacemaker . . .. 155 Anomalous beats . . . . . . . . . 155 Identifying different types of anomalous contractions 157 The algebraic summation of complexes . . , . 158 Contents. XV CHAPTEK XIII.- PAGE -EXPERIMENTAL HEART-BLOCK 162 Methods of producing heart-block in animals. . . . 163 1 . By direct interference with the conducting tracts, namely, the auriculo -ventricular node, the bundle or both divisions of the latter . . 163 2. By stimulation of the vagus . . . . . . 164 3. By introducing toxic bodies into the blood- stream . . . . . . . . . . . . 164 Records of experimental heart-block . . . . . . 166 The susceptible region . . . . . . . . . . 168 XIV.— CLINICAL HEART-BLOCK 171 The signs and grades ■ of heart-block in man . . 171 Prolongation of the '' As-Vs " interval .. .. 172 Dropped beats . . . . . . . . . . 173 Frequent failure to respond . . . . . . 174 Complete heart-block . . . . . . . . 174 The causes of clinical heart-block . . . . . . 178 1. From lesions of the conducting tract . . 180 2. The vagus and clinical heart-block . . . . 181 3. Heart-block as a result of poisoning . . 182 NOTE ON DISSOCIATION OF THE TWO AURIOLBS . . . . 183 NOTE ON DISSOCIATION OF Tilt! TWO VENTfilOLES (HEMI- SYSTOLEl . . . . . . . . . . . . 183 XV. -NEW RHYTHM CENTRES (ATRIO-VENTRICULAR RHYTHM) 184 Evidence that these new rhythms arise in the A- V node 188 Variation in "A-V " rhythm intervals . . . . 188 Influence of nerve stimulation upon "A-V " rhythm 191 Clinical examples' of A-V rhythm . . . . . . 192 XVI.— NEW RHYTHM CENTRES (IDIO-VENTRICULAR RHYTHM, ETC.) . . 196 Origin of the idio-ventricular rhythm . . . 197 Escape of neiv centres ; ectopic impulses . . . . 198 Centres of rhythm production . . . . . . . . 202 xvi Contents. CHAPTER PACK XVII.— VENTRICULAR EXTRASYSTOLES 205 Records of ventricular extrasystoles . . . . . ■ 209 Arterial records . . . . . ■ • • • • 209 Venous curve . . . ■ ■ ■ ■ ■ ■ ■ 210 Electrocardiograms.. . ■ ■■ ■■ 210 Ventricular curves of beats arising from the ventricle 211 Forced beats . . . . . . • • • • 211 Spontaneous beats . . . . ■ . ■ ■ ■ ■ 217 XVIII.— AURICULAR EXTRASYSTOLES .. 221 Records of auricular extrasystoles . . . . . 226 Arterial curves . . . . . ■ • • • 226 Venous curves . . . . . . • . . 226 Electrocardiograms . . . . . . ■ ■ 226 Sinus extrasystoles . . . . . . . . 229 XIX.— PREMATURE BEATS ARISING IN THE JUNC- TIONAL TISSUES ; RETROGRADE BEATS ; PREMATURE BEATS DISTURBING NEW RHYTHMS, ETC 230 Premature beats arising in the junctional tissues. . 230 Retrograde beats . . . . . . . . . . 232 Ventricular extrasystoles and complete heart-block . . 236 Atrio-ventricular rhythm and extrasystoles . . . . 238 XX.— PAROXYSMAL TACHYCARDIA OF SUPRA- VENTRICULAR ORIGIN 241 The point of origin . . . . . . . . 245 Auricular origin' . . . . . . . . 245 A-V nodal origin .. .. . . 248 Paroxysms of supraventricular origin in ivhich the ventricular form of venous pulse is seen . . 250 XXL— PAROXYSMAL TACHYCARDIA IN WHICH THE VENTRICULAR SYSTOLES ARE ABNORMAL.. 254 Clinical examples . . . . . . . . . 256 Contents. xvii CPAPTEB XXII.- -AURICULAR FLUTTER .... Experimental flutter Clinical flutter . . The auricular beats in electrocardiograms Responses of the ventricle Arterial curves paoe 263 263 264 268 270 272 XXIII.— AURICULAR FIBRILLATION (THE CLINICAL CONDITION) The clinical condition The radial pulse curves The venous pulse curves Certain waves which occur in diastole The electrocardiographic curves in limb leads . Complete irregularity of the heart is the result OE auricular fibrillation The irregular oscillations seen upon galvanometric curves are due to an inco-ordinate contraction of some portion of the heart, and are independent of movements of the somatic musculature The oscillations arise in the vicinity of the auricle. The excitation wave travels to the ventricle along the normal paths of conduction Conclusions from the clinical findings 275 27? 277 278 281 283 286 286 290 291 XXIV.— AURICULAR FIBRILLATION {continued) .. .. 293 Auricular fibrillation as it is seen in experiment . . 293 Experimental and clinical arterial curves compared 294 Experimental and clinical venous curves compared 295 Experimental and clinical electrocardiograms compared 296 The harmony between the records obtained in complete irregularity of the heart in man and in experimental auricular fibrillation . . ■ • • • ■ 301 Auricular fibrillation in the horse . ■ . ■ ■ 302 xviii Contents. CHAPTER PAGE XXV.— AURICULAR FIBRILLATION, HEART-BLOCK AND EXTRASYSTOLES 304 Fibrillation and heart-block . . . . . . . . 304 Complete dissociation and auricular fibrillation 304 Partial heart-blocTc and auricular fibrillation. . 306 The effect of vagal stimulation upon auricular fibrillation . . . . . . ■ . ■ ■ 307 Digitalis in auricular fibrillation . . . 308 Fibrillation and ventricular extrasystoles . . . ■ 310 XXVI.— VENTRICULAR FIBRILLATION 313 Records of fibrillation . . . . . . . . 316 Clinical fibrillation . . . . . . . . ■ ■ 318 Clinical records . . . . . . . ■ ■ . 320 XXVIL— THE HETEROGENETIC IMPULSE, AND THE INTER-RELATION OF EXTRASYSTOLE, PAROXYSMAL TACHYCARDIA AND FIBRIL- LATION 321 The hypothesis of heterogenetic contractions .. 321 The heterogenetic impulse as the prime factor in many disorders of the heart's action . . . . . . 324 XXVIII. —OBSERVATIONS UPON THE NATURE OF PAROXYSMAL TACHYCARDIA, FLUTTER AND FIBRILLATION 332 The nature of paroxysmal tachycardia . . . . 332 The nature of flutter . . . . . . . . 334 The nature of fibrillation . . . . . . 335 XXIX.— CAUSES DETERMINING EXTRASYSTOLES AND ALLIED DISORDERS 342 , The role of the heart nerves in causing extrasystolic irregularities . . . . ... . . . . . . 345 XXX.— SINO-AURICULAR BLOCK, SO-CALLED .. 348 General remarks . . . . . . . . . . 352 Contents. XIX XXXI.— CARDIAC SYNCOPE AND UNEXPECTED DEATH The effects of cerebral ancsmia Forms of cardiac syncope 1. Standstill of the whole heart . . 2. Slowing of the tvhole heart accompanied by lowered blood pressure 3. Standstill of the ventricle alone (a) Suddenly developed heart-block (b) Increase of pre-existing heart-block (c) Ventricular standstill in complete heart- block 4. Fibrillation of the ventricles . . 5. Accelerated, heart action XXXII.— THE VAGUS Respiratory arrhythmias Respiratory arrhythmia in experiment Respiratory arrhythmia in the healthy human subject Respiratory arrhythmia associated with patholo gical processes Disorders produced by vagal stimulation 1. Slowing . . 2. Standstill 3. Auriculo-ventricular block Compression of the vagus in the human subject Vagal disorders comparable to those produced by stimulation Phasic irregularity Prolonged sloiving associated with fall of blood pressure . . Prolonged slow action of the ivhole heart Standstill of the whole heart Auriculo-ventricular block Sino-auricular block Complex vagal irregularities . . General remarks PAGE 354 354 355 355 357 358 358 358 360 362 362 364 365 365 366 366 368 368 368 368 368 369 369 369 369) 370' 370- 371 371 371 XX Contents. OHAPTEK XXXIII.— ALTERNATION Changes in the degree of alternation Curves from the heart muscle Electrocardiograms Circumstances in which alternation occurs Alternation in the force of auricular contractions Alternation in auricular fibrillation . . The nature of alternation PAGE 373 374 378 378 378 380 381 381 BIBLIOGRAPHY AND AUTHOR INDEX 387 INDEX (AND DEFINITIONS OF Tl'SRilS AS THEY ARE USED IN THIS BOOK) . . . . . . . . . . . 441 452 Fig 1. A section (magnification, 330 diameters) taken througii the sulcus tcrminalis of a nio's auricle and coloured with ha?.matoxylin and Van Gieson's stain. The plane of the section is in the line of the SLilcus and at right angles to the wall of the auricle. The epicardium lies to the left, the endocardium to the right. The muscle bundles of the auricular musci-ilature are shown to the right, cut in their length and across. To the left beneath the epicardium, is the dense collection of peculiar tissue of the sino-auricular node. It is comjoaot, the nausole fibres are small asid interlaced, and they present frequent nuclei 4 large vessel is seen deep to this tissue, and numerous nerve fibres and ganglion cells are embedded in the node itself. (Published through the Idndness of Dr. Ivy Mackenzie.) Chaptrr T. SPECIAL ANATOMY. The sino-auricular node. Keith and Flack {372), while examining the mtisculature of the auricles, discovered a pecuUar mass of tissue. It Hes at the junction of the superior vena cava and right auricle.* (The position and extent of this node is illustrated in Fig. 2, 34, page 62, and 50, page 86.) " In the human heart, as in most mammalian hearts," these authors write, " an artery or arterial circle lies in the junction ; the artery is surrounded by fibrous tissue in which are numerous peculiar muscle fibres, some nerve cells and some nerve fibres. The nerve cells and fibres we find from dissection to connect with the vagal and sympathetic nerve trunks. "f " Our search for a well- differentiated system of fibres within the sinus, which might serve as a basis for the inception of the cardiac rhythm, has led us to attach importance to this peculiar musculature surrounding the artery at the sino-auricular junction." " In the human heart the fibres are striated, fusiform, with well-marked elongated nuclei, plexiform in arrangement, and embedded in densely packed connective tissue — in fact, of closely similar structure to the Knoten " (referring to the auriculo- ventricular node to be described presently). This description has been confirmed and extended by many workers {383, 385, 387, 692). The special neuro-muscular system lies at the junction of the free border of the appendix with the superior caval termination, and extends downwards along the sulcus terminalis for a distance of about 2 cm. in man. In thickness it is approximately 2 mm.. The muscular fibres are small, being but a half or third the breadth of those of, auricular fibres proper. This structure is termed the sino-auricular node (Fig. 1). * The node lies, as Oppenheimer (579) has recently shown for the human heart, immediately to the venous side of the remnants of the old venous valves. f Ganglia are scarce ; the nerves break into " a plexus of moniliform fibrils in very close relation to its muscle fibres" according to the Oppenheimers (5S0). Argaud (7) describes the nerve supply as coming from a plexus situate at the root of the right coronary artery. The nodal artery is a branch of the right coronary vessel (see Fig. 2). Very beautiful pictures of the nerve endings in the S-A and A-V nodes have been pviblished by Meicklejohn (553) more recently. According to this worker the nerve endings are highly specialised (in the monkey especially) and plexuses are formed around the nuclei of the muscle cells. (See also Eversbusch (154).) B 2 CHAPTER 1 . As I shall show, the heart beat starts in this node, and the contraction travels from it over the walls of the auricle and reaches the structures described in the following paragraphs. Fig. 2. A drawing to scale (nat. size) of the right auricle of a dog, seen from the right side. The preparation was hardened in situ and subsequently dehydrated and cleared with xylol. In this way the chief muscle bands and injected arteries of the auricle were displayed. The precise position of the sino-auricular node was ascertained by cutting serial sections through small blocks of tissue taken from the upper part of the sulcus terminalis. S. V. C. and I. V. G. = superior and inferior vense cavai. P. F.= right pulmonary veins. S.A.N.= sino-auricular node. The exact position of the node, its shape and extent vary in different animals within certain limits (see Fig. 34, page 62, and 50, page 86). The auriculo-ventricular connection. The anatomical connection between auricle and ventricle was discovered after Wooldridge {793) and Tigerstedt {728) had demonstrated functional continuity in the mammal, and subsequent to the elaborate researches .of Gaskell {215), by which the dependence of the ventricular rhythm upon impulses derived from the auricle was clearly established in the cold-blooded heart. The first account of muscular connections between the auricle and ventricle in the mammal, was given by Stanley Kent {375, 376) in 1892. His preliminary paper was succeeded in 1893 by His's description {317) of a clean band of muscle fibres running from auricle to ventricle. In 1904 the foregoing observations were substantiated {39, 619). The next advance in our knowledge was made by Tawara ( 720) . In his book a complete account of the junctional tissues was given, and the anatomy of the whole system SPECIAL ANATOMY . 3 and the connections with the network of Purkinje was described in great detail and in many species. These anatomical observations have been confirmed completely by more recent workers {125, 371, 486, 559). Observers are agreed that a single path of anatomical communication exists between the upper and lower chambers of the heart (see second footnote, page 13). The fibres of the junctional tissues may be traced from auricle to ventricle without break (Fig. 3). The system commences in the auricle in the neighbourhood of the coronary sinus and at the base of the auricular septum, where auricular fibres collect fanwise and interlacing unite with the auriculo-ventricular node. The node hes at the very edge of the auricular tissue at the posterior and right border of the septum. The bundle proper commences at the node, running almost horizontally forward and to the left, ensheathed and isolated in a canal,* and pursuing a course directly to the right of the central fibrous body of the heart to the pars membranacea septi of the ventricles. At the anterior part of this membrane, and a little in front of the anterior attachment of the median or septal segment of the tricuspid valve to the ring, the bundle forks. The left division of the bundle perforates the membrane, and still ensheathed Hes upon the upper border of the muscular septum, and enters the subendocardial space of the left ventricle at a point immediately beneath the union of the anterior and right posterior cusps of the aortic valve. Its further course is downward, and it may be traced as it branches freely under the endocardium of the septum on the left side. The right division soon becomes subendocardial and, coursing downwards, enters the moderator band, or its representative, and proceeds directly to the papillary muscles where it breaks into its arborisation. In its course it marks the old separation between right ventricle proper and infundibulum. The arborisation of the left division (Fig. 4, 5 and 6) starts upon the septum, and passes to the papillary muscles of the mitral valve in two main branches. The arborisation on the right and left side is directly continuous with the extensive and complex subendocardial network of Purkinje fibres, which lines the greater part of the interior of both ventricles (Fig. 5, 6 and 7). From this network direct communication with the ventricular muscle fibres takes place. The smaller ramifications of the network frequently bridge the valleys between the muscular trabeculse, and are there completely enwrapped by endocardium ; these bridges are conspicuous at the apex of the left ventricle in almost every heart, be it human or not. It is said that the bundle and its branches are isolated by connective tissue sheaths beneath the endocardium until the papillary muscles are reached, and that no union takes place with the ventricular musculature during the earlier parts of the distribution ; but this conclusion is based purely upon anatomical studies, and is not upheld by experiment (see page 125 footnote and context). * Which Curran {S4) regards as a, bursa. B 2 G H APT E n t Fig. 3. A specimen in the Royal College of Surgeons Museum, photographed with the kind permission of Professor Keith. A human heart seen from the front and right. The anterior walls of the right ventricle and right auricle have been removed. The inter-aviricular septum, the tricuspid valve, the papillary muscles (G), the moderator band {F) and interior of the infundibulum (H) are exjiosed. A lies in the right aiu-icular appendix. B lies in the jossa ovalis. E is placed below the mouth of the coronary sinus ; directly to the right of it in the figure an area of endocardium has been removed and the upper connection of the auriculo-ventricular node with the musculature of the septum has been exposed. It consists of a fan-shaped piece of mviscle which lies directly below D. A bristle has been placed beneath the fan. From this point the auriculo-ventricular bundle and its right division are traced as they lie on a, series of five bristles between D and F. The strand proceeds in a curved fashion to the membranous septum, which lies directly below C, and at this point the left division passes through the septum. The right division is continued upon the interventricular septum and enters and follows the hioderator band (F) until it reaches the base of the large group of papillary muscles ((?). The arborisation which commences at this point is not clear in the photograph. SPECIAL ANATOMY. Fig. 4. A specimen in the Royal College of Surgeons Museum, photographed with the kind permission of Professor Keith. The heart of a walrus dissected from the left side. The greater portions of wall of the left ventricle and left auricle (A) have been removed and the aorta has been divided vertically at its base (J) and the left half taken away. The inter- ventricular septum and the cusps of the aortic valve are exposed. The anterior cusp of the valve is fully exposed and the mouth of the right coronary artery is seen. Directlv beneath the posterior end (right hand end in the figure) of this segment, the left division of the aiiriculo-ventricular bundle enters the ventricle and immediately splits into two chief branches ; these branches lie upon two horizontal bristles, over which there has been a very small amount of dissection. The further subdivision of the branches is perfectly clear, the arborisations are carried in free strands across the cavity ; several large branches enter the papillary muscles, the bases of which are seen (P). Two long bristles are placerl behind finer branches of the coarse network. / lies on the inferior cava ; G on the pulmonary artery. Note the large collections of nerve tissue at the base of the heart ; bristles are placed behind the thick strands at O, H, and C. CHAPTER I .a s< g <« cS o3 § t3 .2 s ■2 '3 ■- I ^ -S '3 >. "" -^ S" (D CD 'O " 5P 'S S -"■ -c " o -B ^ - T3 0) 2 OS o e S s- ^ o IS El o -a - •^ >■- ° an > to IS 05 aj O SPECIAL ANATOMY. -« Jm "S u QJ d ,fl m J3 VfnouS ,— /\ IV .^V| f\aol(.aL Fig. 11. Physiological forms of the venous pulse. Pnlyqraphic records. — A, simultaneous femoral and jugular curves taken from a dog. B-D, simultaneous radial and jugular curves from human subjects ; a, c, and v waves are clearly shown in each curve. In D, v is split, and several additional waves, including 6, are present. Note the constancy of form from cycle to cycle in each instance. The time in this and subsequent polygraphic figures is marked in fifths of a second. AURICULAR AND VENOUS CURVES. 29 but in most curves the c wave is probably composite, including both venous and arterial elements ; in some curves, especially those which are taken with the receiver pressed more firmly on the tissues the wave is chiefly arterial. Especially in delicate curves two waves may be seen, the first of venous and the second of arterial origin {632, 772). The wave v in the veins is normally a pure affair of stasis, the blood accumulating in the veins while the ventricle is in systole. It is said that the delay in the appearance of v is somewhat greater than that of a {801). In many crude clinical curves it is double. For clinical purposes the onset of a is the chief event in the venous curve, for it indicates contraction of the auricle. The a-c interval, the time distance between the upstrokes of the corresponding waves, is used to indicate the auriculo-ventricular systolic {As-Vs) interval, which in its turn is taken as a measure of the capacity of the tissues to conduct impulses from auricle to ventricle. The a-c interval, which normally varies between 0-1 and 0-2 seconds, is of great clinical value. Its measure is, of course, relative rather than absolute ; it gives no exact measure of the As-Vs interval, but if the As-Vs interval varies materially from case to case or in one case, corresponding differences will be found in the a-c interval. In clinical venous curves, especially when the ventricular rate is slow, the line of the tracing often exhibits a general rise from the end of the depression " ,v " ^o ^^® commencement of the next a wave ; this ascent of the line is clearly seen in Fig. 11, C and D. It is evidently due to stasis and was termed by Morrow the second onflow wave ;* the venous volume increases as the ventricle fills and the pressure in its cavity rises. Often the greater part of the ascent occurs in early diastole and the volume is then maintained ; in these circumstances a distinct shoulder occurs on the diastolic .portion of the curve (see Fig. 11Z>). In early diastole too, a secondary wave known as b or h, first described by Gibson {225) and Hirschfelder {314), is often seen when diastole is long.* It is said to result from a floating up of the tricuspid segments and consequent closure of the valve in early diastole as the ventricle fills with blood. According to Eyster {156, 157) it occurs synchronously with or a little after the third heart sound,| and this sound is also attributed to secondary closure of the auriculo-ventricular valves. * Similar waves are found in intra-auricular pressure curves, t For an account and records of which see 113, 225, 45S, 722. Chapter III. GALVANOMETRIC METHOD. The history of electrocardiographic studies will be found in the original papers referred to in this and the succeeding chapter. The main steps may be described quite briefly. In 1856, KoUiker and Miiller (389) first demonstrated the presence of a current of action in the heart ; and they were able, by laying a frog nerve-muscle preparation in contact with a beating heart, to show the presence of two distinct electrical discharges at each beat of the ventricle. Their observations were "followed by those of a number of investigators {126, 685, 686,) notably Burdon Sanderson, working with the earlier types of rheotome and galvanometer. At a later period the capillary electrometer was used, and, employing this instrument, A. D. Waller (745), in 1887, first showed that it is possible to register the human heart beat. ' The first satisfactory curves from the mammaUan heart were obtained by Bayliss and Starling (27).* In 1903, Einthoven {110, 115) introduced his new instrument, the string galvanometer. The facility with which this instrument may be employed, and the precision of the curves obtained with it, is such that it has rapidly superseded other forms of sensitive galvanometer. It is an instrument which may be used as a routine method of recording the heart beats in experimental researches, and its introduction has brought the systematic electrical examination of patients within the field of practical medicine. The string galvanometer . Galvanometric instruments are based upon the principle of the interaction of a magnet and a conductor of current. In the familiar Kelvin galvanometer of the physiological laboratory, a small magnet to which a mirror is attached is suspended by a fine thread. The magnet is surrounded by coils of wire, and with the passage of currents through the latter the magnet is deflected, and a beam of light reflected from the mirror serves as an index of such movement. The string galvanometer, in its present form the invention of Einthoven, is built on the opposite principle. A straight conducting * In studying the earlier work on the electro-physiology of the heart, the following additional papers (122, 534, 535, 542, 746, 748) should be consulted. GALV AN METRI G METHOD. 31 strand lies between the two poles of a powerful magnet. Currents passed through the string induce deflections of it. The sensitivity of the instrument and the quickness of the movements have been increased by decreasing the weight of the string and by augmenting the strength of the magnetic field in which it lies. The poles of the magnet are closely approximated, so that Fig. 12. The latest model of the Einthoven galvanometer. A A are the poles of the magnet ; B B are the coils of the magnet, M N the terminals conveying constant current to it. D receives and condenses the beam of light which falls upon the string or strings through a mica window in the carrier. C carries the projecting lens, and the same tube has, at its exposed end, an eye-piece which focusses the light on the camera ; i? is a focussing screw ; S moves the microscope from side to side ; I'F is a stand carrying two prisms at V and is used to converge the images of two strings when these are employed. The carrier has a. number of adjusting screws ; H swings the whole carrier on pivots (K and L) moving the mica window to and fro across the beam of light ; J and J' alter the tensions of the strings ; T moves one string and thus adjusts the relative position of the two strings ; O O and P P are the terminals coimecting to the ends of the wires and through them the currents to be tested are conveyed. Carriers may be used which are fitted with a single string or with two strings. only a narrow chink separates them, and the field is saturated. It is in this cleft that the string is fixed, and its movements are observed by projecting its shadow upon a screen. The strings employed are extremely delicate, consisting of finely drawn platinum or of a film of silver over a finely drawn quartz or glass thread. In thickness the fibre is 0-002 to 0-005 mm.. The Cambridge model of this instrument* is shown in detail in Fig. 12 ; in this model two strings are held in a carrier {J to 0) pivoted above and * The galvanometric outfit here described has been specially adapted for clinical studies and experimental work by the Cambridge Scientific Instrument Company acting in co-operation with my laboraljory. 32 CHAPTER III. suspended between the poles of the magnet {A A). The carrier encases the strings and protects these deUcate fibres from air currents and dust ; it has attached to it the terminals which convey the currents to be tested (0-0 and P-P) and milled screws (J and Ji) for adjusting the tension of the strings and their positions in the magnetic field {H). The circuits m practice vary in detail in different institutions and according to the purposes for which they are employed. In Fig. 13 is shown diagrammatically the Cambridge switchboard, which has been specially arranged for rapid and accurate clinical observations. The magnet of the galvanometer is shown, and the string is represented by a fine dotted line passing vertically between its poles ; the ends of the string connect to the wires of the chief circuit (heavy lines). The continuity of the two main wires is interrupted, on the GALVANOMETER Fig. 13. Diagram of the connections of the galvanometer to the three bath electrodes for the limbs by means of the Cambridge switchboard. The arrangement of the patient is shown in Fig. 21. one hand by the calibrating switch and compensator, on the other by the selector switch ; the main wires eventually terminate in the lead switch. The purpose of the lead switch is rapidly to connect the two main wires to any pair of the three electrodes in which the right arm, left arm and left leg of the subject are immersed. In the position shown in the diagram, the right and left arm electrodes are in circuit {RA-LA). The patient's body and his two selected limbs complete a circuit with the string of the galvanometer, a circuit which, when the instrument is recording (position GALV ANOMETRI METHOD. 33 figured), is complicated only by the presence of the compensator. The selector switch has four positions, (a) the recording position as depicted, (b) an off position in which the galvanometric circuit is broken ; (c) and (d) positions in which the galvanometer lies in test circuits, the one containing a 4000 ohm resistance, the other containing no resistance ; these circuits are used when by a preliminary test an estimate of the instrument's sensitivity to standard currents thrown into it is desired, or when an estimate of the internal resistance (chiefly string resistance) is desired. The calibrating switch is used for a number of purposes. When the switchboard is first connected to a patient, this switch stands on the short circuit stop ; thereby the delicate fibre is safeguarded against the receipt of any current from patient to compensator (any such current then fiows through the short circuit wire). When all is ready, the switch is moved to the 1/100 shunt ; in this position approximately 1/100 of the current to be tested is thrown into the galvanometer (the rest passing through the shunt circuit) ; if this current is sufficient to defiect the string from its zero point, it is brought back to the zero by means of the compensator. The switch is now moved to the 1/1 0th shunt (which allows approximately 1/lOth of the tested current to pass through the galvanometer), and any permanent deflection of the string is again compensated. The process of compensation is repeated after the switch has been moved to the first zero position, at which point all the tested heart current passes into the galvanometer. The object of this procedure will be clear when it is reaUsed that currents of two orders are obtained from the patient : first, there is a constant current which comes from the skin and is due to activity of the sweat glands ; this constant current, if it is not compensated by stages, may be of sufficient magnitude to fracture the recording fibre ; it is neutraUsed and the string is maintained at zero by throwing a current into circuit in the opposite direction from the compensator ; secondly, there are the minute fluctuating currents produced by the heart beat, and these it is desired to record with the string at its most sensitive point in the magnetic field, namely, when it hes at zero. The three final positions of the switch 1, and 3 are used to cahbrate the excursion of the recording fibre. Movement of the switch from to 1, or from to 3, changes the E.M.F. on the string terminals by one or three millivolts, respectively. Obtaining standardised electrocardiograms. Electrocardiograms, whether clinical or experimental, should be standardised ; the standard now universally adopted is Einthoven's standard {111, 114), in which one centimetre of excursion in the final photograph is equivalent to one millivolt potential difference at the ends of the recording fibre ; the measurement of standard movement in terms of E.M.F. has the great advantage of neglecting the resistance of patient and the resistance of string, both of which are variable factors. The object of standardisation U CHAPTER til. is to keep records from all laboratories to the same scale of excursion, and to permit a comparison of the amplitude of the corresponding deflections, in a given patient from time to time and in different patients. It is effected most simply and with sufficient accuracy for cUnical purposes, by introducing the patient into circuit, compensating the skin current, and, lastly, by increasing the sensitivity of the fibre (by slackening it) until the introduction of three milhvolts deflects the moving string through three centimetres (see Fig. 14). Testing the properties of the string. — If standard curves of correct form are to be obtained, the response of the instrument to simple currents must be tested from time to time {463, 680) ; for the excursion and shape of the electrocardiogram may be modified by the properties of the string and the conditions of the observation. If an E.M.F. of one miUivolt is thrown into circuit {478, 680) while the patient is disconnected and the tension of the fibre is so arranged as to give an excursion of one centimetre, the curve obtained should be similar in outline to those shown on the right hand of Fig. 15. The string moves when the current enters it and takes up a new position one centimetre away. In arriving at the new position it describes a curve. Now the characters of this curve are important. Fig. 15 shows six electrocardiograms from the same man, and the six corresponding responses to an E.M.F. of one millivolt. The curves differ because they were taken with different resistances in circuit and consequently with different string tensions.* From above downwards, the resistances were increased and the string slackened correspondingly, all the curves being taken so that the standard deflection of one centimetre was obtained when one millivolt was introduced (deflections to right of strips). In the first place, the movement of the string in response to one millivolt should be " dead beat " ; it should not overshoot its new position. Overshoot occurs when the string is too tense (as in Fig. 14, at h), or when the recording fibre is too heavy. The defect of overshoot is not observed in the Cambridge instrument, in the conditions under which it is used for clinical purposes ; but in many instruments on the market it forms a prominent defect, increasing the excursion of the sharp deflections. In the second place, the movement should be of sufficient rapidity, the slacker the string, the more slowly does it respond ; the deflection times for the six strips of Fig. 15 are 0013, 0023, 0-028, 0045, 0-060 and 0-075 seconds, respectively, from A to F. Now the initial changes in the currents to be recorded from the heart are rapid and, if the quickest movement of which the string is capable is too slow to follow these changes, accuracy is lost. The distortion of the curves, especially of their rapid initial phases as the string is slackened, is well illustrated in the accompanying figure. The * Adding resistance to the main circuit decreases the sensitiveness of the recording instrument and the string must be slackened to compensate tills decrease so that the previous excursion may be obtained again. GALVANOMETRIC METHOD. U -p ■». CD f^ rt !§ © S (S p 03 O o m ^ a += O 60 a I O _2 'S § ^ ° O g -p " ■§ s t -o 4-H o r! x> G IS o ^ ^ t^ >c crt X ■a ^ bo (D a ^ CD 5 »,^ p m o I '-+3 1 § i J) -a CD H "^ -^ -O CD H CD CD '^ C t; CD ,a ■ ' -P to "m o .S "" • ^ I — IV £ i ^ ill S ^ "-+3 •^ -p d c " § fi 05 H CD -P p ■ c ^ S2 1^ © h, ^ O CD tn &D ^ C3 (D bo o bo a, rj CO _H -p tH ^ " § ° 2 5^^ i . .. s " ^2^ „ -o 53 ^ ^ 1^ ri CQ ^ CD - 2 o o ^ 3 o -e i-H p CO bb > a s (D CD U CO CD ^tO -p C c9 -^ 36 OH APT EH III. upstrokes and downstrokes are rendered more oblique and their amplitudes are decreased. The first two curves are identical because the degree of slackness reached by the string is as yet insufificient to cause distortion. -r~^ i ii ^^E= ^UA,LLLinjAU.LL Lhn i ^ pStfcgH^fei nrntmuiiiiiii J +l-M4-J-iH-M-t t H_J J^j .J-4 M-4-M-4-+^ Fig. 15. Six electrocardiograms from lead 1 and from a single subject, and the six corresponding deflection curves, to illustrate the distortion of curves when the string tension is too slack. As the string is slackened beyond a certain limit and the deflection time (the time of response to an E.M.F. of one millivolt over the excursion of one centimetre) increases, R and S are materially reduced in amplitude. Time in thirtieths of a second. In this and subsequent figures the lead from which the curve was taken is indicated in the left top corner. OA LV AN MET RIC METHOD. 37 There is an upper limit of tension and a lower limit of tension in sound instruments, the former yielding overshooting, the latter insufficient speed of movement ; between these limits accurate observations are alone obtained. For cUnical purposes the deflection time should be at least as low as 0-02 seconds {478), when a resistance of 4,000-10,000 ohms is added by means of the test circuit. This condition is more than fulfilled by the instrument here described of which the natural frequence is approximately 200-250 per second. Suitable contact electrodes. — The electrodes advocated for making contact with the patient are shown in Fig. 21. A porous inner vessel is filled with warm water, salt and well-washed cotton wool,* to give a mixture of porridge-like consistence ; this is surrounded by an outer vessel containing saturated zinc sulphate in which a sheet of zinc is immersed to which the leading-oif wire is soldered. Each electrode should be insulated from the floor. These electrodes are non-polarisable. That it is necessary to employ non-polarisable as opposed to polarisable electrodes in accurate electro- cardiographic work may be shown readily {473, 681). 1. If an E.M.F. of one millivolt is introduced (Fig. 16, curve 1, in) into the simple closed circuit of a properly tuned string galvanometric recorder, and is cut out later {out), the record is as depicted. It represents, as accurately as modern instruments will permit, the flow of current through the fibre. If the millivolt is introduced into the same circuit, in which is interposed a pair of non-polarising electrodes, separated by a fixed interval of normal saline, then, providing string tension and total circuit resistance remain unaltered, an identical record is obtained (curve 2). But if, in similar circumstances, polarisable electrodes (of platinum) are utilised, the record suffers distortion. If the electrodes polarise slowly, distortion is present (curve 3) ; if they polarise rapidly, distortion is the greater (curve 4). Distortion is due to the rapid development and maintenance of a charge on the surface of the electrodes, while current is flowing through them. In the second figured instance of polarisation (curve 4), (a) the ampUtude of the initial deflection, produced by the E.M.F. introduced, is materially reduced, owing to the rapidity with which polarity develops ; (6) although the E.M.F. is maintained, the string returns to zero, and (c) on cutting out the E.M.F., the string is deflected across the zero line, by the now unbalanced charge on the electrodes. All these effects are clearly defects in the record. If with readily polarisable electrodes a true record of a simple current change is unobtainable, it is unreasonable to believe that complex currents developed by the heart beat will be accurately recorded. In point of fact. * Warm water adds to the patient's comfort and decreases muscular tremor ; salt decreases resistance ; cotton wool supports the limb and decreases movement of the limb and of the surface of the fluid. 38 CHAPTER III. the resultant distortions can be predicted. The upstroke of an initial deflection R will be reduced in amphtude because it will be neutraUsed by the developing polarity. As R subsides, the string will deflect across the zero line and will produce an artefact resembUng 8 (if S is not present) or exaggerate 8 (if 8 is present) ; the final summit T will be similarly reduced in magnitude by neutralisation, and will be followed by an artificial depression as T subsides. That such distortion actually results is clearly shown by curves 5 and 6 in which each of these changes is to be observed. AO titcl7-0^iJi /mjllir/tlt |. ilsstc U t null I veil. Uct "a- T Ys) 1-^ 1 L. T mU hJ . ^ n % MNI r Ulc W/ W} Fi". 16. (Journ. of Physiol., 1915, XLIX. Proc. physiol. soc, L.). 1. One millivolt was introduced into, and cut out of, a simple closed circuit. Circuit resistance = 4800 ohnvs (string 2800 ohms ; added resistance 2000 ohms). The excursion of the fibre is " dead beat " as it should be, and 10 scale divisions in amplitude. 2. One millivolt similarly introduced into, and cut out of, the same circuit containing non-polarising electrodes. Total circuit resistance 4800 ohms (string resistance 2800 ohms ; electrodes and salt solution 1850 ohms ; added resistance 150 ohms). .S and 4. One millivolt introduced into, and cut out of, the same circuit containing polarising platinum electrodes (larger and sinaller electrodes) ; total circuit resistance 4800 ohms (string 2800 ohms ; electrodes and salt solution 250 ohms ; added resistance 1750 ohms). The sensitivityof the string was constant throughout the series. 5. Human electrocardiogram taken by means of the non-polarisable electrodes (of 2) ; standardisation so that 10 scale divisions = 1 millivolt. 6. Curves from the same heart and lead, taken with polarising platinum electrodes (of 4). The total circuit resistance and string sensitivity were identical in the case of the two electrocardiograms. Time in fifths of a second. In electrocardiography the use of electrodes \\'hich polarise appreciably is indefensible, because it is impossible to accept the standardisation of the corresponding curves against a known E.M.F., and because the curves themselves are apt to suffer distortion.* * As Pardee {586) has recently, and quite rightly, said, the polarisation effect is greater when the electrode surface is small (the electrodes in Fig. 16, 3, were larger than in Fig. 16, 4). It large simple metal contacts such as he advocates are used, it is quite true that no material polarisation is usually obtained ; but from time to time, and for reasons not fully determined, the polarisation of ^uph electrodes is considerable and definitely affects the values of the curves. GALVANOMETRIC METHOD 39 To test the electrodes it is sufficient to connect the two inner porous pots by means of a strip of washed cotton wool soaked in brine and, placing this pair of electrodes in the main circuit, to introduce an E.M.F. of one milUvolt ; the resultant deflection should be permanent, the string should show no tendency to overshoot nor to return towards the zero Une.* Photographing. —ThQ string hes vertically and its shadow is projected by an optical system (Fig. 17), consisting of arc light, condenser and Adjusroble Slih Cylindnccl Lens Fibre 9ie Objechve Subshage Condenser '" -ir-y-,^ Z|i, _.. ,, 260 »t — 110 « — 80 — Fig. 17. Sectional view of the optical arrangement, utilised in taking galvanometric records. The light derived from an arc passes through a first condenser and a cooling bath and is focussed upon the fibre by means of a subsbage condenser. The image is projected by the objective and eye-piece on to the cylindrical lens ; en route,, the beam of light is cut at B (where it comes to a focus) by the teeth of a rotating disc, the time-marker. The direction of the fibre movement is across the beam of light as indicated by the arrow C ; the shadow of the fibre is projected upon the cylindrical lens ; it is vertical, cutting the cylindrical lens at right angles ; its movement is in line with the axis of the cylindrical lens and the adjustable slit. The plate which records these movements runs vertically from above downwards (arrow A). microscope. The magnification adopted is usually 600 diameters. The vertical shadow falls upon a cylindrical lens, whose axis is horizontal and which focusses the light, cut by this shadow, upon a photographic plate ; the latter moves inside the camera, its movement being controlled by an oil cyhnder (Fig. 18), and moves vertically behind a slit in which the cyUndrical lens is fixed. The lens is ruled at milUmetre distances, and these lines are photographed on the plate as it travels ; these are the horizontal hues on the illustrations as they are published. f The shadow of the string moves from side to side, leaving upon the developed plate a white line, which varies in thickness according to the rate at which the string is moving. This line is black and its movements are up and down in the reproductions. The speed of the plate is under control, as is also the amount of hght admitted * According to Pardee (587), such overshooting may sometimes take place when non-polarising electrodes are used. It is seen when skin resistance is high, and, according to Pardee, is possibly attributable to the skin acting as a condenser surface. f Certain of these figures have been reduced in size for publication ; the values of the deflection are to be measured therefore against the millimetre lines and not by direct measurement. P 3 40 CHAPTER III. to the camera. For clinical purposes a shutter is employed by means of which three records may be successively taken side by side upon a single plate.* 1203 Fig. 18. The plate camera of the Cambridge outfit. ^ is a cylinder containing oil, B a hollow plunger which as it descends takes up the oil ; the rate at which the descent is accomplished is altered by a screw vernier at Q which opens or closes the communication between A and B. M is a starting release. The plate moves in a carrier inside the camera, its movement being controlled by the oil piston through pulleys {P, J) ; it is exposed behind the shutter. The latter consists of a slide, T, running in the grooves of Z, Z, and a second slide, N, with scales attached. The amount of plate exposed is controlled by the slide JV, the portion of plate by the slide T. As figured, one-third of the plate would be exjoosed, the gap through which the light enters the cylindrical lens being seen in the illustration. Time-markers. — The time-markings in the records of this book are various, because the curves have been taken during the development of the apparatus. The modern time-marker is a rotatory marker. A toothed * With good lighting 1 have found " Process " plates to give the best results. The}' are very slow, and consequently require excellent illumination ; they give great contrast and ensure black and white figures. QALV AN MET RIG METHOD. 41 wheel, driven by an electric current, and controlled by a tuning fork, revolves and cuts the beam of light where it is brought to a focus after it leaves the eye-piece of the microscope. As each tooth passes the eye-piece, it cuts ofE all light from the cylindrical lens for an instant, ruUng a clean line across the plate (this line is vertical and black in the reproductions : see Fig. 26). Simultaneous records. — Many methods have been devised for securing simultaneous electrocardiograms. Bull (44) used two galvanometers arranged in tandem fashion, the images of the two strings being projected upon the camera by the same eye-piece. The optical adjustments in this scheme need to be very exact. Fig. 19. The Lucas coordinate comparator. Upon a rigid cast-iron stand, B, two steel bars ( D D) are fixed, and upon these a carrier, B, which supports the microscopes (.4 A) slides from side to side. Fine lateral adjustments of the carrier are obtained with the screw, K, Through the right-hand microscope a finely divided millimetre glass scale is read; the scale is fixed in a frame, the lower edge of which is seen at E ; this scale may be adjusted laterally by means of the screw, M. The photographic record is placed on » circular glass frame, F, which may be rotated so that the lines of the plate are exactly horizontal. The glass frame rests on a metal carrier, F, which moves up and down against a counterpoised weight over the pulley, iV. The background is illuminated by electrical lamps held in the shield, U. Two galvanometers have been used, side by side, each with its own projection system {461), the light from the two being kept apart up to the point where it falls upon the cyhndrical lens. The disadvantage of this system is that the rotatory time-marker cannot be employed. Recently the Cambridge Scientific Company has introduced a single carrier fitted with two strings, the images of which are eventually brought conveniently near to each other by an arrangement of prismatic lenses 42 CHAPTER III. (see Fig. 12). It is perhaps the most practical device of its kind, and by means of it the simultaneous records shown in this book have been taken for the most part. By placing a moving lever immediately in front of and vertically at right angles to the cylindrical lens, additional records, of arterial pulse, venous pulse, respiration, blood pressure, etc., may readily be obtained, simultaneously with the electrocardiogram. Parkinson {588) has recently projected arterial and venous pulsations directly by placing the patient between the eye-piece of the microscope and the camera. The shadow of the skin, where it pulsates, is thrown directly upon the cyUndrical lens and the movement is magnified in the record by this optical projection. Such records are both accurate and beautiful. Measurement. — Minute measurements are made from the negatives by means of a comparator (Fig. 19). This instrument consists essentially of two microscopes of small magnifying power and fixed in a rigid carrier which slides horizontally on two steel bars. They are moved laterally by means of a millhead screw. Under one microscope is an adjustable plate holder, under the other a divided millimetre scale. Measurements in millimetres are taken of the intervals between the vertical Unes representing time, and between the desired points on the electrocardiogram ; the last named are then converted to seconds. By means of this de^dce, readings from good curves may be obtained with an error of 1/1000 of a second or less. Registration of heart sounds. — The method which I have employed (461) in registering heart sounds is very similar to that described by Einthoven {112) and Fahr {166). A large and sensitive microphone {M in Fig. 20) communicates with the air through thick-walled rubber tubing and a stethoscope end-piece {8). The tubing has a second outlet (0), which remains permanently open ; its purpose is to cut out the changes of pressure in the system of tubing which would otherwise result when the stethoscope is applied over an area of chest wall which is pulsating. The microphone circuit includes a dry cell, a make-break key ( K2), a rheostat {R = 1Q ohms), and the primary coil (10 ohms) of an inductorium or transformer {T). The secondary coil (coreless, 5,300 ohms) of the transformer is connected to a short-circuiting key ( Kl) ; this key, when closed, forms a shunt to the fibre of the string galvanometer {G). Sound vibrations transmitted through the stethoscope fall upon the microphonic plate, and by pressure produce variations in the internal resistance of the microphonic circuit ; the current flowing through the circuit varies accordingly and, varying, induces currents of higher potential in the secondary circuit ; these secondary currents flow directly through the string when the shunt {Kl) is open. A tense string is employed, and sound vibrations having a frequence of from 200-300 per minute are recorded without material damping. GALVAN OMUTR I METHOD. 43 The double fibre carrier is employed, one string being used to record sound, the other to record the electrocardiogram.* Fig. 27 is an example of a simultaneous electrocardiogram and the sounds recorded from the apex beat of a normal subject.! Fig. 20. {Quart. Journ. Med., 1912-13, VI, 441, Fig. 1.) apparatus used in taking heart sounds records. A diagram illustrating the * In the earlier observations of which Fig. 27 is an illustration, I used separate galvanometers, •j- Other papers on heart sound records will be found in references 120, 223, 364, 365, 751 and 752. Chapter IV. THE BROAD FEATURES AND TIME -RELATIONS OP THE NORMAL ELECTROCARDIOGRAM AND CERTAIN PRINCIPLES OF INTERPRETATION. The broad features of the normal electrocardiogram. The leads adopted. — Electrocardiographic curves may be obtained by- leading from various points of the body. In examining clinical subjects, Einthoven (111) adopts three leads, taking in pairs, the right arm — left arm (lead /), the right arm — left leg (lead II), and finally the left arm — left leg (lead ///) ; these leads have come into general use. It should be understood that the type of electric curve obtained from the heart depends largely upon the lead chosen, and that no two leads yield exactly the same picture (Fig. 21). The reason for this change with the leads will be explained at a later stage ; we may be content for the moment in noting it. Each lead is serviceable in given circumstances, for one lead may give information which another will not. All indicate the systole of both the auricle and the ventricle. The physiological type of human electrocardiogram. — The electrocardio- gram of man, like that of other mammalia, consists of two parts, an auricular complex and a ventricular complex. As a whole the electrocardiogram exhibits considerable variations in form from subject to subject even in health, but in a single subject it is constant under given conditions ; thus, electrocardiograms might be successfully adopted as a means of identification (478). In this chapter we shall consider briefly certain variations in the form of the physiological electrocardiogram. The auricular complex consists of a primary deflection in the upward direction. Adopting the terminology of Einthoven, it is termed the summit or peak P (Fig. 29). This summit, which is either rounded or pointed, is succeeded by a horizontal line (an isoelectric* period) or by a less prominent deflection in the opposite or downward direction. * The contacts are isoelectric when no current is passing through the string. NORMAL E LBGTRO CARDIOGRAMS 45 Fig. 21. Three electrocardiograms from a young and healthy subject. I. Leading from the right arm to the left arm. II. Leading from the right arm to the left leg. III. Leading from the left arm to the left lee. The three curves were taken separately and each standardised so that one centimetre of excursion represents one millivolt. The horizontal lines are ruled photographically at distances of one millimetre in the original curves and serve as a, convenient scale of measurement; each scale division is equivalent to 1/10 millivolt. This standard has been used for all the curves of this book except where it is definitely stated to the contrary. The time records thirtieths of seconds. Note the varying amplitude of the several deflections in the separate leads, and the appearance of a deflection Q in leads II and III in this instance. P represents auricular, Q, R, S and T ventricular, activity. 46 CHAPTER IV B'lg. 22. Photograph of a subject as connected for observation. The two arms and the left leg are used, and curves are taken from the three loads which are represented by the arrows drawn upon the figure. The zinc sulphate is placed in the outer vessels of the electrodes shown in this figure. NORMAL ELE CTRO CARDIOGRAMS . 47 The ventricular complex is, as a rule, triphasic or quadriphasic, and is constituted by the deflections R, S and T, or Q, R, 8 and T, of which R and T are directed upwards, while Q and S are directed downwards. J? is usually the most conspicuous summit in the curve and its duration is brief (usually 0-03 seconds or less). It is often preceded by a small and brief Fig. 23. dip, Q; it is often followed by a brief dip, S, which is of very variable amplitude. The opening phases of the electrocardiogram consist, therefore, of a summit, P, associated as we shall see with auricular systole, and summits 48 CHAPTER IV, and dips, Q, R and S, associated with the initial events of ventricular systole. The Q,R,8 group of deflections is of special importance and is one of the distinguishing features of a normal or physiological curve ; as I have pointed out, this group of deflections has a total duration of no more than 0-1 of a second, and usually constitutes less than one-third of the full ventricular complex (463). It is followed by a larger or shorter line which is horizontal,* during which the contacts are isoelectric, and the whole complex ends in a broad and prolonged deflection, T.^ 3? ^=ESE: " ^?F - - -'ur-' m:\mmu^\iAA M.\.\.\mA.u.ux.mi.n.WtU 'i ilti 1 i !• 1 1111 1 i 111111111 11 i 1 1 i 1 1 1 iTi 1 1 1 1 i 1 1 nil Fig. 24a. n.mn\\\xmimmumm\ti.^mm^ ^ Fig. 246. Fig. 24a. Normal human electrocardiograms showing the tallest R in lead III, and the shortest R and deepest S in lead I. Time in thirtieths of a second. Fig. 246. Normal human electrocardiograms showing the tallest R in lead I, and the shortest R and deepest .S in lead III. Time in thirtieths of a second. Variations in the physiological type in the several leads (477). — As a general rule R is most prominent in lead // ; but it may be largest in lead ///, in which case it is short in lead I, and *S is conspicuous in the same lead. It may also be most prominent in lead /, in which case it is short in lead ///, and S is conspicuous in this lead (see Fig. 24a and 246). Normal electrocardiograms often exhibit notching of P, and the summit of R, and base of S, are not infrequently split in leads // and ///. In lead III bizarre types of initial deflections (the Q,-^,*^ group) are not uncommon; examples are shown in Fig. 25. * Not infrequently absent. f In many normal curves T is followed, especially in lead II, by a. further summit V of small dimensions. It falls in early diastole and its meaning is not understood. NORMAL ELECTROCARDIOGRAMS. 49 '^==;^Emi^ n.UU.UUH,U.l.l.l.l.t.f.i.Rl.l.l.Uilll.l.l.U,IJ.I.Iil.l:M.I Fig. 25. Normal human curves taken from lead /// in three different subjects. Illustrating the curious arrangements of the initial ventricular deflections which sometimes occur in this lead. They are often associated with inversion of T. Time in thirtieths of a second. In the accompanying table the minimum and maximum and average values of the several deflections of the normal electrocardiogram, as they were found in 52 healthy subjects, are given (478). Lead /. P Q R S T U Minimum Trace 00 1-5 00 —0-5* 00 Average 0-52 0-51 516 2-06 1-93 010 Maximum 10 20 12-0 Lead II. 6-0 5-5 Trace Minimum Trace 00 40 00 Trace 00 Average 116 0-73 10-32 2-23 2-46 0-16 Maximum 1-7 2-5 16-5 Lead ///. 4-5 5-0 0-8 Minimum Trace 00 2-0 0-0 —2-0 00 Average 0-81 0-86 6-61 1-73 0-61 006 Maximum 1-5 2-5 14-0 4-0 3-0 0-3 The chief time-relations of the electrocardiogram. — Electrocardiograms have been taken simultaneously with records from the auricular and ventricular musculature (Fig. 26), with intra- ventricular pressure curves and with heart sound (Fig. 27) and polygraphic curves (Fig. 28) on numerous occasions. The records taken by different methods and by different workers are in fairly close agreement. They are portrayed in the diagram (Fig. 23). * The — sign indicates inversion of the deflection. 50 CHAPTER IV. m mi m 1 = i 1 — ~1^ i m M if !* __ - 1 1 ^ b 1 E F- 1— [ ^ ■i 5 s *^" J k f i 1 r 1 — — j _::;: *- ■ 1 - 1 — a ] 'i / 2 il- 1^1-^^' 1^ ^ ^ ^^^P i^ ■'1 l« ' wwm ^^^ ^ Fig. 26. Simultaneous electrocardiogram and myocardiographic curves, the latter taken by levers attached directly to the wall of the exposed heart. Taken from a dog. The vertical lines represent the time-marker, which records 1/5 and 1/25 seconds. The lines were ruled photographically while the curve was taken, and each cuts the three curves at precisely the samd instant in time. The relation between P the auricular summit and the upstroke of the auricular myocardiogram {A) (upstroke representing systole) and the relation between R the first ventricular deflection and the upstroke of the ventricular myocardiogram ( V) is displayed. Fig. 27. Simultaneous electrocardiogram and heart sound curve from a normal human subject. The figure shows the time-relations of the electrocardiogram to the beginnings of the 1st and 2nd heart sounds. All points on a vertical line are simultaneous. Time in thirtieths of a second. NORMAL ELECTROCARDIOGRAMS 51 Fig. 28. Simultaneous electrocardiogram, venous and radial curves from a young man. On account of the conduction of the polygraphia curves through the air of the rubber tubing, the venous and radial records are displaced to the right equally and by approximately 0-03 of a second. This delay may be allowed for all similar curves in this book. The relations between the several waves and deflections of the curves may be gauged if this delay is allowed. Time in fifths of a second. P stands in relation to auricular systole and its upstroke precedes R by from 0-13 to 0-21 of a second, this time interval constituting the P-R interval (observations upon 52 healthy young men) {478). Usually the interval lies between 0-13 and 0-16 of a second.* In dogs, the P-R interval is between 0-08 and 0-10, and in cats between 0-06 and 0-08 of a second. The upstroke of P precedes the upstroke of a in the human jugular curve by from 0-1 to 0-15 of a second. In six dogs the upstroke lay from -024 to -043 of a second before the commencement of the curve of auricular shortening. The upstroke of R precedes the upstroke of c in the human jugular by from 0-1 to 0-15 of a second. The beginning of the initial ventricular deflection (i2 or Q) usually precedes the onset of ventricular contraction (as estimated from myocardiograms from the front of the ventricle in six dogs) by from 0-020 to 0-038 of a second. This interval has been regarded as a measure of the latency of the contraction. But it has been shown by Kahn {363), and the writer is in agreement with him, that when the muscle is artificially excited in the neighbourhood of the attached myocardiographic lever, the interval is less (according to this author it amounts to 0.002 of a second or even less). There is a delay of approximately 0-02 to 0-03 of a second between the commencement of ventricular activity and its appearance and record on the surface of the heart. Kahn has recorded simultaneously * Schrumpf {694}, analysing the curves of 342 subjects, gives the normal interval at from 0-05 to 0-15 of a second. An interval of more than 0-15 of a second he regards as abnormal. In placing the upper limit of normality as low as 0-15 sec. Schrumpf is certainly in error. 52 CHAPTER IV. R of the electrocardiogram and the intra-ventricular pressure rise and finds a delay of approximately this extent, but Piper {602) using more exact methods has recently stated the delay to be less. The summit T falls during the systole of the ventricle, while the whole mass of the muscle is shortened. In comparison with intra-ventricular pressure curves it has been found to subside a few hundredths of a second before or after the end of the plateau (212, 213, 358, 602, 777). The relations of the electrocardiogram to the heart sounds are perhaps the most valuable which we possess ; the heart sounds are the only accurate standards for comparison in man {25, 166, 359, 364, 458, 461, 778). TABLE I (HUMAN). Heart rate. 77 77 74 100 95 108 75 68 81 72 102 90 78 90 75 The error in measurement in this table is probably no greater than 0-005 of a second in any instance. In the dog I find similar relations : TABLE II (DOG) Heart Beginning of Q Beginning of R End of T to rate. to 1st sound. to 1st sound. 2nd sound. 131 0-017 0-010 or less —0-010 130 No Q 0-00-t —0-051 121 0-024 0-018 0-023 130 0-029 0-027 —0-011 The measurements of Table I are in general agreement with those of other observers. The 1st sound begins in man from 0009 to 0-039 of a second after the deflection R begins. The relation of the end of T to the 2nd sound is variable ; it may fall 0-03 of a second before or after the sound, or in an intermediate position. beginning of Q Beginning of E End of T to to 1st sound. to 1st sound. 2nd . sound. 0-0.39 0-026 0-0 0-036 0026 - -0-015* No Q 0009 - 0-014 0-015 0-008 0-0 No Q 0-005 0-002 0012 0-006 0-005 No Q 0015 00 0-015 005 — 0-011 0-002 — 0-024 0-011 0-005 No Q 0-026 00 0-028 0-018 - -0-013 0-013 0-008 0-019 0-025 0-018 0-028 0-030 0-025 — -0-035 * The — sign indicates that the 2nd sound precedes T. NORMAL E LEGTROGARDI OORAMS. 53 If we take the upstroke of R and the end of T as evidences of the beginning and end of systole, the error in the former will probably not exceed 0-02 of a second and in the latter will not exceed 0-03 of a second. That P constitutes the auricular representative in the electrocardiogram is clearly shown by its time-relations to curves taken directly from the auricular muscle and by its occurrence whenever the auricle contracts, whether the ventricle responds or not (see Chapter XIII on heart-block). The deflections Q, R, S, are known to correspond to the initial processes of the ventricular contraction because they are related to this contraction in time, and because they are to be recorded when the ventricle contracts, whether the contraction follows an auricular contraction or not. The last evidence is to be emphasised in respect oi Q ; it is beyond question a ventricular event, occurring as it does in curves of complete dissociation of auricle and ventricle as part of the ventricular complex, as was first pointed out by Einthoven. Gertain principles of interpretation. Leading directly from the muscle. — When a simple strip of muscle is stimulated at one end, a wave of contraction passes from the proximal or stimulated end to the distal end. Associated with this wave of contraction there is what is termed a wave of excitation, and this follows the same course as the contraction wave, but actually precedes contraction by a very brief time interval. It is this wave of excitation which is responsible for the electrical changes as we record them. If a strip of muscle (Fig. 29, P-D) is connected to a galvanometer by means of non-polarisable electrodes in contact with its two ends, and if the muscle is then stimulated at P, the galvanometer exhibits two deflections. These deflections when recorded form a diphasic curve ; they are produced by what are sometimes termed " action currents." The first deflection is associated with activity of the muscle strip beneath the proximal contact ; the second deflection, which is of opposite sign, is associated with activity under the distal contact. The first deflection has the same direction as one produced when the proximal contact is placed on the zinc, and the distal contact on the copper of a copper-zinc couple. When muscle passes into a state of activity it becomes relatively negative to inactive muscle, in the same sense that the zinc of a battery is negative to the copper. This electrical change in the muscle induces a current which passes through the galvanometer from the inactive (-I-) to the active point ( — ), and the instru- ment records the passage of the current by a movement which in our photographs is upward. As the excitation wave travels along the muscle from P to D, sooner or later it reaches D and subsides at P ;* thus, D * According to the length of the strip and the duration of the active phase, activity at P subsides or begins to subside before or after activity is awakened at D. The diagram shows the second time- relation, for that is the relation which obtains in the heart. 54 CHAPTER IV, eventually becomes more active than P, and therefore relatively negative to it. The swing of the galvanometer is reversed as a consequence, the current flowing through it in the reverse of the original direction. The two phases of the curve are due to a change in the relation of the active point to the contacts as the wave travels from one end of the muscle to the other. The culmination or summit of the first deflection can be shown theoretically and by demonstration to coincide with the earliest arrival of the excitation process under the distal contact (Fig. 29&), for that is the instant at which electrical balance between the two ends of the muscle begins to be restored ; a little later, when both ends of the muscle are equally active (Fig. 29c), Fig. 29. A simple strip of muscle is stimulated at its right-hand end (P). The diagram shows (at d) the resultant curve and relates its phases to the events in the muscle. the two ends of the muscle are isoelectric ; at this stage no current flows through the galvanometer and the first defiection has been completed. There follows a variable period of equal activity, the isoelectric condition continuing. As activity begins to subside at the proximal end, the second deflection commences, the current flow through the galvanometer being reversed, and is completed as the muscle becomes quiescent (Fig. 2M). Now this experiment is a simple one and is readily understood, once it is known that active muscle is relatively negative to inactive muscle. It is fundamental, nevertheless ; interpretations of curves taken from the heart are largely based upon it, and in the next chapter we shall see more particularly how it is apphed in direct examinations of the heart in experiment. NORMAL E LE CTROCARD I OGEAMS. 55 This first law leads us to a second, which has a wide apphcation. It is that the direction, which the excitation wave takes* in travelling, governs the form of the corresponding curve. This second law in so far as it applies to direct leads may be illustrated by means of the same muscle strip, for if the latter is stimulated at D and the contraction wave is forced to travel from D to P, a diphasic curve is still obtained, but the phases as compared to those of the first curve are reversed (Fig 30&). The reversal must clearly occur, seeing that the order in which the ends of the muscle are activated is reversed. In interpreting abnormal electrocardiograms, which are taken by means of indirect leads, the known relation between the form of the curve and the direction taken by the wave is also constantly applied. A change in the direction of the contraction wave usually implies that the wave starts from an abnormal point ; thus, electrocardiography may acquaint us with the points from which the heart is excited to contract. Fig. .30. A simple strip of muscle is stimulated at its right (a) or left end (6). The diagram shows the resultant curves reversed. Indirect leads. — In human electrocardiography the electrodes are not in immediate contact with the heart muscle, but with the limbs, and through these with the thoracic tissues. The leads are indirect, and this fact is constantly to be remembered.^ A lead from the right arm to left leg in human electrocardiography is by no means equivalent to a lead from contacts placed in experiments on the actual base and apex of the heart. In human work the contacts with the heart are large ; they are constituted by the tissues surrounding the heart on all sides. When small electrodes are placed directly upon the exposed muscle the resultant curve chiefly expresses the events in those small sections of muscle which lie immediately in contact with the electrodes. In human work, in which the contacts are broad, the resultant curves are composite and represent much more in their due proportions the activity effects of all the muscle of the chambers. * Relative to the leading- off electrodes. tit 'may be emphasised by employing Samojloff's terminology {6S2), speaking of curves taken from indirect leads as electrocardiograms and direct leads as electrograms. E 2 56 CHAPTER IV. Those who regard a lead from the right arm and left leg in the human subject as a simple base to apex lead, assume that an upright deflection indicates relative negativity of that portion of the whole chamber which is nearest to the arm contact (i.e., where the ventricle is responsible for the deflection, the base of that chamber). This assumption is seriously open to question or misinterpretation. In the case of an upright deflection in a human electrocardiogram, it is more correct to assume that that 'muscular section of the chamber to which a single excitation wave is confined is so placed that the first part of it to be activated lies nearest to the arm contact. Thus, where the ventricle is concerned, it may perhaps be agreed to consider the point first activated as equivalent to P in Fig. 29. The usual physiological conception considers the whole of the remaining muscle of the chamber as equivalent to D from Fig. 31. Electrocardiograms from the three leads in a, case of transposed viscera. All the deflections of lead / are inverted ; the deflections of the remaining leads are natural in direction. Time-marker in thirtieths of a second. the very first. This conception is rendered difiicult, if indeed it is accurate, for, as we shall see, the excitation wave starts almost simultaneously at a large number of points. If a given excitation wave starts at a point in one small section of the ventricular muscle, that point may be considered as equivalent to P ; D, to my mind, is not represented by the remainder of the ventricular muscle but by that section of it through which this particular excitation wave will travel. The excitation wave in question has a limited territory, at the imaginary boundaries of which it meets neighbouring excitation waves, and meeting them is unable to pass further. Thus it would be a matter of indifference whether the section of muscle lies near the base or apex of the ventricle, providing that that part of it which first becomes negative Ues nearer the right shoulder, and that part of it which becomes NORMAL ELECTROCARDIOGRAMS. 57 negative a little later lies nearer the left thigh ; such deflection as is seen during the spread of the excitation wave will be upright. In other words, if the direction in which the excitation wave travels is, on the whole, away from the arm contact and towards the leg contact, the resultant deflection will, by its uprightness, proclaim the fact. The manner of ascertaining the more precise direction in which the excitation wave travels by means of the indirect leads of human electro- cardiography will be described in a later chapter when the occasion arises. For the moment we may consider the broader question, that the electrocardiogram indicates the course taken by the excitation wave as a whole. A simple clinical demonstration first accomplished by Waller {745) will suffice to show that the curve obtained from the heart is controlled by the direction taken by the excitation wave in its relation to the leading-off contacts. When the heart is normal and normally situated in the body, the deflections P, R and T, are upright in direction (see Fig. 21) in curves taken from lead / (right arm to left arm). But when a patient who presents transposition of the heart is similarly examined, the direction of all deflections in this lead is inverted (330, 574) (Fig. 31). The differences in potential between the right arm and left arm in a normal subject, during the progress of the heart beat, are precisely the same as the differences of potential between the left arm and the right arm in a dextrocardiac subject. A little reflection will show (as does Fig. 31) that this inversion will occur only in lead /, for lead / is the only symmetrical lead of the three used {583, 683). Inversion of all the deflections in lead / in the absence of reversal in the other leads is a reliable physical sign and probably the most reliable sign which we possess of dextrocardia. Chapter v. THE NORMAL PACEMAKER OP THE MAMMALIAN HEART. Iisi the hearts of the cold-blooded vertebrates, the great veins, systemic and pulmonary, terminate in the first chamber, the sinus venosus and the sinus and the true auricle into which it leads are clearly defined. Consequently it is a matter of little difficulty to show that the heart beat arises as high up as the line of union in these animals. A ligature or clamp placed upon the sino-auricular junction (1st Stannius ligature), thereby isolating the sinus, brings the lower chamber to a condition of standstill, while the upper portion of the dissociated musculature preserves its original rhythm. In the mammalian heart no additional chamber, in the form of a sinus, exists. It is natural to seek the remains of sinus tissue at those points at which the great veins enter the heart, and for many years the sulcus terminalis, a groove on the outer surface of the auricle, has been regarded as the line separating the representatives of the two chambers. Earlier workers were content to differentiate two portions of the auricle, one above and one below this line, and their experience, coloured as it was by contemporary morphology, spoke simply of a pacemaker in the region of the great veins. But recent anatomical research has sharpened our knowledge. Keith and Flack hold that there are sinus remnants in the region of the mouth of the superior vena cava, in the coronary sinus, in relation to and in the auricular septum, and possibly also at the mouths of the pulmonary veins, for they find masses of differentiated muscular tissue in these situations. Anatomical and morphological researches led Keith and Flack {372) to suspect that the pacemaker lies in one of the collections of specialized tissues found in the auricle, and directed their attention more particularly to the sino-auricular node ; for this is a relatively large collection of peculiar tissue, and it stands in intimate relation to the rich supply of nerves entering the heart in its neighbourhood. The statement of these views has been responsible, directly or indirectly, for the most recent observations. No sooner was it known that a mass of tissue, remarkable for the peculiar form and arrangement of its elements, exists in the neighbourhood of the superior cava, than experimental pathologists turned close attention to this region. THE MAMMALIAN HEART. 59 Of the many experimental methods which have been adopted in locating the position of the pacemaker in the mammahan heart, the chief may be summarized. Many of the observations have been upon hearts subjected to considerable mechanical or chemical injury, or upon hearts in which normal nutrition was disturbed. I propose, therefore, to deal first with the electrical methods, which have been undertaken in a manner free from these objections, for they are, to my way of thinking, the most perfect we possess for the purpose and are suiHoient to show the region of the auricle which first becomes active. Forcing an excitation wave to follow the natural path. It is known that the electrical curve yielded by an unexposed and naturally beating auricle exhibits very great constancy from animal to animal and from one species of animal to the next. We may conclude, therefore, that wherever the pacemaker lies it has a similar position in different animals of the same species, and in animals of separate species (man, horse, pig, dog, cat and rabbit) ; so that if there are several widely separated centres, from any one of which the excitation wave might be supposed to originate, it does not spring from this in one animal and from that in another. If we are satisfied that the sino-auricular node is the pacemaker in the dog, we shall be satisfied that it is the pacemaker in all mammals in which this node is present, and in which the heart yields the same type of auricular electric curve in a given lead. Now auricular contractions may be excited from any desired part of the auricular musculature by means of single induction shocks, and such experi- ments may be performed under extremely favourable conditions, with natural respiration, with the thorax closed and with the heart beating at a normal rate (445). Successive excitation of many points of the auricular musculature (right and left) shows with certainty that the auricular electric complex obtained when a heart beat is excited from the neighbourhood of the superior vena cava precisely resembles the normal auricular complex. The meaning of this fact is clear, for the form of the electric curve indicates the direction of spread. The path taken by the normal wave in the auricle is similar to that taken by a wave excited artificially from an area surrounding the mouth of the superior vena cava ; in other words, the normal and the excited waves are propagated from the same area. It is also found that auricular curves of a similar nature are yielded by excitation of no other area, but that in any given animal the type of curve obtained from inferior vena cava, coronary sinus, pulmonary veins, etc., bears no resemblance to the normal auricular curve.* These facts are illustrated in Fig. 32. * Rothberger and Winterberg state their belief that the curves of these experiments of mine were materially complicated by the break shocks of stimulation {665, p. 601). This was certainly not the case. Such deformity of the curves is usually avoided if threshold stimuli are employed, and if the electrodes are close together. Any deformity arising from this source is easily recognised and allowed for, and should but rarely occur. 60 CHAPTER V . Fig. 32. Three electrocardiograms from a dog, each taken by lead II. The last three cycles in each strip of curve correspond to natural heart beats. The early cycles of each strip are responses to stimulation in the region of the sino-auricular node (S. A. N.), the inferior vena cava {I.V.O.), and the tip of the right auricular appendix (R. App.). In each instance the forced beats resemble the normal beats in so far as the ventricular elements (-R and T) of the curve are concerned, but only when the beats are forced from the region of the sino- auricular node are the auricular elements (P) the same for forced and normal beats. Time in fifths and twenty-fifths of a second. The point of primary negativity estimated by the direction of deflections. We have seen that when a strip of muscle becomes active at a given point, this point becomes negative relatively to all other points. This law has been familiar since its statement by Hermann. Negativity is readily shown by a galvanometer, for, if the recording instrument is connected by means of non-polarisable electrodes* to two points upon the surface of a muscle, primary negativity of one contact is shown by a deflection of the * These electrodes consist of small glass tubes plugged with salted kaolin and half filled with saturated copper sulphate solution in which copper wires are immersed. The contact with the surface of the heart is effected by means of cotton threads smeared with kaolin and imbedded at gri? end in th? kaolin plug. THE MAMMALIAN HEART. 61 string in a Known direction. This principle permits us to test our hypothesis that the first active point of the auricular muscle is in the region of the S-A node. If one contact is placed over this region and the other is placed successively upon points lying around it, then, if activity is always first developed under the central contact, this contact should always become primarily negative to the outlying contact. That is to say, if such radiating leads are taken, maintaining the central contact upon the point at which the excitation waves arises (Fig. 33), a series of electrograms should be obtained, in each of which the first deflection is in a given direction ; the direction of this deflection ought always to indicate primary negativity of the central point. POLd. _|_ ve/w. I -h s.i/.c. sdPTun. 4- AUR-WALL. Fig. 33. A diagram illustrating the electrical condition of the S-A node at the beginning ot auricular activity. The muscle in this region becomes relatively negative to all other points on the auricular surftice. Now there is only one superficial region of the mammalian auricle which exhibits these electrical relations when the heart is acting normally. If one contact* is placed over the region of the sino-auricular node and the other contact is moved to any other point on the auricular surface, it matters not where the second contact is placed, the first deflection obtained with auricular systole is upward in direction and indicates relative negativity of the centre point. A diagram illustrating the contacts in an actual experiment is shown in Fig. 34. Examples of the electrograms are shown in Fig. 35. * That which gives, when connected to the zinc terminal of a copper-zinc couple, an upright deflection. 62 CHAPTER V. This type of experiment goes a long way to convince, for the muscle of the auricular waU is thin and the S-A node lies in what may be regarded as the centre of a muscle sheet, that is to say, points may be chosen for examination all around it. There is little or no chance of this region receiving the excitation wave from a deeper structure, for all the muscle bands running towards the node may be tested. The conclusion that it is the centre in which the excitation wave starts is strongly suggested by these observations, which were first undertaken by Wybauw {794, 795), and in my own laboratory {440, 485), nine years ago. Since the original reports appeared. Fig. 34. ( Heart, 1910-11, II, 158, Fig. 4.) A diagram of a dog's right auricle showing a series of paired contacts. The contact T^ was maintained at the cephalic end of the sulcus terminalis, the contact paired with it was moved to different points of the surface of the auricles and great veins. In each instance the first deflection corresponding to auricular activity was upright, indicating primary negativity of T^. The general direction of spread of the excitation process, as indicated by this method, is shown by means of arrows. The distribution of the sino-auricular node, ascertained histologically, is shown by the dotted line. I have repeatedly confirmed them ; they have also received recent confirmation by Eyster and Meek {161). In the pig, the sino-auricular node Ues higher up the sulcus than in the dog, and in one animal of this species examined the point of relative negativity was found in a corresponding position (464). It lay, as Dr. Ivy Mackenzie was subsequently able to show histologically, immediately over the sino-auricular node in this animal, as it had lain in all our experiments upon dogs. The mammalian heart. 63 Extrinsic and intrinsic deflections. Before pursuing our discussion, it is desirable to consider what may be termed " outlying " leads — leads in which neither contact lies over the head of the 8- A node. Such a lead is illustrated in Fig. 34, S'^-S'^. In leading directly from the heart muscle, the chief deflections occur when the excitatory process is produced or arrives immediately beneath the contacts ; the contacts are exposed to the full force of the electric discharge in the active tissue underlying them. Such leads are essentially different from those used in human electrocardiography, for in these the electrode contacts are upon the limbs. Curves of the excitation wave may Fig. 35. (xfj) Three photographs, each showing a simultaneous electrogram and electrocardio- gram in a dog. In each photograph the lower curve was taken by a constant lead (lead //). The upper curves (or electrograms) were taken directly from the exposed auricle of the dog. In the left hand figure the contacts both lay near the cephalic end of the S-A node ( T^ to T'' in Fig. 34). In the middle figure the upper contact was maintained { T^), the lower being placed near the midpoint of the sulcus ( T''). In the right hand curve the upper contact was maintained at I"^, the lower being placed at T" on the inferior vena cava. In each curve of the direct leads the first deflection starts sharply and is prominent and upright. Each of these deflections is almost simultaneous with the beginning of P in lead II. At the beginning of the auricular systole, therefore, the region of the S-A node was primarilj' negative to the three other points on the sulcus. Time in fifths of a second. be obtained by both methods, namely, when contact with the active structure is direct or when it is indirect through inactive tissue, and the electric effects of the two orders should be distinguished. Now, when contacts are placed on the muscle of the auricle they produce deflections of two kinds {483). The chief deflections are those which result from the arrival of the excitation process immediately beneath the contacts ; these are termed intrinsic. They are deflections, which represent relatively large electrical potentials and they have correspondingly large amplitudes.* The deflections of the * Exceptionally the intrinsic deflection is not the most prominent in the electrogram. 64 CHAPTER y. second order are those yielded by the excitation wave travelling in distant areas of muscle. These are qualified by the adjective extrinsic. Intrinsic and extrinsic deflections may be illustrated by a simple experiment. If two contacts are placed upon the sulcus terminalis, then at each beat of the auricle a large intrinsic, deflection is produced as the auricular excitation wave reaches one contact. But the same contacts also record the subsequent discharge of the ventricle. These last effects are extrinsic, representing the activity of distant muscle elements. A similar double effect is noticed in respect of the auricle itself. If we lead from two contacts lying over the right auricular appendix, an example of an outlying lead, we obtain a curve ^1 {MS^~= lJEf<^/J - ^= ^^* #^ _j — • 'jo,. jk/^^-> lAtir ^=: i^^h MPmiiiP" Elf^ViiVHlM f^MllliHI "gj^ \- (=> !^! © "S g ^; i c g e 3 OS £ m "-< ;^ S 0-. 1 fe, h-n-^ ■-I ^^ i-i (S ^ =« ° (3 m o a, SS.2 C > "i ■** 5 TJ 3 7! !^ S J= O CI -^ O a> g- . J3 .S OS o ^ l-l n 4-< 10 ^ .. -t? r^ i-i ctf ^. i-H J^ t-H (B ■n fi CJ S-^ node lies at the one end of the series, and the time at which the excitation wave appears at each contact of the series is estimated by leading from each succeeding pair. It is found that the time increases steadily, and, when the contacts are equidistant, by equal increments as the leads recede from the S-A node. arrives at contacts 2 to 6 are estimated, relative to its arrival at contact 1, these readings are found to increase in a regular order. The excitation wave moves at a uniform rate from S-A node to the end of the taenia. HEADINGS FOR CONTACTS AKRANGED SEHIALLY AND IX THE LINE OF THE S-A XODE ON THE T.ENIA AND C'AViE. Tfenia contacts Superior caval contacts Inferior caval contacts* 8 mm. apart. 5 mm. apart. ,'5 mm. apart. •0000 -0131 -0201 •0083 -0194 -0238 •0159 -0265 -0287 •0200 ^0325 -0343 •0251 ^0357 .0394 •0447 •0519 Headings for contacts a to g, in Fig. 48, SPREAD Of THE EXCITATION WAVE. 83 T.T, Kg. 48. {Phil. Trans., 1914, B., GOV, 37a, Fig. 14.) A diagram of a dog's auricle to scale, showing two series of contacts on the inferior cava. The muscle bands are shaded. T.T = taenia terminalis. The curves obtained from each pair of contacts, a-h, b-c, etc., to h-i, are charted relative to the same standard time instant in Fig. 49. A similar proof is forthcoming that the excitation wave moves up the superior cava and down the inferior cava. Thus in Fig. 48, the contacts a to g, lying on the taenia and inferior cava, gave the readings shown in the last column of the above table. The corresponding electrograms are illustrated in Fig. 49 ; all these curves are plotted relative to the same standard, namely, the time reading of the excitation wave at its onset in the S-A node ; the gradual recession of the intrinsic deflection as the lead is moved away from the node is clearly shown. It is noteworthy that the intrinsic deflection is always inscribed when one contact lies over auricular muscle, but when the edge of this muscle is passed (as in lead h-i, Fig. 48) the intrinsic deflection disappears (Fig. 49). 84 CHAPTER VII. b-c c-d I I I I — r~r-i I e-f Fig. 49. A series of electrograms charted in relation to the S.A.N, line (the line representing the onset of the excitation wave in the auricle). They were taken from the corresponding leads a-b. etc., depicted in Fig. 48. Investigating the auricle in this fashion, and knowing the distances between given contacts and the 8- A node, we may estimate the rates of conduction to all parts of the auricle. SPREAD OF THE EXCITATION WAVE 85 TIMES AND RATES OF TUAXSMISSION OF THE EXCITATION WAVE TO VARIOUS PARTS OF THE AURICLE, Region. Average distance ill mm. Average transmission, time in seconds. Average transmission, rate in mm.per sec. Number of observations. Intercaval region 15-2 ■0139 1232 18 Intra-aur. band 12-9 •0126 1252 6 S.V. cava 8-2 ■0136 588 11 Septum (mid & low) 31-5 ■0305 1059 11 Rt. appendix 28-0 ■0314 955 11 Rt. auricle 16-0 ■0206 859 7 Rt. pulm. vein 24-0 •0254 1121 4 I. V. cava 31-5 ■0325 998 18 Coronary sinus 43-9 ■0412 1096 5 Left pulm. vein 45-2 •0412 1118 5 Left appendix 44-6 ■0446 996 7 Average heart rate 158^4. These calculated rates, are, with minor exceptions, wonderfully uniform, such differences as occur being explained by error of measurement and by the arrangement of the muscle bands. The chief variant is the superior cava, where, apparently on account of the obliquity of its fibres, the rate is notably slower. There is no evidence of slow conduction from the node itself to adjacent muscle fibres, neither can I accept the observations of Eyster and Meek {161, 163), as showing that there is more rapid conduction from 8-A node to A-V node.* The wave spreads between the two nodes through the muscle of the septum, as it does through other auricular muscle bands, at a rate approaching 1,000 mm. per second, * If we are to accept the findings of these workers, it must be allowed that the spread of the excitation wave in the auricle varies fundamentally in its course through the auricles of different dogs. Such variations are opposed in general by physiological experience, and in particular by the considerable uniformity of the P summit in lead 77 in different animals. Eyster and Meek's conclusion is based chiefly upon studj' of curves obtained from paired contacts on the auricle, A contact is placed on each of two points under observation and the two contacts are led to the galvanometer ; the point first showing relative negativity is ascertained. This method is fallacious, as I have pointed out {483), if it takes no account of extrinsic effects ; in practice the distinction between extrinsic and intrinsic effects is often difficult or impossible, if the lead is not in the line taken by the excitation wave, G 86 CHAPTER VII. The spread of the wave may be likened to the spread of fluid poured upon a flat surface (474), its edge advancing in an ever widening circle until the whole surface is involved (Fig. 50). SUP.VZN.CM. SUP. INF.CAV, APP. Fig. 50. A diagram to illustrate the manner in which the excitation wave spreads over the surface of the right auricle. The spread is from the upper part of the node and follows the chief muscle bands at almost uniform rates. The arrangement of the auricular muscle guides this mode of spread. The chief bands run from the region of the S-A node. This node is placed in the most advantageous position for quick distribution, the muscle fibres run from its neighbourhood in distinct bands (see Fig. 2, page 2) ; the tsenia runs from the top to the bottom of the sulcus ; the strong intra-auricular band runs from the top of the sulcus, behind the aorta to the tip of the left appendix ; the chief muscles of the right appendix radiate from the tsenia ; other fibres run from the sulcus down the septum. To sum up : the excitation wave, originating in the S-A node, spreads at once and at rates ranging around 1,000 mm. per second along the chief auricular muscle bands and these radiate from the neighbourhood of the node. It spreads as fiuid does when poured on a flat surface, involving an ever increasing area, and finally progresses against the blood stream at the mouths of the veins to end upon these where the muscular sleeves end. It reaches the A-V node by spreading through ordinary auricular muscle, and passes into the A-V bundle. Its course through this bundle was proved in the last chapter. In the ventricle The problem of spread in the ventricle is more difficult ; for the events succeed each other with greater rapidity. It has been studied in the first instance by examining the surface distribution of the wave (486), and for this purpose a single contact is placed upon the epicardium, while a second SPREAD OF THE EXCITATION WAVE 87 Fig. 51. {Phil. Trans., 1915, B., GOV I, 181, Fig. 21.) A projected drawing (x tV) of the interior of a dog's right ventricle, showing the relations of the large papillary muscle and free arborisation of the right bundle division to the surface. The overlying tracings are projections of the superficial muscle fibres and of the surface of the heart with the contacts and readings of the experiment. a 2 88 CHAPTER VII. rests upon the body wall. The intrinsic deflection, consequent upon the arrival of the excitatory process beneath the epicardial contact, is always upright in the corresponding curves and may be clearly recognised and timed if this method is constantly employed.* The surface distribution in the dog may be exemplified by Fig. 51. If a series of readings from contacts overlying a superficial band of muscle fibres is studied, for example the muscle band which sweeps from the conus across the upper part of the interventricular groove and around the left border of the heart to the apex, it "is at once evident that the wave of excitation does not follow the band. Fig. 51 serves as an illustration ; in this instance the readings -0241, -0231, -0198, -0150, -0146, -0187 and -0196 sec.,f are found along the muscle band in question. It is activated almost OOfg ■0001 J), ex 03.66 o/gt 0/1(2 0/8S Fig. 52. {Fhil. Trans., 1914, B., GCVI,196, Figs. 6 and 9.) ( xf.) Surface distribution in two dogs over right and left ventricles, seen from the sides, i . Fo = Vortex of left ventricle. simultaneously throughout its whole length. This and other observations clearly show that the arrangement of the muscle bands does not materially influence the spread, as iS the case in the auricle. * Erfmann [HI) has used Clement's differential electrodes (51) {i.e., closely paired contacts on the ventricle), but the results obtained by this method are open to objection in that while intrinsic and extrinsic deflections are recorded, these often cannot be distinguished in leading from the ventricle, because the direction of the intrinsic deflection is not actually knoum beforehand. In practice the use of differential electrodes has not proved satisfactory when exploring the ventricle. I These readings are the time intervals between the beginning of B in lead II and the arrival of the excitation wave at the corresponding contact points. SPREAD OF THE EXCITATION WAVE. 89 If readings are taken from the whole front surface of the heart, then the region of the right ventricle that lies parallel to the ventral attachment of the free wall to the septum is always found to be activated first ; but this area, in which the excitation wave appears early is of considerable extent ; many points are activated almost simultaneously and, if the underlying structures are examined, it will be found that the area corresponds to the free arborisation of the right bundle division, an arborisation which is very prominent in the dog (Fig. 51). Now the rest of the right ventricular surface (front and back, see Fig. 52) is activated later and in a constantly increasing degree as we pass towards the base of the heart. Although this order of the readings is constant, yet the time differences between them are so small that the order cannot be explained by direct surface spread at the transmission rate ascertained for ventricular muscle. (This, as we shall see, is about 400 mm. per sec.) In the case of the whole auricle, the progress of the wave from start to finish occupies some four or five hundredths of a second ; in the ventricle, despite the larger size of this chamber and despite the slower conduction rate of its muscle, the surface is supplied in three-hundredths of a second or less. The order over the left ventricle is quite as definite (Fig. 52) but it need not be described in detail. Suffice it to say, that the vortex (L.Vo) is the first surface point activated, while the central and basal parts are excited later, and that over a large part of the central region the wave appears at a number of points almost simultaneously. No system of spread from point to point of the muscle fibres can be imagined to explain this distribution. We are compelled to assume that different parts of the surface are activated along distinct channels. These channels, as can be shown, are constituted by the arborisation and network of the bundle divisions, and the excitation wave spreads in the ventricle from within outwards. The Purkinje pathway. — In a previous chapter, changes in the type of the electrocardiogram have been stated to accompany lesions of the chief branches into which the bundle divides. The reason for this change, namely, altered distribution of the excitation wave, can now receive proof. If readings are taken from a series of contact points on right and left ventricle (Fig. 53) before (upper readings) and after (lower readings) the termination of the right branch (R.B.B.) has been divided, it is found that the readings over the left ventricle remain unaltered.* But over the front of the right ventricle the change is profound and the magnitude and order of the readings is such that the right ventricle is clearly shown to be activated by spread from the left ventricle. Similar, but even more profound changes are seen when the left branch is divided at or near its point of origin. * Except for a very small and constant difference, which is attributable to alteration in the shape of the standard curve of measurement. 90 CHAPTER VII. Fig. 53. {Phil. Trans., 1914, B., CGVI, 200, Fig. 12.) Natural size projection of ventral surface of a dog's heart, showing readings for a series of contacts before (upper reading) or after (lower reading) cutting the right division of the bundle. The line of union of the two ventricles is marked by the chief blood vessel. The cut and its relation to the branch (R.B.B.) and papillary muscle {P.M.) is shown below. The main divisions of the bundle are thus proved to be concerned in distributing.* It can also be shown that the network of Purkinje is concerned ; for if a reading is taken from the surface of the conus, a transverse cut or even a scratch, placed on the apical side of the contact, conspicuously delays the reading, providing the cut or scratch is upon the endocardial lining of the heart ; a similar cut in the epicardial surface, even though it penetrates * The branches are capable of conducting in either direction (486) though normally they conduct only in one. SPREAD OF THE EXCITATION WAVE. 91 the muscle deeply, is without effect; the spread takes place along the endocardial surface. Furthermore, if four deep cuts are made on the surface so as to surround a surface contact, then from this contact the readings before and after the interference are the same. No experiment points more definitely to the progress of the excitation wave through the muscle from within outwards; none shows more distinctly that the wave is independent of the direction of the muscle bands. Conduction rates.— li an artificial wave is promoted by stimulating the surface of the heart in line with two contacts (Fig. 54), the time interval Fig. 54. (Phil. Trans., 1914, B., OCVl, -01, Fig. 14.) A diagram illustrating an experiment in which it is shown that the conduction interval between two contacts is uninfluenced by cutting the superficial muscle fibre lying between them. Each muscle contact is paired with its own chest wall contact, and the records are taken simultaneously. The stimulating electrode is placed in line with the two heart contacts. between the activation of the muscle under the two contacts can be accurately determined, and a rate of transmission can be calculated. For different parts of the ventricular surface this rate varies. It is highest and approaches or surpasses 2,000 mm. per second where the muscle wall is thinnest ; is lowest and approaches 400 mm. per second where the muscle is thickest. The reason for this variation is that the rate of conduction in muscle proper is slow, while in Purkinje substance it is very rapid. If the wall is stimulated where it is thin, the artificial wave penetrates the whole muscle thickness and is conveyed along the Purkinje network, from which it spreads outwards through the muscle again to reach the contacts ; if, as in the left ventricle, the wall is thick, the wave may be conducted across the superficial contacts before it travels to the lining of the heart and out again. These conclusions follow from experiment. The border of the right ventri^3le is stimulated (Fig. 54) and an artificial excitation wave is propagated 92 CHAPTER VII. from B to A ; the times of arrival at the two points is ascertained. A deep cut in the muscle between the contacts does not affect these readings ; a shallow cut or scratch on the endocardial surface at once delays the arrival of B. Clearly the wave is carried along the endocardial lining over the greater part of its course. If contacts are placed on the pericardial surface (P) and on the endocardial lining (E) of the wall (Fig. 55) and the surface is stimulated r .P£r-^. 1 ■ ->^ - - - ■> " ^7 a. cu \Tr 1 f-n(7n --> — 2^- _J E Fig. 55. {FMl. Trans., 191i, B., GGVI, 208, Fig. 15.) A diagram showing the alternative paths (a-a, 6-6) which an artificial!}' induced excitation wave may take in the heart wall. at some distance from them, the excitation wave is found to reach the internal contact first and, if the stimulating electrodes are far enough removed, the interval between the readings at the two contacts is precisely the same as for the natural heart heat (natural readings). In brief, the excitation wave reaches P by the path b-b, for that path is the quicker on account of the high rate of conduction in the lining. But if the stimulating electrodes are nearer to the contacts, the wave may reach P along the path a-a. If the thickness of the muscle wall is known and the distance from the point of stimulation is also known, then, in an experiment in which the natural readings at the two contacts are just maintained, the relative rates of conduction in Purkinje substance and muscle can be ascertained. The rate is at least five times more rapid in the network than in the muscle. The experiment is arranged so that the excitation wave reaches P along the two paths a-a and b-b simultaneously ; the length of muscle traversed along a-a and the length of muscle and network traversed along b-b is ascertained and from these data the conduction rates are calculated. When the results of experiments of this and other kinds are considered, it is calculated that in muscle the rate of conduction is in round figures 500 mm. per second ; in straight (as opposed to the usual wavy) Purkinje strands it is approximately 5,000 mm. per second. SPREAD OF THE EXCITATION WAVE. 93 Further arguments. — When the heart beats naturally, readings from the lining of the heart are always earlier than readings from the surface at corresponding points, and the difference between such readings is controlled by the thickness of the intervening muscle. Readings from the lining of the conus are later than readings from other parts of the lining of the right ventricle ; and this is so because the Purkinje path to the conus is the longest. Thus the distribution at any point on the surface of the ventricle is controlled by two factors, the length of the Purkinje path, and the thickness of the muscular wall ; if these data are ascertained for a number of surface points, and calculations are made on the basis of estimated transmission rates in the two classes of conducting tissue, then these calculated readings are found to correspond in a surprisingly exact manner with actual readings from the surface. The chief controlling factor is the muscle thickness, and it is partly for this reason that the attached border of the right ventricle is activated early, for here the muscle is thin ; it is wholly for this reason that the vortex of the left ventricle is activated early, for here the muscle is often extremely thin. Certain time-relations between the spread of the excitation wave in the heart and the deflections of the axial electrocardiograms are of importance. Activity begins in the auricle at an average time interval of 0-01 of a second before the beginning of the deflection P. Activity in the ventricle begins approximately 0-005 of a second before the beginning of Q in an axial lead. The galvanometric string, arranged at tensions such as it is customary to employ, fails to record the progress of the excitation wave in a limb lead, until a considerable mass of muscle is involved. A comparison of the surface and lining readings shows clearly that the whole period of the initial phases, Q, E and S, is occupied by the spread and development of the excitatory process in the ventricle. The total duration of these deflections corresponds to the activation of the muscle* and it constitutes an approximate measure of the duration of the spread. A final word completes this chapter. The conduction rates in the heart muscle increase as we pass from ventricular muscle to auricular muscle and from these to Purkinje tissue. The glycogen content of these tissues and the breadth of the fibre increases in the same order. The A-V node has the finest fibres and the glycogen content is poor (569) ; there is much evidence in proof of its having the lowest conducting power. Thus, there is a suggestive relation between the power to conduct and structural and chemical constitution. Distribution in the auricle is expedited by the central position of the pacemaker, by the arrangement of the muscle and the relatively high rate of muscle conduction (i.e., 1,000 mm. per second). The muscle of the ventricle conducts slowly [i.e., 500 mm. per second) because its function of distributing * A conclusion which applies not only to the dog's heart, but to the amphibian, reptilian and avian heart (475). 94 CHAPTER VII. is a minor one ; but this, the driving chamber, is provided with a special system of distribution, clearly arranged to provoke almost simultaneous contraction of all parts of the walls ; these special fibres are endowed with conduction powers of the highest order (5,000 mm. per second).* Note. — In this chapter it has been possible only to summarise recent observations ; for further details of the work the original papers should be consulted. It has also been considered undesirable to expand the chapter by including a description and criticism of many earlier attempts to trace the course of the wave of excitation and contraction. A number of the earlier writers sought to unravel the excitation wave in the ventricle by analysing the curves obtained from paired contacts resting on the ventricle ; the last exponent of this method was Gotch {230, 231). The method is insufficient and Gotch was led astray, as were his predecessors, by treating the whole ventricle as a simple strip of muscle. These workers were either unacquainted with or failed to grasp the high significance of the Purkinje conducting system. This historical aspect is dealt with in a recent paper (475). It is also unnecessary to describe the Relatively crude attempts to determine the precedence of contraction in the two auricles or two ventricles by mechanical devices (677), or of different portions of one ventricle by the same means (139, 281). There are parts of the right auricle which contract before parts of the left auricle, there are parts of the left which contract before parts of the right ; the same statements apply to the ventricles. Mechanical methods of recording, even as they are now developed, are impotent to solve such questions. The order of contraction may be presumed to be the order of activation, and that is described in the present chapter. Note. — For recent observations on the spread of the excitation wave in the ventricle of the lower vertebrates the following papers may be consulted (51 and 475). * Some direct measurements of the rate of conduction in Purkinge strands have been o'utained by Erlanger (146). The rate as estimated by him is 750 mm. per second. The observations were undertaken in the perfused heart of the calf, and therefore in all j)robability underestimate the Purkinje conduction rate very considerably. The same writer has shown that Purkinje fibres are excitable to artificial stimulation, but whether they contract or not is still unknown. Their striation certainly suggests this power, as does also the elaborate protection by sheaths which they enjoy in the ungulates. Chapter viii. THE MEANING OF CERTAIN VENTRICULAR DEFLECTIONS. Q, R, S, THE INITIAL DEFLECTIONS. The duality of the normal electrocardiogram. In the last chapter it was shown that the excitation wave in spreading throughout the ventricle follows the branches of the A-V bundle and the arborisation of Purkinje ; the wave courses from the endocardial to the pericardial surface of the musculature. During the stage of invasion it gives rise to the initial deflections of the electrocardiogram. In the first part of the present chapter, the constitution of these deflections will be described in so far as it is at present understood. It may be reiterated that we have proof that Q, R and S correspond to the stage of invasion or spread, for the time interval covered by these deflections corresponds with that covered by direct readings from the ventricle. If the diagram (Fig. 29, page 54) maybe used for comparison,* these deflections correspond to the first phase of the curve there illustrated. Inasmuch as the excitatory process passes to the ventricle by two distinct channels, formed by the right and left divisions of the bundle, and since no anatomical union is known to exist between the arborisations of the two sides, it might be assumed that the spread to a given ventricle is a distinct process and is confined to that ventricle ; that is to state that the excitation wave does not cross the interventricular groove during the natural heart beat, but that the right and left waves meet somewhere in the septum. This assumption is supported by two considerations. First, it is supported by the actual readings from the surface of ventricles, for these show an abrupt change of Order in the region where the two ventricles meet {486). Secondly, it is sujjported by the effects of branch lesions already described. When one division is cut, the distribution to the contralateral ventricle is unaffected right up to the septum, for the excitation wave spreads in normal fashion through the uninjured division ; but the spread through the homolateral ventricle is completely altered and now proceeds across the septum from the ventricle whose supply is uninjured. * strictly speaking, I do not think this comparison is quite justified, in that Q, R and S are written in an indirect lead and the diphasic curve of Fig. 29 is obtained by a direct lead. Actually, however, a very similar diphasic curve is yielded by leading indirectly from a simple strip of muscle. 96 CHAPTER VIII. This being the case, we are justified in supposing that the natural ventricular electrocardiogram is in reaKty a composite picture, depicting the superimposed effects of the separate ventricular activities.* This assumption has recently been proved correct by means of experiments devised to that end (475). When a single division of the bundle is transected, the excitation wave passes naturally to the opposite ventricle ; it courses through it in a normal fashion but yields a highly abnormal curve, for the ventricle is then activated in the absence of a balancing activity in the other chamber. But the curve may be termed normal in the sense that it represents the electrical events so far as the activation of this normally excited ventricle is concerned. This statement clearly applies to that portion of the curve which represents spread in the ventricle directly supplied from the auricle ; it does not apply to that portion which represents cross spread (abnormal spread) in the other ventricle. It applies, as has been shown, to the prehminary phases, or the sharp deflections, of the electrocardiogram. If the prehminary deflections are produced by spread confined to the right ventricle (as when the left division is cut), I term the curve a dextrocardiogram -."f if by spread confined to the left ventricle (as when the right division is cut), a levocardiogram.'f Now it is possible to obtain both the dextrocardiogram and levocardiogram of the same animal in experiment by alternately squeezing the left and right divisions of the bundle, and temporarily throwing them out of action as conducting structures. It is also possible, by modifying a method of mensuration described in a previous chapter, to chart these curves, so that corresponding time phases he vertically the one above the other (Fig. 56). When such a chart is accurately constructed, we have before us a complete and accurate statement of the electrical forces which combine to form the physiological electrocardiogram ; and being complete and accurate, both in respect of voltage and time, the physiological curve may be constructed from them accurately and in detail by a simple mathematical process. The two curves are combined by algebraic addition. Thus, if on a given time line, the voltage in one curve is 10 and in the other curve — 8, the corresponding voltage in the combined curve is 2. The chart (Fig. 56) shows the dextrocardio- gram (Rt.) and levocardiogram (Lf.) of a dog as recorded electrocardiographic- ally ; it shows the calculated normal curve ( C) and the actual normal curve ( N) in the same animal. The calculated and actual curves are alike in shape, they are alike in their amplitudes and have similar time relations. The proof is forthcoming, therefore, that the first portion of the normal electrocardiogram is a composite of two curves ; it represents the summated effects (for other examples of summation, see page 158) of right and left * An idea which has been mooted already as a hypothesis by Selenin (696', 697), Rothberger and Winterberg {666), and others. t In my original paper (475), I spoke of dextrogram and levogram. The present terms are modified to suit Samojloff's terminology. VENTRTGULAR DEFLECTIONS A Z\ 3\ 4\ s\ 6l 7\ b\ ^~\ ■CIsecond 97 ^'■0^- ;^ ■^^ .^ ^*. ^ |a iAM 1 1 0600 ■060O4dZSl 0450 ■030O •0050 ^-^ ,f/QO -0150 •0200 -OZSO Fig. 61a. ■005C 0300-0100 ■0250 Fig. 61a. (Phil. Trans., 1916, B., CCVII, Part IV, 284, Fig. 1.) A chart of the initial phases of human electrocardiograms, taken from a heart presenting signs of deficient conduction in the right division of the bundle. Ordinates, 1 cm. = 1 millivolt ; abscissse, 1 cm. = 0-02 sec. A diagram showing the angles of the corresponding electrical axes and the rotation. Fig. 616. (Phil. Trans., 1916, B., GCVIl, Pari IV, 284, Fig. 2.) A similar chart and diagram, taken from a heart presenting signs of deficient conduction in the left division of the bundle, Ordinates and abscissse as in Fig. 61a. VENTRICULAR DEFLECTIONS. 107 a number of human electrocardiograms, charted and analysed in this fashion, and by utilising as a basis the fuller data obtained in experiment upon the dog, it becomes possible diagrammatically to represent the distribution of the excitatory process in the human ventricle, recording against the endocardial and pericardial surfaces of the heart, as represented in section, the approximate times at which the electrical disturbance reaches the various regions of the musculature during the normal ventricular cycle. And this may be done, as in Fig. 62, with some pretence to accuracy. ■0300 •0350 b'ig. 62. (xj.) A diagram of the human heart as seen in section. It represents the author's considered view of the directions in which the excitatory process spreads in the human ventricle, and the times in seconds at which, after its commencement in the ventricle, this process first reaches the various regions of the ventricle. Now if the general conclusions here set forth be accepted, and the considerable body of evidence in our possession at the present time is in their favour, they satisfactorily explain a number of phenomena. Some of these may be briefly sketched at once and elaborated in the next chapter. According to my view, the human electrocardiogram, taken from whichever lead, is a composite of left and right effects. Its earliest events are those associated with activation of the septum on right and left side ; the activity 108 CHAPTER VIII. of this part is responsible for Q* and the commencement of R in all leads. Q and the chief part of the limb R are right-sided events in leads II and ///, left-sided events in lead I. 8 is & right-sided event in lead / ; a left-sided event in leads II and ///. Corresponding deflections are to be found in the dextrocardiogram or levocardiogram as the case may be. Briefly, we have an explanation of all the initial deflections of the electrocardiogram, which not only harmonises with physiological observation, but which is in complete accord with contemporary anatomical discoveries. We are also provided, as the next chapter will show, with a clear conception of those abnormal clinical curves which are associated with preponderating hypertrophy in one or other ventricle, an explanation which we have lacked hitherto. The end -deflection " T." Upon a written discussion of T in the electrocardiogram I enter with some reluctance ;t while desiring to refrain from airing such conceptions as have passed through my mind, yet it is hardly possible for me to do so without leaving an obvious break in the continuity of this book. In the circumstances I can but refuse to commit myself further than to indicate in a general way my reflections on this subject and the basis from which, as it seems to m.e, further observational work might begin. Nevertheless, I fancy I have this advantage over writers who have previously taken this course, namely, that the detailed path of the excitation wave in its spread through the ventricle seems clear to me, while from them it was, at all events in large part, hidden. But there are fundamental points in respect of which we still lack definite or final knowledge, and until such knowledge is won a full explanation of the deflection T and its variations does not seem possible. In a previous chapter the usual conception of the electrical events in a simple strip of muscle, forced to contract from one end, is represented diagrammatically (Pig. 29, page 54). Muscle, when it becomes active is known to show relative negativity to inactive muscle. In the muscle strip of the illustration the wave of activity passes from P to D, and, as I have described, when it arrives and reaches its full force at D, P and D, assume the same electrical state and remain for a variable period isopotential. But it is supposed that the point P maintains the electrical state relative to the inactive point which it assumes at its first activation, and that the point D after its own activation behaves likewise ; for this * In the amphibian and reptilian heart, in which there is no septum, Q is not seen in the electrocardiogram. It has been asserted repeatedly that the frog's electrocardiogram presents identical initial deflections to these of the human electrocardiogram. This is not the case, for in the former Q is always absent. t In speaking in this chapter of T, I mean to refer only to the end-deflection in the electrocardiogram and not to end-deflections in direct leads from the heart. VENTRICULAR DEFLECTIONS. 109 supposition is compatible with a continued isopotentiality of the two points* and explains the final deflection of the string. It is supposed that the electrical disturbance is maintained at its height at both points and that isopotentiality is continued, until the excited state of P begins to decline ; a deflection of opposite sign to the original deflection is then produced. Although this hypothesis is very possibly in the main correct, yet there is no proof that the electrical state of an active muscle point is constant in degree throughout any prolonged period of the systole. It occurs to me that the full change, which is undoubtedly manifested by muscle as it enters the active state, may be a more transient effect than is commonly supposed and that Burdon Sanderson's classical diagram {685), which is reproduced in modified form in Fig. 63, may fail, in that respect at least, to convey a true impression of what happens. It seems quite possible that the electrical change in the initial phases of activity may be of a different order quantitively to that of the late stages and that the plateau of the diagram should not be represented as the highest point reached. In testing active and uninjured muscle we obtain no measure of the electrical change at one point ; it is the difference of potential between two points, the one active, the other as yet inactive, which we observe. Moreover, we have no measure even of the approximate duration of electrical activity at one point. We assume that the duration is from the instant when the excitation wave arrives to near the termination of the end-deflection ; but even this assumed interval can be ascertained only approximately, for the end-deflection subsides and reaches the base line very gradually. In expressing the following view of T's constitution, I do so subject to these reservations which I have touched upon. The state of activity in a simple muscle strip may be described as comprising three phases : — (1) a stage of invasion, a stage during which the excitation wave is spreading ; (2) a stage of 'possession, or a stage in which isopotentiality is found as a rule between all parts of the muscle and during which activity is suf posed to be maintained in full force ; and (3) a stage of retreat, during which activity is thought to subside in the order in which it came. With the first stage as it applies to the heart I have dealt fully and in so far as this portion of the hypothesis is concerned we may rest content that it is sound ; we now know from direct observation, what has hitherto frequently been supposed, namely that the initial phases of the electrocardiographic curve correspond to the spread * This period of isopotentiality is not always to be observed in direct leads even in the frog and tortoise heart, where its production would seem to be especially favouredby thelong duration of the contraction. Actually there may be during the middle period of systole an almost constant deflection of the string to one side of the zero line. Further, before the isopotential state is reached, there is often in direct leads from the heart, a second deflection in an opposite direction to the first. The isopotential period in human electrocardiograms is usually brief, and in many cases non-existent. no CHAPTER VIII. of the excitation wave and that the presence of excited areas, alongside of unexcited areas, disturbs the balance of potential. It is in regard to the second and third stages that there is uncertainty ; there are some writers indeed who believe that when the last phases of the electrocardiogram {T deflection) are inscribed the lack of potential balance is due, not to the decline of the same excitation process which in its spread produced the ^' a- ■^-^ Initial PHasE Eni PKaSE Fig. 63. Sanderson's diagram (modified) to explain the two chief deflections given by the cold-blooded ventricle. The actual curve obtained from the heart, and shown below in the diagram, depends upon differences of potential at the two contacts (33 = basal ; A = apical) ; these are expressed by corresponding differences between the two curves (B and A) measured vertically. The two curves B and A are drawn of identical duration and form, but B is placed somewhat earlier in the figure, to represent the earlier arrival of the excitatory process at the heart's base. In Sanderson's diagram the curves B and A rise abruptly until they reach the level of the plateau and do not rise above this level ; the rise represents the stage of invasion. The plateau is represented by a long continued horizontal line ; this represents the stage of possession. Finally, the curves B and A descend gradually to meet the zero line ; this represents the stage of retreat. The curves are theoretical ; we have no direct knowledge of their forms. It is possible that the highest jioint of each curve is reached, not in the plateau, but in the stage of invasion, as indicated in the dotted lines of the diagram. The presence of these early summits would alter the res\.iltant curve as indicated by the dotted line in the lowest curve. initial deflections of the curve, but to the decline of a totally distinct process, namely, the contraction of the muscle (160, 325, 682, 695). The chief source from which the last view seeks support is that the end deflection is said to be more pronounced the stronger the ventricular contraction, and that where there is no visible contraction T may disappear while the initial phases persist (325, 682). But the waxing of T in the VENTRICULAR DEFLECTIONS. Ill electrocardiogram with the strengthening of contraction is not without notable exceptions ; in alternation of the heart (see Chapter XXXIII) T frequently alternates in height, and the large T may correspond with the weaker ventricular beat. The view that T is unrelated to the excitation process is not supported by observation ; on the contrary there is some direct evidence that they are intimately connected. Thus, Mines (555) observed what he was convinced was complete cessation of contraction in the heart, while the galvanometer showed strong regular deflections of the usual general form. This result he obtained by perfusion with a calcium- free Ringer solution, and he pubhshed a figure showing a conspicuous end-deflection. Other evidence we shall examine directly. Misconception has not infrequently arisen in that many writers appear to have tacitly or openly regarded the end-deflection as a manifestation pecuhar to the ventricle or to a particular part of it. This is not the case ; a similar end-deflection is found in records of the auricular systole, though, as the deflection is a sHght one and usually buried in the initial phases of the ventricular curve, it is not often seen [159, 292, 555) . A similar end-deflection is produced by the truncus arteriosus of the tortoise heart {475). But we may go much further than this and assert that the end-deflection is not peculiar to a single chamber of the heart or even to heart muscle ; a corresponding deflection appears in the final phases of the contraction process in a strip of muscle, a fact which has been recognised for very many years ; and it is a matter of indifference whether the muscle is somatic or cardiac* The deflection comes from a lack of potential balance and this, according to the most generally accepted present day view, is attributable simply to the retreat of the excitation process. But it is to be emphasised that of this retreat, and especially of the path it pursues and of its rate of progress, we have little definite knowledge. It is considered probable that the rate of descent of the excitatory state is more gradual than is the rate of its ascent. The steepness of the intrinsic defiection is the measure of the latter in curves taken direct from the muscle. The rate of descent is judged from the rate at which the end-deflection declines. The initial deflections arise abruptly, T fuses almost imperceptibly with the base line ; so it is felt that the excitatory process dies away slowly, and this view is expressed in such diagrams as that of Fig. 63. A simple strip of muscle gives two deflections, an abrupt deflection in one direction and a more leisurely one in the other. These two deflections correspond to the stage of invasion and of retreat, and for that reason, although they both correspond to a movement in one direction through the muscle, it is the rule that they are. opposite in sign (see Fig. 29, page 54). The more detailed application of this view to the mammalian electrocardiogram may be introduced by citing the relatively simple ventricle * In suitable circumstances it appears in both direct and indirect leads. 112 CHAPTER VIII. of the amphibian or reptile. In this single ventricle the excitation wave spreads from the A-V ring along the almost straight muscle bands of the interior to reach the main wall of the ventricle at points midway between base and apex (Fig. 64). Subsequently it courses downwards to the apex and upwards to the base. In the simplest instance, the base is first reached and from first to last the general movement of the wave is from above 0800 0600 0700 Fig. 64. {Phil. Trans. Roy. Soc.,19ir,, B., GOVII., 236, ParlI.,Fig.6.) (X5.) A diagram of the toad's heart, seen in coronal section, and of the manner in which the excitation wave is supposed to spread through it. A = auricular muscle passing into A - V ring. C =■- endocardial cushion. S = auricular septum. T.A. = commencement of truncus arleriosus. downwards. The electrocardiogram opens, as we should expect, with a single prominent upward initial deflection R in leads II and III (representing an electrical axis of approximately 90°). As in this example the last point to become active is the ventricular apex, so it is to be expected that the apex will form the last point in the retreat. In other words it is assumed, though it is not proved, that the order of retreat is the same as the order of invasion. As in the case of the simple strip of muscle, it is to be antici- pated that the end deflection (T) will have the reverse (i.e., downward) direction to the initial deflection (R). This is found actually to be the case.* But T is not always a downward deflection in the amphibian electrocardio- * It is to be pointed out that this and subsequent statements applies to the elecvrocardiogram , a curve obtained from the heart by contacts placed on the surrounding tissues. I am not speaking of the electrogram, which, taken by direct leads, chiefly comprises the intrinsic deflections from small areas of muscle in immediate contact with the small electrodes. VENTRICULAR DEFLECTIONS. lU gram ; more often it is upright ; indeed, this seems invariable when the heart is httle disturbed. In such curves, in my experience, R does not stand alone, but is followed by a distinct downward deflection S (representing an electrical axis of approximately — 90°), and the ascertained order in which the excitation wave spreads shows a corresponding difference. In such examples, although during the greater part of the invasion the average direction of spread is downwards* through the heart, the last points to be activated are discovered by direct observation to be at the extreme base and the average direction of spread at the last is upwards (as in Fig. 64). If we assume the same order of retreat, the base will leave its last impression on the electrocardiogram and the upright T is attributed to this dying activity. f According to this view both S and T in the amphibian and reptilian electrocardiogram are basal effects, the one produced in the stage of invasion, the other in the stage of retreat. They are therefore opposite in direction, S being downward and T upward. The rule may be formulated for these electrocardiograms that the direction of the end deflection (T) is opposite to that of the last initial deflection. That is precisely what we are led to expect from the simple muscle curve, and from the hypothesis which attempts to explain it. The rule may not be absolute, but it is nearly if not quite so. Personally I have seen no exception to it. Thus, in the cold- blooded heart, beating naturally, we possess evidence that the direction of T is related to the original path of spread of the excitation process. That is of much importance. Is there a similar relation in the mammalian electrocardiogram ? In the human electrocardiogram, of which we have most knowledge, it is the rule to find an upright T if the last deflection of the initial group is downward. It is also the rule that when T is a downward deflection, the last of the initial deflections is upright. There appears to be a similar relation to that obtaining in the cold-blooded heart, though it is not so rigid, for an upright T is not uncommonly seen in curves which present no S in the initial phases . Thus, there is in the human electrocardiogram sufficient evidence of a close, though not constant, relation between the direction of T and the form of the * When the average direction is downward, the points of relative negativity (or activity) stand more closely related to the base of the heart, while the points of relative positivity (the inactive parts of the heart through which the wave will subsequently spread) stand more closely related to the apex. I again use direction of spread as a convenient expression, knowing that technically it may be open to objection (see footnote, page 104). •j- Thus we come in modified form to the view expressed by Gotch {230, 231), that, in the naturally nourished and undisturbed heart, activity is first developed in a region related to the base (though not at the actual base as he supposed), that the first spread is to the apex, and that subsequently there is movement towards the base. The movement is not, as Gotch supposed, a relic of the S shape of the embryonic heart, for the excitation wave does not spread as a continuous wave from base to apex with a return to base ; originally flowing downward in the inner musculature, the wave proceeds downward and at the same time is reflected up towards the base. If it reaches the actual base a little later than the actual apex, and this is the rule, the final unbalanced movement is upward at the base. 114 CHAPTER VIII. initial group of deflections Q, R,S; the relation is particularly to the direction of the last initial deflection, as in the case of the amphibian and reptilian heart.* In the illustration of the simple muscle strip, the deflection produced in the stage of retreat is opposite in direction to that of the stage of invasion. We might therefore be led to expect that the final and initial phases of the human electrocardiogram would show equal complexity, but that the direction of the several end-deflections would be the reverse of the initial deflections. But there is at least one good reason why the end-deflections should be of simpler form than the initial deflections ; if we are correct in supposing that the decline of the excitation process is more gradual than its ascent, we should expect the final defiections to blend with one another : we should also be led to anticipate that the current produced in the last * The directional relation of the end-phase and the initial phase would be emphasised, if a corresponding relation were found in beats forced by stimulation from the ventricle. Now, in so far as the electrocardiogram of the mammalian heart is concerned, this corresponding relation is discovered. A forced beat, whose initial phases end with an upright deflection, has a downwardly directed end-phase ; a forced beat whose initial phases end with a downward deflection has an mijaght end-phase ; thus, it is shown clearly that the end-phase in these curves varies according toJlShe direction of the original spread (see Fig. 170, page 215). But an experiment by Samojloff {682) seems for the moment to be interpretable in the opposite way. Leading from the base and apex of the cold-blooded heart and recording natural beats or beats forced from various parts of the ventricular surface, he finds that while the initial phase varies in direction, the end phase is constant in direction. Mines (5/>5) interprets this observation as meaning that the excitation process constantly lasts longer in a particular region of the heart, independently of the original order of excitation. I am not convinced that this explanation is valid. Samojloff's experiment should be repeated with the use of indirect leads ; these would in all probability yield a different result, as they do in the mammalian heart. As Seemaim {695) has shown, the Samojloff phenom.enon is not constant even in direct leads. The interpretation of curves taken by means of direct leads from the heart constantly gives rise to difficulties ; these curves contain much that is mysterious. I am prepared to grant that Samojloif's experiment requires exjilana- tion, and that I am unable to give that explanation. His observation and the one which I shall cite will presently form a chief reason why a full acceptance of any present day hypothesis is impossible. In direct leads from the heart, in which the two contacts are very close together, an end phase always appears (51). It appears independently of the region of the lead, thus clearly shovsring that there is a disturbed balance even in the shortest strip of ventricular muscle towards the end of systole. The finding of this end-phase universally over the ventricle, however, does not forbid the interpretation of T in an axial and indirect lead as mainly a basal or apical effect according to its direction ; for the end-phase of the direct lead is not the same end-phase as T of the electrocardiogram, though the cause of the first may be comprised in the cause of the last. This appearance of T in all direct heart leads is not more mysterious than the appearance of a similar deflection as a final effect in all strips of muscle which pass through the contraction process. Neither is the variable direction of the end-phase, in different direct leads in which two contacts are placed on the ventricle, an essential difficulty. The chief difficulty is to explain why the end-phase varies in its direction in direct leads, when one contact is placed on different parts of the ventricle while the other is maintained on a constant point on the body wall. This is the case in the cold and warm-blooded heart and I have been unable to determine what influences the direction of the end-phase in these curves. The upright end-phase or downward end-phase, as the case may be, is not particularly associated with one region of the ventricle, neither does it show constancy of direction for one ventricular area from animal to animal. This phenomenon like that witnessed by Samojloff lacks explanation at the present time. VENTRICULAR DEFLECTIONS. 115 phase of the retreat would leave a relatively strong impression upon the curve,* for at this stage it would be less opposed by conflicting currents. Thus, on theoretical grounds, we might expect a relatively simple form of end-curve, but an end-curve whose direction was especially influenced by the last phase of the retreat and therefore especially related to the last phase of the stage of invasion. It should be understood that a quite constant and simple relation between the initial phases a,nd the terminal phases of the curve will appear only if the path of retreat and the rate of retreat is exactly or almost exactly the same as the path of invasion and the rate of invasion. To put the same matter in a different form, this constant and simple relation will appear only if the duration of the excitation process f in muscle substance is quite uniform in all parts of it. Whether there is ever quite such uniformity we do not know and at present there seems no means of determining, though a 'priori it seems improbable. We might assume that as the muscle structure is similar in different regions of the ventricle, the duration of the excitation process is the same in different regions ; but such an assumption would be at the best precarious. It is well within the bounds of probability that the duration varies in different regions under several influences, for exampies nutritional or nervous influences. A simple experiment probably illustrates this very point. If heat is applied to the apex of a heart which in the base-apex curve or electrocardiogram yields a downwardly directed end- deflection, then this deflection becomes upright without any material change in the initial phases^ (Bayliss and Starling, Mines, etc.) [27, 136, 555). This reversal of the end-deflection ( T) may be explained if we suppose that heat, by quickening the processes at the apex, speeds the retreat in this part of the heart, rendering the basal parts relatively electronegative to those of the apex at the last. The initial phases of the electrocardiogram are much less liable to change of form and direction than is y in a given instance. This rule conforms to the experience of all workers, and there are m-any conspicuous examples of it. One noteworthy example, namely, the effects of local change of temperature, has been discussed. Another almost equally prominent example is found in the effect of nerve stimulation. Stimulation of the vagus or sympathetic nerves has profound effects on the end-deflection T ; but the effects on the initial deflections are relatively slight {664, 682). Many poisons act in a similar fashion. As Cohn and his fellow-workers [66, 67) have shown, an early effect of digitahs intoxication in the human subject is an inversion of T ; * At all events upon the last part of T, which is the most constant in direction. t The time from which the invasion begins at a given point to the time at which the retreat leaves it free. t Heat applied at the base or cold applied at the apex, produces the opposite effect on T. The experiments of Eppinger and Rothberger (136), in which various parts of the ventricle were frozen, are to be classed with these cooling experiments and are not to be viewed as experiments in which large parts of the wall are thrown out of action. 116 CHAPTER VIII. this effect of digitalis appears to result from a direct action of the poison upon the muscle. To sum up these reflections, it may be said that the direction of T seems to be related to the manner in which the excitation process spreads in the ventricle, and there is suggestive evidence that the invasion and retreat follow more or less the same path when the heart beats naturally. If that is so, then the hypothesis, which supposes the excitation process (with its accompanying state of relative negativity) to continue throughout the whole of systole, sufficiently explains the chief experimental observations. But if this hypothesis is adopted, it would also be necessary to suppose that from time to time certain influences, such as changed nervous control or the action of poisons, may prolong or shorten the excitation process in one part of the ventricle to a greater extent than in another, and thereby promote dissimilarity between the path or rate of advance and retreat. In such fashion the curious relations of the deflection T under altered innervation or poisoning of the heart, might be explained. The elucidation of T might conceivably be approached from a slightly different standpoint. As we have seen, the initial phases of the electrocardiogram are comprised by the superimposed effects of right and left ventricle. If we are right in assuming that the duration of the excitatory process is. uniform in different parts of the heart, then T is also a dual effect and is to be regarded as the product of the end-deflections of right and left ventricle. Unhappily we have no certain means of ascertaining either the direction or value of the end-deflections of the true dextrocardiogram and levocardiogram. There is some reason to believe that the end-deflection of the human dextrocardiogram and levocardiogram* are opposite in direction to the last and chief initial phases in these curves ; thus, in lead III and usually in lead II the end-phase of the dextrocardiogram would be directed downward, and in the levocardiogram it would be directed upward ; whereas in lead / the directions would be the reverse. In that case the upright T of the normal electrocardiogram would be attributable to a preponderance of the left ventricular effect in lead I and of the right ventricular effects in lead II and III. But the evidence is no more than suggestive and cannot in this place profitably be pursued further. f Alterations in the value or direction of T, both of which are commonly seen in oases of heart disease, have not as yet received any adequate explanation. Note. — ^The following additional ]5ublications naay be consulted for further discussions and views on the constitution of the electrocardiogram {117, 118, 325, 392 and 573). * 1 am speaking of the directions of the real end- deflections and not of end-deflections which appear when the Left and right bundle branches are divided ; for these, as I have explained, are not reliable indications. t Experiments in which one or other ventricle is removed and the resultant curve observed {138), are perhaps too crude to possess material value. Chapter TX. ABERRANT CONTRACTIONS AND THE ELECTROCARDIOGRAMS OF HYPERTROPHY, ETC. In Chapter VI, the electrocardiograms portraying defects in the chief divisions of the dog's A-V bundle have been described, and it has been shown that the changed form of the curves results from altered distribution of the excitation wave in the ventricles. In the normal heart beat, the spread to each ventricle is through the bundle divisions and their arborisations ; the spread is simultaneous in the two ventricles in these circumstances. On the other hand, when one division of the bundle is cut across, the spread is at first confined to the contralateral or unaffected ventricle and, so far as this ventricle is concerned, it is normal ; later, the spread is to the affected ventricle and in this the distribution is wholly abnormal. Thus the spread of the excitation wave to the ventricles, considered together, is abnormal by reason of a defect in a chief distributing channel ; consequently I term the resultant contractions aberrant, for they are ,the product of impulses which have gone astray. In the last chapter certain human curves were used on the assumption that they express similar aberration in the human heart. The reasons for regarding these electrocardiograms as representing aberrant systoles may now be set forth. It will be shown in a later chapter that defective conduction through the A-V bundle is not an uncommon incident in cardiac disease; and histological studies have proved that the A-V conduction system, including the main bundle and its branches, is particularly susceptible to certain forms of degenerative and infiammatory change ; these injuries fall upon the conducting system selectively. The co-existence of discrete lesions of similar type in the main bundle and in one or both of its chief divisions has been observed on many occasions ; in other and numerous instances a single lesion has been found to involve the main stem and one of its two offshoots. Therefore, it is natural to anticipate that curves representing aberrant contractions will be obtained from the human subject. Now in any large collection of clinical electrocardiograms, a number of peculiar curves, having the distinctive forms of those used in the last chapter, are to be found and these same electrocardiograms show, as often as not, defective A-V conduction. The frequent association of A-V I 118 CHAPTER IX. block in its various degrees with the curves which we are now considering, may be taken as a first and important evidence of their origin. It is to be hoped that ultimately, a sufficient number of cases may be adduced in which these curves have been obtained and in which corresponding lesions have been seen microscopically ; but at the present time, owing to the difficulty of collecting such material, only a few instances can be quoted. In describing the curves now spoken of as levocardiograms, Eppinger and Stoerk (140) regarded a sclerosis of the septum, involving the right division of the bundle as the responsible lesion ; they had indeed diagnosed such a lesion during life. In a case examined by Cohn and myself (72), in which curves of the same kind were published, a lesion was found in the right division.* A second example has recently come within my own experience, the expected lesion being discovered by Dr. Butterfield who examined the heart for me. On the other hand, in several hearts examined for me (75) by Cohn,f no such lesions were found. Briefly, there are instances in which the expected lesions have been discovered, and these support our conclusion. There are also instances in which no lesion has been seen and these do not materially weigh against it J ; similar discrepancies between functional damage and anatomical discontinuity are frequent when A-V block has been discovered, discrepancies which will be described in the proper place. But the strongest reasons for regarding the human curves as representing defects of the bundle divisions is still the comparison of the human curves with those obtained experimentally. Curves held to represent defects of the right division {the human levocar- diogram). — The human levocardiogram (Fig. 65a) comprises three deflections in lead /. An inconstant deflection Q' is succeeded by a tall and broad summit R' (usually notched) and this is followed by a deep and rounded depression T' . In lead ///, the first deflection is a small summit R' and this is followed by a broad and extensive dip ;S", which is often notched, and by a large and rounded elevation T' . The curves of lead II are variable, and having smaller amplitudes are of less consequence ; usually they are similar in outline to those of lead ///, or consist of a number of small phases. The amplitudes of the chief deflections in leads I and /// exceed, and often greatly exceed, those of natural electrocardiograms. The initial phases * Associated with a lesion in the main stem and complete A-V block. f Who has reported upon them in our joint names. The electrocardiograms in these patients were in each instance characteristic of defective conduction in the right stem. The clinical notes (Hearts 42, 44, 73 and 91 of Cohn's report) are reported by Carter (46), as cases 5, 18, 20 and 19 respectively. The weight of the separated ventricles in the patients are given in a separate paper of my own (487), and Cohn's numbers correspond to those in the column " case nvimber " of my tables. My reason for referring to these weights is that they show that the curves were not the result of preponderating left hypertrophy. { They weigh against the conclusion less in that the division in its whole length was not examined. ELECTROCARDIOGRAMS OF HYPERTROPHY. 119 as a group present an exaggerated duration in all leads ; the Q, R, S group in the normal human electrocardiogram has an average duration of 0-08 sec, the initial phases of the human levogram average 0-14 sec. in duration, and always exceed 0-1 sec. (453, 463, 475). It is also noteworthy that the chief deflection, R' or ;S" as the case may be, is always opposite in direction to the final wave, T' . =-p-- jr::^^.^ ^^^ — •f-d^r^:. _j_4_.;:_„ \. T.~' T=; ^l"^^ t#t- _^^^ L ' ^~~1H §-Wi - J^ S c— A ^Q.-r~S--r 1 — ti rJL-^- 1 P^ ' -y-l ■f H ^ ■ N - r 1 ^ita P< -z:'-—\: ^rp— - —-:v ^^_^_ ■^71 :;:;: JT-Trrr: IJTT - — - ;^: '- ■--.---\l -^:m_ ^ = -~ - ■ ~ IIZZZ =t7 — - ^^^-■-—~ ...:--: i-i= _t4 -^^^ ^ — - ; 4- % i ^ —-3 riz?7=^ «^ - r — -^ ^\^- ■n^fa — ^MHl^ t '^^l m --- : ~ '7 ;— P--_---^--_-_J _ — f ^ — :-.-- Fig. 6Sa. Fig. 656. Fig. 65a. {Phil. Trans., 1916, B., CCVII, 2S4, Part IV, Fig. 11.) Electrocardiograms from the three leads in a patient. Curves of this type are associated in the human subject with defects of the right division of the A - V bundle. Time in fifths and twenty- fifths of a second. Fig. 656. {Phil. Trans., 1916, B., CGVII, 2S4, Part IV, Fig. 12.) Human electrocardiograms from the three leads. From a case of aortic disease with great hypertrophy of the left ventricle. These curves are almost identical with those of Fig. 65a in their initial phases ; they differ from them in their final phases {T' and T). Time in fifths of a second. These two series of curves have been specially selected for their resemblances. Curves held to represent defects of the left division {the human dextrocar- diogram). — The first curves of this kind to be described {463), are shown in Fig. 666.* In lead /, a diminutive deflection R' is succeeded by a deep and * The rarity of human dextrocardiograms and the comparative commonness of the human levocardiograms is to be explained chiefly by an anatomical difference. The right division has a long course as an unbranching strand, the left quickly subdivides and after the first part of its course the complete division of this portion of the conducting tract would necessitate a very extensive lesion. I 2 120 CHAPTER IX broad deflection 8', and an upwardly directed T'. In lead III a diminutive Q' and a tall, broad and notched R' are followed by a depression T'. The curves in leads I and III are similar to leads III and I, respectively, in levocardiograms ; they are in the main diphasic, T' having an opposite direction to the chief initial deflection. The initial deflections are of long duration as compared to the normal. " T-J —— ^ 1 ' - :::p. N F i ^s^ Pitt^^^^B ■ 1 ^ „ — _^ - - - Fis. 06a. Fig. 666. Fig. 66o. {Phil. Trans., 1916, B.,CCVII, 284, Part IV. Fig. 13.) Three electrocardiograms from a case of mitral stenosis; illustrating preponderating hypertrophy of the right ventricle. Time in fifths of a second. Fig. 66.i. {Phil. Trans., 1916, B., CCVII, 384, Part IV, Fi;]. 14.) Electrocardiograms from the three leads in a human subject. These anomalous curves were probably produced by a defect of the left division of the bundle. Time in tliirtieths of a second. Comparison ivith experimental curves. — Comparing the general features of the human curves and those obtained experimentally from the dog, we find the following features in common, whether we are dealing with levocardiogram or dextrocardiogram, and irrespective of lead ; — 1. The amplitude of the chief deflections is more than normal. 2. The initial phases have an unusual duration. ELECTROCARDIOGRAMS OF HYPERTROPHY. 121 3. The final deflection is always opposite in direction to the chief initial deflection ; each curve is, as a whole, broadly diphasic* If the human and canine levocardiograms are compared in detail, very close resemblances are noticed in form (see Fig. 44, page 77, and 65a, lead III). Complete correspondence between the number and directions of deflections is frequently to be found, and even the notching of the chief initial deflection is usual in the human curves. "j" The dextrocardiogram of the dog and of man may present notable differences in their detail and this is to be expected, seeing that there is a difference in the distribution of the Purkinje strands to the right ventricle. Levocardior/ram Bicardiogram Dexlrocardiof/ram. ( CaJc ) T ^ _, i^s^.^ 2 V T^ A M-y ^ ' ^i^^^"*^^' I s -^ •rrr t- t- 1^-^ \ _(V H^ 7- o I I 3 ^ S 6 T r- -^ S 1 ^ -> -> A- / / / / / / m / (1 y '^ ~. / / / 2f / \ N / , A 3 'f ■i I / X 3 ^ S i Fig. 67. (Phil. Trans., 1916, B., CCVII, 287, Fig. 3.) A chart. showing three sets of curves obtained from a large Rhesus monkey. The series of initial deflections shown to the left is that obtained after clamping the right division of the bundle. The central series comprises the natural curves obtained from the undamaged heart. The right-hand series shows the calculated dextrocardiograms. Abscissae and ordinates as in Fig. 61. For this reason it seemed expedient to examine monkeys. Between the monkey's dextrocardiogram (Fig. 67) and the supposed human dextrocardio- gram there are no conspicuous differences, and a similar statement is applicable in the case of the levocardiogram. *This feature is usually common to all three leads, but exceptions are found in lead // when the initial deflections are multiple or small. t In some series of curves taken from the dog, the curves of lead I have different outlines, a fact which has led to misapprehension (671). 122 CHAPTER IX. Considering the evidence as a whole, there can be no question that the curves described, and supposed to represent human levocardiogram and dextrocardiogram, are respectively due to defective conduction in the right and in the left divisions of the bundle. Curves associated with preponderance of one or other ventricle. It has been stated by Einthoven {111, 114), that certain variations in the form of ventricular curves, as seen in the several leads, result from hypertrophy of the left or right ventricle. Einthoven appears to have based his view on the types of curve obtained in patients suffering from aortic disease and mitral stenosis. He described an exaggerated R in lead /, and an exaggerated S in lead ///, as the effects of left hypertrophy (Fig. 68&) and the converse picture, namely an exaggeration of S in lead I and of R in lead 77/ as the effects of right hypertrophy (Fig. 68a). :-fp:- -X ^EE=M^ ^3^ 11 I MJ 1 1 MM I M IJ 1^. ■ I ■ ■ ■ ■ I ■ 1 ■ ■ I 1 1 . ■ 1 ■ ■ ■ 1 ■ ■ f ^E=F^:q^=^T^- HMMMMMMMI MMllllI IllM I Mil IM IM M Ml MIMIMI M 111! Tn^TT-riiiriiMiniMMiniMmTmiiMiurT Fig. 68a. MMMMMMMiMMMi Fig. 686. Fig. 68a. Curves from the three leads in a patient in whom there was preponderating hypertrophy of the right ventricle. The chief deflection inlead / is .s'. in lead III it is R. Time-marker in thirtieths of a second. Fig. 686. Curves from the three leads in a patient in whom there was preponderating hypertrophy of the left ventricle. The chief deflection in lead I is H and in lead III it is S. Time in thirtieths of a second. From time to time {399, 467) further evidence in support of Einthoven's view has accumulated, and it is now sufficient to estabUsh his signs as the most rehable which we possess of preponderance of one or other ventricle. First, if a series of cases of mitral stenosis (or pulmonary stenosis) is examined electrocardiographically, the average curves obtained exhibit Einthoven's signs of right ventricular preponderance. The weights of the separated ventricles in a similar series shows in the average preponderance ELECT ROGARDI 00RAM8 OF HYPERTROPHY. 123 of the right ventricle {467). Secondly, if a series of cases of aortic disease is examined, Einthoven's signs of left preponderance are found in the average curves, and the weights of the separated ventricles in a similar series show preponderance of the left ventricle.* Thirdly, the heart of the child possesses a relatively heavy right ventricle from the time of birth up till three months after birth. The signs of right preponderance are always present in the child at birth (Fig. 69) and disappear about the third month of extra-uterine life {399, 467). Finally, in instances in which electrocardiograms have been obtained, and in which the weights of the separated ventricles have been taken, the correspondence between the electrical signs and the ratio of weights of right and left ventricles, has been remarkably exact {467). \—rr - - »-i«tp*g^ 1 m 7- 1 p mm^frn mmm^fKmmmm^fmmtm ..a.... C?5 Fig. 69. Curves from a child two hours after birth. The relative heights and depths of the peaks is such as is expected where there is relative preponderance of the right ventricular muscle. Einthoven's conclusions are confirmed by these observations, and there is httle reason to doubt that they are vahd.j The complete series of electrocardiographic signs, corresponding to preponderance of one or * In some instances of rnitral stenosis or aortic disease, preponderance, as estimated by weighing, is not discovered in the expected ventricle. Neither are Einthoven's signs of right hypertrophy discovered in all instances of mitral stenosis, nor those of left preponderance in all cases of aortic disease. t A recent criticism [42] of my conclusions, in regard to the value of electrocardiograms in recognising preponderance of one or other ventricle, appears to be based on the assumption that if the electrocardiographic and clinical evidence seem incompatible, the electrocardiographic evidence must be at fault. Further misunderstanding appears to have occurred in that the distinction between the terms "hypertrophy" and " preponderance " is not fully grasped. I have endeavoured to make this distinction unmistakably clear. The signs described are not signs of enlargement of the heart, but of a disturbed balance between the mass of muscle on right and left side. The left ventricle may be hypertrophied and the signs of right preponder- ance appear if the right ventricle is hypertrophied in greater proportion- Moreover, the signs of one or other preponderance may be seen when the heart as a whole is not enlarged. 124 CH AFTER IX . other ventricle, as they are now known to us {475) may be tabulated. Right preponderance. Left preponderance. Lead /. Lead III. Lead /. Lead ///. Q R S Q R S Q R S Q .R S + ++— ++— + Not only are there changes in the amplitude of R and S, but also in the amplitudes of Q ; and these changes in right and left preponderance are qualitatively identical with the changes discovered when respectively the left and right divisions of the bundle are defective. Briefly, there is a close resemblance between the initial phases of the electrocardiogram when there is left preponderance and when the right division is defective ; there is a similar resemblance between these phases when there is right preponderance and when the left division is defective. This observation has suggested the explanation of the curves associated with hypertrophy (475). To exemplify ; if one ventricle (the left) preponderates, then in the dual curves taken from both ventricles, the levocardiogram will preponderate ; and inasmuch as the left ventricle is the more massive, in so much will the levocardiogram impress itself upon the combined curve ; in extreme instances of left preponderance, the electrocardiograms in their initial phases resemble levocardiograms in the closest detail (see Pig. 65a and b). Such curves, when analysed in respect of the rotation of the electrical axis, show a similar rotation to that observed in the case of the levocardiogram. The Initial phases of the levocardiogram, as opposed to those of the physiological curves, exceed 0-01 seconds in duration as has been stated ; this increase is due to the defect in the right division of the bundle and to the consequent delay of spread in the right ventricle. A similar, though less pronounced, prolongation is seen in the curves of left hypertrophy, though here it is attributed to delay in the spread of the excitation wave consequent upon the thickness of the muscular wall of the left ventricle. Usually, curves of hypertrophy may at once be distinguished from those of defects of the right division of the bundle, which they resemble, by comparing the initial phases of the curves. The characters of the levocardiogram are not fully displayed, and the increase in the duration of the initial phases is inconsiderable (compare Fig. 65a and 68&, and also Fig. 70 and 71). I have fortunately obtained (465) two series of curves showing the differences in a single patient (Fig. 70 and 71). They are striking and characteristic. Although the chief initial deflections in leads I and 7/ 7 are similar in direction in both series, the amplitudes of these deflections and their durations are much greater in Fig. 70. Moreover the direction of the chief deflection and the end deflection are opposite in leads I and 777 of Fig. 70 ; this is not the case in Fig. 71. The last distinction is one of considerable practical importance ; for where (as in Fig. 65&) the initial characters of the levocardiogram are fully developed in curves which are, in reality, the result of left muscular preponderance, these curves are to be distinguished from curves expressing aberrant heart-beats by the amplitude and direction of the final deflection and by these criteria only. ELEGTROC AUDI 0GRAM8 OF HYPERTROPHY. 125 ■ . . . ^t ....-■ . ... .-^^^■,.,^^-^-.. Fig. 70. Fig. 7]. Fig. 70- Curves taken from the three leads in a case of aortic disease during a febrile atta(3k;. They show defective conduction along the right division of the auriculo-ventricular bundle. Fig. 71. Curves from the same patient, taken a day later and during the subsidence of the fever. The ventricular portions of the curves have changed profoundly ; there is now no evidence of bundle defect, but of relative preponderance of the left ventricle, Minor forms of aberration. Since it is known that the normal ventricular electrocardiogram is controlled in its form by the course which the excitation wave takes, and that the course of the wave is in turn governed by the arrangement of the A-V conducting system ; and since it is known that a lesion completely transecting one whole bundle division profoundly modifies the form of curve, it is to be supposed on a priori grounds that lesser interferences with the channels which conduct the excitation wave in its descent to the ventricles will also leave clear impressions upon the resulting electrocardiogram. This conjecture is completely upheld by experiment. I find in actual experiment that a lesion which interrupts any of the larger end branches of a bundle division will influence in greater or lesser degree the corresponding curve, according as it deflects the excitation wave in greater or lesser degree from its accustomed path. A lesion which cuts the right division of the bundle as it reaches the papillary muscles produces a conspicuously modifled curve, though such a curve does not depart from the normal so greatly as does the curve resulting from a break in the division high up on the septum ;* section of the chief end branches of the right division, where these are clear of the papillary muscle and bridge the cavity of the ventricle, produce lesser changes ; lesions involving the right network as it lies beneath the endocardium do not greatly change the curves. Similar conclusions apply to the arborisation in the left ventricle. * Hence it is to be recognised that there are early outgoing branches from this bundle division. 126 CHAPTER IX. It is certain that lesions in the lower reaches of the arborisations occur in the human subject ; there is every ground for the belief that such pathological damage modifies the human electrocardiogram more or less in the corresponding patients. To identify lesions in particular regions of the arborisation with corresponding forms of electrocardiogram is a task which lies in the future. But there is another way in which the electrocardiogram may suffer change in form, namely, by variations in the relative rates of conduction along the two divisions of the bundle and their branches. When the left division is so nipped in the jaws of a clamp that conduction through this tract is abolished, the curve in its initial phases is that of the right ventricle only (the dextrocardiogram). If the clamp is now relaxed, the power of the division to conduct will frequently return. But the recovery takes time, and during this stage of recovery a form of electrocardiogram is seen which is often transitional between the normal curve and the dextrocardiogram. Clearly, since the actual form of the initial deflections is due to a summation of dextrocardiogram and levocardiogram, the form of these deflections and their amplitudes will depend largely upon the times of events in right and left ventricle in relation to each other. A delay in the appearance of levocardiogram or dextro- cardiogram and consequent -distortion of the bicardiogram is strongly suspected to occur in certain patients. Some years ago I advanced this hypothesis of delay in the recovery of conduction in a single bundle division to explain a striking and remarkable series of transitions between a complex having the full features of a levocardiogram and a normal complex {451). The changes occurred over a series of eight ventricular cycles, each ventricular complex being intermediate in type between those adjacent to it. This curve (Fig. 29 of my paper) was taken from an asphyxiated cat in whose auriculo-ventricular system conduction changes were known to be occurring. Clinical examples of an almost parallel kind have been published by Cohn (61) and are probably to be explained similarly. Many other clinical examples of temporary disturbance of a bundle division have since been recorded (182, 543, 581, 783). Without exception, so far as I am aware, curves showing the gradually changing complexes (from dextrocardiogram to levocardiogram, from levocardiogram to dextrocardiogram, or from one or other to a curve of the normal type), such as are here described, have been obtained from patients in whom there were unequivocal signs of deficient conduction in the main bundle, or strong evidence of its presence.* "^ in Frederioa and MoUer's case {liS:i) of auricular Hbrillation, 1 judge complete heart-block to have been present, although the authors seem not to have been of this opinion. Their curves show the gradual development or gradual disappearance of the dextrocardiogram. At autopsy a lesion of the septum, involving the left bundle division was discovered. The curves are not those of right preponderance as the authors would suppose, for the initial phases are prolonged and the direction of T is opposite to that of the chief deflection in all leads of the curve. Note. — Another mechanism by which a gradual transition of curves from one type to another is produced, is described at page 217. ELECTROO AUDIOGRAMS OP HYPERTROPHY . 127 Displacement of the heart's anatomical axis. In Chapter VIII methods of calculating the heart's electrical axis have been considered, and it has been seen that this axis varies very greatly during the progress of the cardiac cycle. It is to be emphasised again that none of these electrical axes can be assumed to represent the anatomical axis. But as it is certain that the electrical axis bears to the anatomical axis a certain relation, inconstant though it may be, it is true that the former may be employed within certain limits in calculating the actual lie of the heart in the body. There is no better example to illustrate this fact than the transposed heart ; for, as in most normally placed hearts the direction of the electrical axis corresponding to the phase when R is inscribed is inclined to the horizontal at approximately 75°, so in transposed hearts it is usually inclined at approximately 105°, being as much inclined (by 15°) to the right of the vertical in the former, as to the left of the vertical in the latter. But inasmuch as there are other factors which influence the direction of the axis when R is inscribed, such as a relatively greater development of one ventricle or the other and individual differences in the distribution of the excitation wave within the ventricles, in so much is this method of ascertaining the relative position of the anatomical axis in different individuals uncertain.* Nevertheless, in calculating changes in the direction of the anatomical axis in one and the same individual, changes such as may be induced by posture, breathing or other act which displaces the heart, the method is not unserviceable. Conversely, a chief merit of the method is that it explains many of those changes in the form of the electrocardiogram which are seen to result when the lie of the heart becomes altered. To simplify my description of these changes of form I have con- structed the accompanying diagram (Pig. 72). If we suppose a manifest potential difference of fixed value to be developed in the chest and imagine further that it rotates through a full circle in the plane of the leads, then this potential difference, as it rotates from 0°to 180° and again through to 0° will be represented in the three leads in the fashion shown in the diagram. Thus, in lead / when the axis lies at 0° (see left hand edge of eMn Fig. 72), and therefore in the line of the horizontal lead, the difference in potential will be fully represented and will be shown as a full up- ward deflection. As the axis rotates from 0° to 90° the difference of potential shown by lead I will gradually decline until it reaches zero, for at 90° the axis is at right angles to the lead. On further rotation in the same clockwise direction, the potential difference will again gradually increase, though the direction of the deflection will now be reversed (downwards) until at 180° it will again reach its maximum, for again the axis and the lead will be in the same hue. During * That there is a relation between the type of electrocardiogram and the lie of the heart, estimatedorthodiagraphically orskiagraphically, is stated {233, 747), and no doubt this relation exists ; but the lie of the heart will not account for many conspicuous differences in individual curves taken from healthy people. 128 CHAPTER IX. the further rotation from 180° to — 90°, there will be a second gradual decline to zero, and during the last part of the rotation, from — 90° to 0° there will be a gradual increase until the original upward maximum is reached. Similar changes will occur in leads // and ///, though at different phases of the rotation, as shown. This diagram is so constructed that the relative values of e^, e^ and e* are accurately shown for all positions of the electrical axis in its revolution. Thus, when the axis is 60° the relative values of e^, e^ and e^ are 5, 10 and 5 respectively, where 10 is the value of the manifest potential M (see Fig. 58a, page 101). Now in, the most usual form of normal electrocardiogram the angle which the axis, while B is inscribed, makes with the horizontal averages approximately 75°. If the heart is tilted so that its anatomical axis moves in an anti-clockwise direction through some 30°, the electrical axis will follow through 30° and will move from 75° to 45°. During this movement (see Fig. 72) B in lead I may be expected to increase considerably from its original small value, in lead // it will at first increase a little and later decline a little, while in lead /// it will steadily and conspicuously dechne. The chief change, be it observed, will occur in leads I and ///, and the changes will be in opposite directions. These are precisely the changes which are seen to occur in the height of B in these leads, when the apex of the heart is tilted upwards and to the left, as when gas is introduced into the stomach {323), or when the diaphragm rises in expiration. In the same type of electrocardiogram the angle of the axis while T is inscribed, averages approximately 55°. If the heart is tilted through 30°, this angle will be decreased from 55° to 25°. In lead /, T should increase conspicuously in size, in lead II it should show some decline in size, while in lead III it should diminish to zero and even become slightly negative. These also are precisely the usual changes seen in T in fuU expiration or when the stomach is distended. The axis while S of the usual electrocardiogram is inscribed is very variable, and the changes which have been observed in its magnitude during the phases of respiration are also variable. The data which we possess are insufficient to formulate any precise rules. Writers have for the most part not taken into account the original angle while S is inscribed, and to this the lack of constancy in the changes of 8 is very probably in large part attributable.* In a less usual type of electrocardiogram B is originally as high or almost as high in lead / as in lead II, while it is inconspicuous in lead III. The axis corresponding to it lies in the neighbourhood of 35° (see Fig. 72). In such cases, anti-clockwise rotation of the heart leads to a sUght increase of B in lead /, a steep decline in lead //, and inversion in lead ///. These are again precisely the changes to be anticipated. * Lack of constancy, so it is thought, may also result because rotation of the heart on its vertical axis is thought to be particularly prone to affect S. ELECTROCARDIOGRAMS OF HYPERTROPHY . 129 o l^HHI ^Hi ^H^ 5? ^^^^^H ^S ^^^^ '^ ^^^^^H ^ ^^^^^ 1 ^^^^^1 W ^^^^^^ ^^^^1 ^k I^H^^A ^^^^H ^^^1 ^^ ■^ ^ <^ A & rt 1 1 %■ E-. ^ rt 1- «J N 5 -p 03 '♦H »T^ X c3 •- ' 1 ^ -P .5 g g ° I g.S g o .S ~ :S _§ ^ .2 (D t^ m -r; CD ' 5 H 0) X! " i -p ^ o -S d o ts -^ S £ V, * O w ft '^ 5^" -J ."S °3 .13 p a o © . £3 5f b6 0) © Tl r1 +j f ce F-4 > 5 d ^ +3 +3 -p & CO CD M o •a (S 3 d 3 -C ;3 p -p :^ .^ © o T3 c c8 >>^ S o II c -gi. 3)^ a c3 » — M W t M " 'tf tW " VH tf'W V V WW I*** V ** V V ' V H • i « V V V V %« iH « I ir~ >»»»l'MliH> ■»»■■» . W > » W V » Fig. 76. (X |.) Two radial curves from a single patient. The upper one is accompanied by a venous curve (the two venous curves were identical in every respect, the lower one is therefore omitted). To illustrate the presence of a dominant rhythm ; most of the prominent radial beats belong to it. The lengths of the intervals between prominent beats fall into a few categories. A period of irregularity of moderate length is duplicated. The irregularity resulted from premature contractions arising in the ventricle (the actual events were determined electrocardiographically). rhythm may be assumed whenever an irregular pulse shows periodicity of any kind. These facts are of special importance in excluding irregularities of the ventricle which are due to fibrillation of the auricles (see Chapter XXIII.). ARTERIAL PULSE CURVES. Signs of an undisturbed dominant rhythm. 136 After ascertaining that a dominant rhythm is present in a given case It IS often possible to carry the analysis a step further, and, as we have seen' to show that the dominant rhythm is unaffected by the ventricular irregularity (Fig. 74, top curves). If a pulse in which the beats are paired or grouped occasionally becomes regular and there is a simple mathematical relation between the intervals separating groups and the intervals separating regular beats (Fig. 77, top curves.), the same dominant rhythm controls each portion of the curve, and this rhythm is unaffected by the irregularity. An example in which the controlhng rhythm is affected, is shown in the bottom curve of the same %ure. The three curves of this iigure are more complex instances of what we have seen in Fig. 74. 9vvilvv9¥¥¥¥¥vmtrM¥aw9^v » » » » »'W ¥ ¥ ¥ ¥ ¥ ¥ ' Fig. 77. Three arterial curves from patients showing coupled pulse beats and transitions to regular heart action. In the upper two curves there is a simple m.athematical relation between the lengths of the paired beats and the lengths of the regular beats. The same dominant rhythm controls the arterial curve, whether its beats are in pairs or are regular, and this rhythm is undisturbed. In each of these curves the intervals a and 6 are equal. In the bottom curve this relation between the paired and unpaired beats is not found ; the presence of a dominant rhythm is known because the two groups of paired beats are equal to each other in length, and because the three succeeding beats are equal to each other in length ; but the measured intervals a and 6 and c and d show that this dominant rhythm has been disturbed. The actual events in these and succeeding tracings were determined polygraphically and electrocardiographically. Further, when the pulse beats occur in uneven groups and simple mathematical relations may be established between the lengths of all the groups, the same conclusion holds good, even if the individual groups have not the same detailed construction (Fig. 81). The stability of a controlling rhythm may be demonstrated in irregular curves of very great complexity ; examples are shown in Fig. 78-80. K 2 136 CHAPTER X. The same principle is applied to curves in which there are abrupt changes of rate, the heart beating regularly before and after the change ; for if there is a simple relation between the slow and fast periods, the fundamental •rhythm of the heart has remained unaltered over the change ; while, if no such relation can be established, the faster rhythm is a new development and interrupts and supersedes the fundamental rhythm (Fig. 82). Fig. 78. (X |,) An arterial curve from a case of flutter, showing an undisturbed dominant rhythm. The stretches of curve a and b are equal ; and these two stretches are dovetailed by three equal groups (c, d and e). Fig. 79. ( X |.) An arterial curve from a case of flutter showing two equal stretches of curve one of which (a) includes two groups of paired beats and » group of three beats, the other consisting of five regular beats. The periods c and d are also equal. The dominant rhythm is undisturbed. XJXJVJ Fig. 80. (X f.) An arterial curve from a case of flutter. The stretches a and 6 are equal ; the stretches c, d and e are also equal. The whole curve is covered by measured stretches which adjoin or overlap. The whole curve shows evidence of the same and undisturbed dominant rhythm. It should be noticed especially that measurements of this kind are only justified when the beats, lying at the extremities of any given bracket, have equal pauses preceding them. In Fig. 78, 79 and 80 the irregularities resulted from heart-block. When the dominant rhythm is undisturbed, a pulse irregularity is due to heart-block or premature ventricular contractions. A regular sequence of heats originates from a single source. When a pulse is regular for long periods it may be taken that the rhythm is promoted by impulses arising in a small and limited area of the musculature, and probably from a single point. And this statement holds true whether the rhythm originates in the normal pace-maker or not, ARTERIA L PULSE CURVES. 137 ^ a += A S^ latio dom hapt t-^o a C (s a. 3) m 1 °^ §^1 H 4i c3 •2-S ^ -H = ^ ,0 n -1^ 2 § C3 n O ^ -d cS t rt a « 3 ^ ^ (S of h 5 not tim, OJ .s -t> M , >v iS g ^ CS ^ !- rup the logi ^ . o e ^11 t-^ "^ tC fH p .a c3 o 1 s ^ ^ M O ?s a o ^ cS Q c "3 .3 ^H O © 5 r, ^3 ^ fl s '§ > ^ O « ai «4H X 'rti C/J > C! ^ o tS tf B tH M •^ ;? 138 CHAPTERX. Observation permits the further statement that, where a number of pulse beats are found, for example six or more, and they are equal in amplitude and regular in their succession, the impulses from which they originate are generated in a single focus, whether the rhythm of which they are an expression is dominant or interrupting. An example of an interrupting rhythm of 10 beats arising from a single point in the ventricle is shown in Fig. 81. The method of examining an arterial curve now described I call the " spacing " of curves. Phasic variation of the dominant rhythm. The reverse statement, that a rhythm generated at a single focus is regular, is not necessarily true. For if such a rhythm is built up at the pace-maker it frequently shows periodic variations of rate. It is often possible to identify these variations in the arteriogram. They consist of a gradual waxing and waning of pulse rate, which is repeated and is usually synchronous with the acts of breathing. Phasic variations of rhjrthm which are independent of respiration also occur (Fig. 83), but cannot be identified certainly without the aid of records from other heart chambers. (Variations of this character are more especially considered in Chapter XXXII.) The rate of the dominant rhythm as an index of its source: The rate of the dominant rhythm, usually to be ascertained in the arteriogram, offers a clue to the source of such a rhythm, though it is not decisive in this respect. The dominant rhythm may have its origin in the normal pace-maker, and under these circumstances the rate usually approaches 72 to the minute ; but the limits of variation are great and pass from 30 to 240 to the minute. The dominant rhythm may have its origin in the ventricle, and under these circumstances its rate approaches 30 to the minute (the known limits in man are from less than 1 to 90 to the minute). Not infrequently a new rhythm of pathological type may prove dominant, and such rhythms may arise at many points in the musculature. The rate of these new rhythms is variable and the limits approach 110 and 343 per minute. Thus if the dominant rhythm has a rate approaching 70, there is presumptive evidence that it is generated in the pace-maker. If the rate is approximately 30, the rhythm is usually of ventricular origin. Lastly, if the rate lies constantly and continuously between 130 and 340 new or pathological impulse formation should be suspected. The length of the cycle following a premature heat. This measurement is of value in indicating the source of the premature contraction and will be referred to in more detail at a later stage. At present the following categorical statements may be made in regard to such cycles : — ARTERIAL PULSE CURVES. 139 1. If the cycle is of a length equal to that of a cycle of the dominant rhythm the source of the premature beat and the source of the dominant rhythm are the same. 2. Further, it may be said that the greater the distance between the points at which dominant rhythm and premature contraction arise, the longer will be the cycle following the premature beat. 3. And finally it may be stated that if the cycle following the premature beat is of full length, so that it compensates for the curtailment of the previous cycle re-establishing the spacing of the dominant rhythm, in all probability the premature contraction has arisen in the ventricle* (Fig. 74). Complete irregularity of the pulse. When a pulse is completely irregular, that is to say when the intervals between beats are of very varying lengths, when there is no regular lengthening from beat to beat or subsequent regular shortening from beat to beat, when phases of irregularity are not repeated periodically, when the lengths of beats do not fall into a few simple categories, a dominant rhythm is absent, and the heart's mechanism is one known as fibrillation of the auricle (Fig. 84). The value of the arteriogram to the student of disordered mechanism can hardly be exaggerated. The early observations of Cushny (86), and especially of Wenckebach {754-758), in which arterial curves were critically studied, laid the basis of our present knowledge. Wenckebach's insight led him to differentiate a large number of the more important disturbances in these curves alone. My electrocardiographic studies have enabled me to extend his method (447, 455, 457). "V— y-y- y v ' y y-Tr-vvv V * r-r-ifvw 1 •»-» ■»■ > ■ If ti JS 10 id 2.3 3.0 ^.s |.J 3 ,) i-K i-i ]_.i I'l 11 111-^' 3J 21- I'l 2-« iO .-7 l-l ii 3'1. Fig. 84. (X |.) Complete irregularity of the pulse. From a patient with fibrillation of the auricle. The intervals from pulse beat to pulse beat show the wide variation characteristic of the condition. The lengths of the beats are marked in fifths of a second below the curve. * Premature contractions arising in the auricle and junctional tissues occasionally give rise to this picture, but these instances are exceptional. Chapter XI. THE ANALYSIS OF DISORDERED MECHANISM. VENOUS CURVES. The arterial pulse signals the ventricular systole ; the venous pulse signals the auricular systole ; these are our chief objectives in using the arteriogram and the phlebogram, respectively. The simultaneous or polygraphic record, indicates the time relations of these systoles to each other, and displays the sequence or lack of sequence in chamber contraction. The polygraph was the first instrument employed for this purpose ; it is to Mackenzie [499, 500), that we chiefly owe a method which still maintains its lead as a routine device for clinical purposes. Identifying the " c " wave. Generally speaking the radial curve is taken as the standard from which all measurements are made, for the artery at the wrist presents natural advantages, rendering it the most serviceable point of arterial pulsation. The radial upstroke is used in calculating the onset of ventricular systole, and it is first corrected from a simultaneous carotid curve in order to allow for the difference in transmission time from heart to radial and heart to carotid, respectively. In this way a point is obtained upon the radial curve a little (about 0-1 sec.) before the upstroke, which represents the time at which the upstroke of the carotid occurs in the neck. In simultaneous venous and radial curves this point is transferred to the venous curve, and in the latter it represents the onset of ventricular systole. (These measurements may be made in Fig. 85.) Briefly, the measurements identify the upstroke of the c wave. The two measurements may be accomplished by a single transference. A short strip of venous and radial curves is taken ; the clock is stopped, and index marks are written ; short strips of carotid and radial curves are then obtained. The interval {r r) between two radial upstrokes, one lying to the left (Fig. 85) the other to the right of the index marks is measured. A corresponding carotid upstroke to the right of the stops is determined, and also a point upon the venous curve, separated from the carotid rise by the distance {r r) ascertained in the previous measurement. The transferred measurement is justified when the clock is running at the VENOUS CURVES. 141 same rate both to the right and left of the index marks, and when the curves to right and left are level, and only under these conditions. The question of level has always to be considered in using levers which describe an arc* Where accurately measured intervals are desired, correction from the actual index marks is essential, and the points obtained should be transferred to the time marks. Where the waves of the jugular curve are simple in form and clearly inscribed, a single measurement is taken from the index mark to the upstroke of a radial beat to the left of it, and is transferred from the upper index mark to a point on the venous curve ; the upstroke of the c wave will .-JV4N..jv.^jv.j\4NNMVsNA.f^^ Fig. 85. (x -|.) Polygraphio curve, illustrating a method of identifying the c waves. Simultaneous radial and jugular curves are shown to the left, and simultaneous carotid and radial curves to the right of tfie central index marks. The transferred measurement is from ?• to r in the radial curve and from c to c in the upper tracings. lie at approximately 0-1 sec. to the left of the point thus ascertained. The wave c is used as an index of the onset of systole ; the actual instant of onset may be obtained from a simultaneous cardiogram, mechanical or electric, or may be gauged approximately in a venous curve by allowing for the transmission intervals from heart to aorta (presphygmic interval) plus aorta to carotid ; a total allowance of about 0-1 sec. to the left of the carotid upstroke. Identifying the "a" wave. In a venous curve, where a group of clearly inscribed waves accompanies each cardiac cycle and where one of these waves lies directly before (to the left of) the upstroke of c, this wave is known to be a, and its upstroke represents the onset of the auricular systole in the venous curve. Its onset can be ascertained with certainty when the wave is prominent and clean cut from start to finish. Where the wave shows division, or where the upstroke lies at more than 0-2 sec. from the onset of c, the curve should be lettered with greater caution. Constancy in the shape and position of the wave from cycle to cycle or from one phase to a similar phase of the same curve is most * No such correction is needed when the curves are taken photographically, as in Fig. 28, page 51. 142 CHAPTERXI. desirable ; it is a golden rule only to accept those curves of which the interpretation is transparent, or of which repetitions in detail are secured. Where the limits of a are dubious, they should be checked from a waves at other points in the neck, from similar waves upon the apex curve, or lastly from evidences of auricular systole in electrocardiograms. The " As-Vs" interval. The interval a-c, measured between the commencements of the corre- sponding waves and taken as an index of the As- Vs interval (the true interval separating the commencements of auricular and ventricular systole), is customarily 0-1 to 0-2 sec. It is usually longer than the corresponding As-Vs interval as shown by electrocardiograms, probably because c is more delayed than a. But this does not appreciably detract from the value of the a-c interval as an index. The a-c and P-R intervals never differ from each other by more than 0-1 sec. (generally the difference is less), and consequently a prolongation of the a-c interval to 0-3 sec. or more is a reliable guide to similar lengthening of the As-Vs interval. Even greater reliance may be placed upon a notable change in a-c interval from beat to beat in one and the same case. The value of the summit "v." The upstroke of c is determined and checked from the arteriogram. The onset of a cannot be checked in a similar manner, therefore the measurement of a-c intervals should be circumspect. To determine the limits of ventricular systole is important in cases where v may be mistaken for a, or in instances in which a and v are suspected to coincide (in cases where the heart beat is rapid and v is insignificant, or in cases where the a-c interval is prolonged). The end of ventricular systole is represented in sensitive carotid and radial curves by the dicrotic dip, and in the jugular curve by the apex of v. The summit of the latter is synchronous with a point lying a little (usually 0-1 sec.) after the bottom of the dicrotic depression in the carotid. The apex of V in the jugular is synchronous with the depression of the dicrotic in the radial curve. The summit of v may be also checked by measuring the length of the systole in a single carotid beat and transferring it to the jugular curve. The measurement is of practical value because the summit has so constant a relation to the end of systole ; a relation which is well shown in Fig. 28, page 51. An example of its application may be studied in Figs. 91-94.* * Theoretically the use of the dicrotic notch in this fashion in polygraphic work is open to criticism, for in crude curves the dicrotic dip is not properly represented ; the actual dip which is seen is often largely an artifact. In practice, however, only a rough gauge is usually required, and experience proves these measurements to be sufficiently exact. They are of special value when used for different cardiac cycles in one and the same tracing VENOUS CURVES. 143 "^^w \f y^--^' » n ir^ Ca^DM, ■■(^" V V 1^ »^ ^'%» V 1^ ^ ^ ^^'V ^ y..^-.^. V ^ ^ ^l^^lflf or ■VCTLOOS vyaucUaX- Fig. 86. This figure includes four polygraphic tracings taken from a single patient. The lower curve in each strip is from the radial artery. The upper curve in all tracings is from a single and fixed point in the neck. The differences in the neck curves are dependent upon the pressure at which the receiver was applied. In the uppermost curve the pressure was heavy ; in the lowermost curve light ; in the middle curves moderate. The figure shows transitions from the arterial to the venous type of curve as pressure upon the neck is relaxed and the receiver recedes from the artery. The first tracing exhibits a curve of purely arterial character ; the second tracing demonstrates a small a wave directly to the left of the carotid upstroke ; the third tracing shows an almost perfect, the fourth a fully developed, phlebogram. Vertical lines have been drawn at approximately corresponding points in the four tracings. The figure illustrates a simple method of proving the precarotid time-relation of the wave o. 144 CHAPTER XI. The interpretation of oxjrves in which there are disturbances of sequence. Since the identification of " a " waves is the aim of venous pulse work, no wave should be so lettered until all other sources of wave production at the given instant are excluded. The interpretation of abnormal curves hangs largely upon our knowledge of experimental irregularities. Where the usual form of v is modified from point to point in a curve, and where an unusual and synchronous event is observed in the arteriogram, the source of v's variation should be sought dihgently. If the amplitude of a given v appears to be exaggerated, it may be suspected that an auricular systole has fallen during the ventricular systole, a view which is confirmed if the time-relations of the suspected v wave are not the customary ones (Fig. 91). •••ry ^ " ^rnrir >■»' t v x-v «'if * » ii'nr *» * «"

» » » » »!>» Fig. 91. A clinical polygraphic curve showing two premature contractions of auricular origin. The radial pulse is regular at first and is then disturbed by two unusually long pauses. Opposite each pause an exaggerated wave marked with an asterisk is seen in the venous curve. These two waves are of like nature. To take the former as an example ; it might seem natural to assume this wave to be the v wave of the corresponding ventricular systole ; but it is much larger than the last v wave and differs from it in form. These features should themselves draw attention to the wave in question ; especially if, as in the present instance, it is repeated whenever the unusual pause is seen in the radial tracing. Its nature is determined to be auricular, when its relations to ventricular systole are examined. The latter begins in the jugular on line 3, it ends at line 6, lines which coincide with the beginning of c and the summit of V. But the line 6 for the particular cycle considered falls clear of the wave marked by an asterisk. The latter is therefore no v wave and must have been produced by an interference, such as contraction of the auricle. Further and conclusive evidence is to be found when the auricular contractions are spaced, for it is then ascertained that the dominant (auricular) rhythm has been disturbed. interval is partly responsible for the coincidence of contraction in many of these instances, so that even with little acceleration of the heart's action an auricular contraction of one generation may fall with the ventricular con- tractions of the preceding generation, as Wardrop Griffith has aptly described VENOUS G U BV ES 147 Fig. 92-94. Three curves from one patient and taken at one sitting. They illustrate changes in the jugular curve resulting from alterations of the a-c interval. Fig. 92. In this curve the c waves in the jugular were readily determined in the usual fashion. The remaining waves are those marked by asterisks. What is their origin ? Are they w or a waves ? If the limits of ventricular systole are ascertained approximatelj', these waves are found to lie in ventricular diastole, rising to summits which lie nearly two-fifths of a second after the end of the ventricular systole. They are a waves, therefore, and the a c intervals are prolonged to nearly half a second's duration. cl. ■'» ) II ■ \i—r^f~v V r y~* » > » * « — v— v—^r- v~r-v--»--»--^<--v— v— -y—v— ir-w~v~v~v Fig. 95. Polygraphic curves from a clinical case. The c waves are each preceded by presystolic waves a, and there are certain additional waves of similar appearance, falling outside the limits of ventricular systoles, and pi aced equidistantly from preceding and succeeding a waves. They are also the result of auricular contractions. The auricle is contracting at exactly double the rate of the ventricle. Fig. 96. Venous and arterial curves from a boy in whom the heart's action was slow and somewhat irregular. Each heart cycle is accompanied by a, c, v and 6 waves ; the intervals between v and 6 are constant, between 6 and a they are inconstant. Supernumerary waves. If in a curve from a patient, where three clearly inscribed waves accompany each ventricular systole (waves which may be identified as a, c and v), a fourth wave occurs in the diastolic period of each cycle, it may be due to an auricular systole yielding no ventricular response. The interpretation is confirmed if the fourth wave is of similar form to the known a waves in the same curve, and if it is placed equidistantly between the preceding and succeeding a waves (Fig. 95). In these circumstances the pulse is generally slow. When the fourth wave is inconspicuous, and particularly when the remainder of the curve shows no evidence of impaired conduction, it may possess a different origin. V E NO US U nv ES. 149 When the pulse is slow, a physiological wave, described by A. G. Gibson (225), is sometimes found in mid-diastole (see page 29). He termed it b, and is not difficult to recognise if different heart rates are studied in the same case ; sufficient change of heart rate may be obtained by slight exertion, or, in some instances, by the suspension of respiration. It is also easy to identify when the lengths of ventricular cycles vary naturally from time to time. In Fig. 96, there is in addition to the usual a, c and v waves of each cycle, a prominent wave in each diastole. But the pulse is not quite regular and the diastoles have a varying duration. If the positions of the super- numerary waves, relative to the preceding c and v waves and relative to the succeeding a waves are noted, it will be found that, when diastole is short, the extra wave (&^) lies much closer to the next a wave, than when diastole is long (6^). But the supernumerary waves lie at practically a constant distance from preceding c and v waves ; they belong, then, to the preceding systole of the ventricle.* Changes in the frequency of the heart rate or the chance occurrence of longer pauses in the ventricular rhythm are also helpful in discovering impaired conduction where the heart beat is regular for long periods and auricular and ventricular systoles are fused ; for at the longer pause the two waves separate. When in addition to clearly established c and v waves there remains a series of further waves, falling at regular intervals but bearing no fixed relation to ventricular systoles, the regular and added waves are due to auricular systoles (Fig. 97). In such instances, it will be noticed that where an a wave falls during the limits of ventricular systole, its prominence is enhanced (Fig. 97), Records of premature beats. In the interpretation of curves in which single premature beats are transmitted to the radial tracing, the points in the venous curve at which the ventricular systoles commence are fixed first of all. In rare cases it may be necessary to calculate the actual transmission interval for the premature beat (Fig. 98). If the wave c corresponding to the premature ventricular systole is immediately preceded by another wave, this last wave is due to auricular systole (Fig. 99). The presence of a premature auricular contraction is confirmed if the dominant rhythm of the heart is disturbed. But if no such wave is found and an exaggerated peak occurs at the instant of ventricular systole, then the auricle and ventricle have contracted together (Fig. 87 and 88, page 144). Absence or apparent absence of " a " waves. When the venous curve presents no distinct a waves, and when the chief summits of the curve fall without exception within the limits of ventricular * The nature of b waves was discussed on page 29. 150 C HAPT E R XI. Fig. 97. Polygraphic curve from a caise of Adams-Stokes' Syndrome The c waves have been identified in, the usual manner, and there remains a series of waves, scattered at approximately equal intervals in th& curve, some of which are presystolic in time. They are all due to auricular contraction. Where, as happens in one instance, u, and c fall together, a very exaggerated wave results. The auricular and ventricular rhythms are dissociated ; the auricle is beating somewhat more than twice as rapidly as the ventricle. Fig. 98. A curve taken from a dog's carotid, by means of a Hiirtlile manometer. It shows two weak pulsations marked with asterisks. These resulted from premature contractions arising in the ventricle. Yet the carotid beats are not premature, but delayed. The increased delay may be calculated fromi the short ordinates which represent the actual times of onset of ventricular systoles, determined from a simultaneous curve taken direct from the ventricle. This fallacy in estimating the onset points of ventricular systoles from arterial curve is one which shoiild be kept in mind {256), though it is but rarely responsible for an inaccurate interpretation. Fig. 90. Clinical polygraphic tracing, in the radial curve of which a single premature beat occurs. It is accompanied by waves c' and v'. Each of the rhythmic beats is associated with a, c, and v waves. An additional wave (a') at the point marked with an asterisk is attributed to premature contraction of the auricle. VENOUS G U RV ES . 15! systole, the curve exemplifies the ventricular form of venous pulse. This form of venous pulse is encountered in m^ny different circumstances. The arterial pulse accompanying it may be regular or irregular. A. When the arterial pulse is regular the ventricular form of venous pulse is due to one or other of the following conditions : — (a) engorgement of the auricle (Fig. 100). In such curves, c and v tend to fuse and form a plateau ; the movements of the distended auricle are incapable of causing a noteworthy impression upon the curve. Fig. 100. Venous, radial and electrocardiographic curves from a case of cardiac failure with engorgement of the auricles. The venous curve is of the plateau variety and the a waves are not at all distinct. Yet the electrocardiogram shows that the auricle is contracting in early diastole. Time in fifths of a second. (b) simultaneous contraction of auricle and ventricle, which may itself be produced in one of several ways {4 and 446). 1. Auricular "contractions faUing with ventricular contractions of the preceding generation (long a-c interval) (4, 762). 2. Retrograde contraction of the chambers {434) (see Chapter XXI). 3. Simultaneous response of auricle and ventricle to impulses formed in the A-V node (see Chapter XV). (c) slow and regular action of the ventricle, associated with fibrillation of the auricle (a condition fully described in Chapter XXV). L 2 152 C HAP TEE XI. \b[n\\0 Fig. 101. Fig. 102. Fig. 101 and 102. ( Heart, 1910-11, II, 130, Fig. 1 and 2. ) Two polygraphio curves from a case of paroxysmal-fcaohycardia. The first figure shows the normal curves of the slow periods. The second figure shows the ventricular form of venous pulse. The prominent waves fall during the period E, which marks the limits of ventricular systole. Resulting from simultdneovis contraction of auricle and ventricle. B. When the arterial pulse is irregular, the ventricular form of venous pulse is almost always due to fibrillation of the auricles (see Chapter XXIII). Chapter XI I. THE ANALYSIS OF DISORDERED MECHANISM. ELECTRO CAB DIOGB A PHIC CURVES. The analysis of most disorders of cardiac mechanism may be accomplished by studjdng arterial or polygraphic curves, and the principles underlying these methods are summarised in the two preceding chapters ; they are practical diagnostic methods. The final path to accurate knowledge is the electrocardiographic method. For purposes of investigation its precision is unrivalled. It presents numerous advantages over other methods. The curves are directly inscribed by the muscle of heart chambers whose activity it is desired to record. The auricular and ventricular systoles* are clearly depicted and their relations to each other are accurately expressed. The curves are uncomplicated by transmission intervals. One fibre yields the double record, and transference of measurement from one curve to another is obviated. The electrocardiogram displays all the events in auricle and ventricle. Except on rare occasions, it provides a full analysis of the disorders examined. When, as sometimes happens, certain auricular complexes are obscured, it may be necessary to add a simultaneous record from the veins. The electrocardiograph not only records the activation of auricles and ventricles ; it speaks clearly of the direction of the excitation wave in these organs. It has been repeatedly laid down that the form of the galvanometric curve, corresponding to the activity of a given muscle mass, is controlled by the order in which the muscle elements become excited. This, the first grammatic rule of a new language, is to be borne constantly in mind. If the left ventricle is excited at its apex, the corresponding axial electro- cardiogram is in the main diphasic (Fig. 103) consisting first of a downward phase and finally of an upward phase. If the right ventricle is stimulated at its base, the curve is of a diflierent type ; the first phase is upwardly directed and the second phase is downwardly directed (Fig. 104). It is true that the form of an electrocardiogram may be affected in a minor degree by change in the position of the heart relative to a given pair of electrodes ; it is also true that hypertrophy by producing preponderance of one or other chamber may influence the amplitude and even the direction *The term systole is used as a convenient expression, the actual record is, of course, not of contraction but of the electric change which very slightly precedes and accompanies systole. 154 CHAPTER X f I . Fig. 103. Fig. 104. Fig. 103. Electrocardiogram from lead II. A dog's heart is responding to rhythmic stimuli (see signal marks at bottom of each curve) applied to the apex of the left ventricle. Time in fifths and twenty -fifths of a second. Fig. 104. A similar curve from the same animal lead, showing the effects when the stimuli are applied to the conus arteriosus. These two figures illustrate the rule that the direction of spread regulates the shape of the electrocardiogram, which is consequently dependent upon the point at which the impulse originates. of deflections. The curve is outlined from moment to moment by the direction of spread relative to the lead, and by the energy of the combined electrical discharges ; but as the anatomical relations of the heart to a customary lead do not often vary very materially even in pathological conditions, and as the changes in the curves form this cause and from an abnormal balance of the right and left musculature are rarely profound, the direction in which the excitation wave spreads and is distributed constitutes the chief factor. Auricular and ventricular contractions of physiological types. Ventricular contractions of supraventricular origin. — When an impulse descends to the ventricle through the A-V bundle and is distributed through the uninjured arborisation of Purkinje, the excitation wave spreads normally throughout the ventricle. The corresponding electrocardiogram invariably consists of a group of initial deflections {Q, R, S group) which represents this normal spread, followed after an interval by a more or less prominent T deflection. It is a matter of indifference whether the impulse originates in the normal pacemaker, in the substance of either auricle, in the ELECTROCARDIOGRAPHIC CURVES. 155 A-V node or in the bundle down to the point of its subdivision ; the spread in the ventricle is precisely the same, the resultant ventricular complex is the same, in each instance. Whenever the form of the natural electrocardio- gram is known for a given patient, any systole of the heart which yields an electrogram of this type is promoted by what is termed a supraventricular impulse. That is to say, its impulse arises above the point of division of the A-V bundle and is consequently distributed to the ventricle in a fixed and orderly manner. This rule is absolute, whether the beat in question belongs to a rhythmic series or is the cause of an irregularity. Even when the normal electrocardiogram is not available for comparison, if the ventricular complexes — individual ones or all — conform to the general type of natural complexes, their supraventricular origin may be assumed. The converse conclusion, that a ventricular systole of supraventricular origin will yield a natural electric curve, usually holds good for pathological as for normal conditions ; but as will presently be seen there are important exceptions. Beats arising in the vicinity of the pacemaker. If a ventricular complex of supraventricular type is preceded by an auricular complex of normal outline — a rounded or peaked summit — the spread of the excitation wave through the heart as a whole may be assumed to be natural and the origin of the beat may be stated to be from the vicinity of the pacemaker. Clearly this conclusion is especially serviceable when applied to isolated systoles the origin of which is to be ascertained ; and it is absolute if the form of the electrocardiogram for such beats is in all detail a replica of beats known to be physiological in the same subject. Anomalous beats. When those minor changes in the complexes associated with malposition of the heart or hypertrophy of its chambers are excluded, departures from the natural type are the result of a single cause. This cause is abnormal spread of the excitation wave. As a natural electrocardiogram portrays a natural spread, so, with the exceptions named, an anomalous electrocardio- gram portrays an unusual distribution of the excitation wave. This unusual distribution may be confined to auricle or to ventricle. When it occurs in the auricle it is due to an unusual starting point of the heart beat. If systoles are propagated serially from the pacemaker, and this series is disturbed by a systole arising in some other portion of the auricular musculature, all the ventricular complexes of the curve will be alike, but the auricular complex corresponding to the disturbance will be of unusual form (Fig. 105). Arising in a new focus the excitation wave is propagated along abnormal auricular channels and a change in the corresponding portion of the electric curve results ; but as the focus is 156 CHAPTER XII supraventricular the spread to the ventricle takes place through normal channels and the succeeding ventricular complex therefore suffers no distortion. ri=3r:.r: ;P^-.4A^.: -- -^-'i¥ ";_:rz:-Pr-7:. L_ _ p I T . p^ T ', r p -■ T :■'" I -— » 2 f' --^^& -r - ^ ' lA ' -* * " """^^ — "~a~4ti m^^mm HHPVb-yV^B-^MHMMM^H i^'^^— ^'^W^ ',^^^ si pifc.^i,,-' t^S: :r-;^---„4>-S__ _t°j? i' ,^S _ _^-::- Fig. 105. A clinical electrocardiogram, showing a normal heart rhj'thm, is disturbed by a, premature beat {the fourth). This fourth systole presents a ventricular complex of normal outline ; it has arisen therefore in a supraventricular focus. It is preceded by an auricular complex -which is inverted, showing that the excitation -wave has taken an abnormal course in the auricle and that the corresponding impulse arose in an abnormal situation in the auricle. Time in thirtieths and fifths of a second. If the systole disturbing the normal series is propagated from the ventricle, the corresponding ventricular complex is altered (Fig. 106), because, when the excitation wave starts at any point below the bifurcation of the bundle, the spread of the excitation wave in the ventricle is abnormal. While this is the cause of most anomalies in the ventricular complexes, the same explanation does not always hold good. For as the shape of the electric niiiiimHiiiniiiMNniiiijhiillilllllniiMiiiHiiunii ujiiiiiiuMnHj .,|,,..,,.|m^„ Fig. 106. A curve showing two anomalous beats interpolated between natural heart contractions in tt patient. These beats arose at some focus in the ventricle, and the corresponding ventricular complexes are abnormal, since the excitation wave has pursued an abnormal course in the ventricular muscle. Time in thirtieths of a second. curve is governed by the direction in which the excitation wave travels, this direction may be modified, not only by an alteration of the point from which the contraction originates, but by changes in those special channels of conduction which exist in the ventricle (Fig. 107). An impulse entering the ventricular segment of the cardiac tube through the auriculo-ventricular ELECTROCARDIOGRAPHIC CURVES. 157 bundle passes into the two bundle divisions ; but conduction may be defective in one of these, or it may be defective in more distal branches of the Fig. 107. An electrocardiogram taken from a cat during asphyxia. The ventricle responds to each regular auricular impulse (P) ; all the ventricular beats are therefore of supra- ventricular origin ; but the first, third, and fifth responses of the curve are unnatural. Tliese complexes show deepening of S and lifting of T, changes which are due to aberration. Time in thirtieths of a second. arborisation. The course of spread in these circumstances is aberrant (see Chapter IX) and according to the degree of aberration the electric curve is correspondingly modified. So far as we know " aberration " is peculiar to the ventricle ; the auricle is exempt because in its structure it possesses no special conducting paths. Identifying different types of anomalous contraction. If an electrocardiographic curve exhibits beats of physiological type, and beats of anomalous type occur as occasional interruptions, the contrast between one type and the other is generally conspicuous and the anomalous beats are at once recognised. But though the dissimilarity is usually manifest, the custom of comparing the anomalous beat with those of physiological type in the same case, cannot be too strongly urged. For what is a somewhat anomalous type in one case may be physiological in another ; the departure from the physiological type is at times inconspicuous ; this is the case where there is but slight alteration in the spread of the excitation wave. The cause of a given anomaly is usually known as soon as the curve is inspected, for the common anomahes are characteristic and speedily become memorised. Nevertheless, each curve should be treated on its own merits. When we deal with a premature beat, the ventricular complex may be anomalous. In that case, and if it is not preceded at a natural interval by an auricular systole, the ventricular origin of the beat is known. If the ventricular complex is natural in outhne and is preceded by an anomalous auricular complex, we recognise that a new focus in the auricular muscle has originated the disturbance. But if the ventricular complex is abnormal, and at the same time it is preceded by an auricular complex of unusual type, then not only is there a disturbance in the site of auricular impulse formation, but the paths of conduction through the ventricle are also defective (see Fig. 112a). 158 CHAPTER XII. Aberrant contractions — which as previously stated are confined to the ventricle — may be identified if they conform to a recognised type ; or if, departing in form from physiological beats in the same subject, they are preceded by auricular systoles, for these betray their supraventricular origin. The chief forms of aberrant contraction are described in a previous chapter ; other forms will be described at a later stage. The algebraic summation of complexes. In many examples of disordered mechanism it happens that ventricular and auricular systoles fall synchronously, and in discussing venous curves, we saw that simultaneous contraction of the upper and lower chambers produces a special form of phlebogram, dependent upon hindered discharge of the auricular contents. Individually the electric complex of the auricle and of the ventricle are not appreciably changed by such coincidence of the contractions, and the curve which results is consequently a simple composite of auricular and ventricular complexes. The auricular systole may fall at any point upon the ventricular systole, and its representative (P) will be found, superimposed upon R, S or T. The summation of effects is exact algebraically. When complete dissociation of the auricular and ventricular rhythms is present, the deflections P are of normal form and the deflections R, S and T show a close resemblance to the normal type. Auricular complexes, representing beats arising in the vicinity of the pacemaker are superimposed upon ventricular complexes of supraventricular type (Fig. 108). =^ Fig. 108. A curve from a j)atient who displayed complete heart-block, or dissociation of the auricular and ventricular rhythms. The auricles and the ventricles are beating regularly, but at the independent rates of 78 and 29 per minute, respectively. The exact superimposition of auricular and ventricular summits should be remarked. Time in fifths of 11 second. If the individual cycles from such curves as those of Fig. 108 are excised, they may be re-arranged above each other, not in the order in their natural sequence, but in an order which renders the superimposition more manifest. Such a re-arrangement is seen in Fig. 109. Traced from above downwards, the first auricular summits pass gradually into, ELECTROCARDIOORAPHIC CURVES. 159 through and beyond the opening phases of ventricular systole ; the second row of auricular summits continues the tale, and shows the passage over and clear of the broad summit T. The summation is always exact ; the height of R in the fifth curve of this figure is especially noteworthy: the ampHtude is that of the natural R and P combined. Fig. 109. A figure constructed from the electrocardiograms of the patient whose curve is shown in Fif. 108. Single ventricular complexes have been rearranged above each other, not in the order of their natural sequence, but so as to display the varying relation to them of the auricular complexes. When a premature ventricular contraction coincides with a rhythmic auricular systole, a P of normal type is superimposed upon an anomalous ventricular complex (Fig. 110 and 111). It is not always easy to identify such auricular beats, for the exact and uncomplicated form of the anomalous 160 CHAPTER XII. ventricular complex may not be known. But where the prematurity of the ventricular contraction is variable, even if only slightly variable, as in Pig. 110, the auricular portion of the curve can often be identified by noting the slight differences in outhne between the two anomalous ventricular complexes. The buried auricular summit, when found, is seen to occupy a midway position between preceding and succeeding auricular contractions. The buried auricular summits can also be identified when two premature beats occur in succession, for here too the anomalous ventricular complexes vary, the auricular summit falHng with one only, and falling equidistantly between neighbouring auricular deflections (Fig. 111). Fig. 110. A clinical electrocardiogram showing two premature contractions of ventricular origin. The auricular contractions are regularly dispersed throughout the curve ; but two of them (P^ and P^) are buried in the anomalous ventricular complexes. That these summits are not part of the anomalous complexes is proved by the slightly different time-relations in the two instances. Time in thirtieths of a second. MT'illB." ii_r.._LL___ I £j _.„:ii ii iii i iiuituiui-mimiJiJumniTiiuuu^ifi' i uuijiinmiii i iuimmuiijm i Fig. 111. Two premature contractions arising in the right ventricle of a jiatient. They occur together and replace a single normal ventricular contraction. The rhythmic auricular contraction falls with the first premature beat. Time in thirtieths of a second . When a premature and anomalous auricular contraction coincides with a rhythmic ventricular systole, an anomalous P is superimposed upon a ventricular complex of normal outhne (Fig. 112a and b]. ELECTROCARDIOGRAPHIC CURVES. 161 Fig. 112a. A premature contraction of auricular origin in a patient. The first three cycles are normal, but T in the last of these is notched at its beginning by the j)remature P which is of inverted form (compare the 2nd and 3rd T summits of the figure). This premature P is followed by a ventricular complex which differs from the normal on account of "aberration" in the ventricle. All the beats are of supraventricular origin. Time in thirtieths of a second. Fig. 1126. Two premature contractions arising in the auricle are shown in this patient's record. The auricular complex P is inverted and notches the preceding T in each instance. The ventricular complex of the second premature beat shows slight aberration. Time in thirtieths of a second. The original observations upon which the principles of analysis discussed in this chapter are based, are described for the most part in the iollowing original publications {111, 390-392, 428, 433, 445, 447, 463). In the preceding chapters I have dealt with the principle rules guiding the analysis of disordered mechanism and have spoken of the more important phenomena observed in studying arterial, venous and electrocardiographic curves respectively. The observations outlined in those chapters will serve as a basis in proceeding to examine the chief disorders of the heart beat as they are met with in clinical work. Thus, while the first section of this book is devoted to anatomy, physiology and method, the second section will be devoted to the disorders themselves and to correlating them with experi- mental findings. It will be my endeavour to present the experimental and clinical facts side by side, as far as possible, and to identify the naturally occurring disorder with that produced by deUberate interference with the heart in the lower animals. The second section will also contain the evidence upon which many of the conclusions of the first section, and in particular the general rules of interpretation, are founded. Chapter XTir. EXPERIMENTAL HEART-BLOCK. In Chapter VI, we saw that the sequential contraction of auricle and ventricle depends upon the integrity of the auriculo-ventricular bundle, and that experimental lesions of this neuro-muscular tract produce heart-block. We have further concluded that the impulse is conveyed to the ventricle from this bundle through its branches and their arborisations. It is permissible to make the more sweeping statement that the proper sequence of contraction hangs upon the functional integrity of the junctional tissues as a whole. Unless there is an intact pathway, traversing the node, the bundle and one or more of its branches, the responses of the ventricle are unpunctual or fail. When a clamp is applied to the bundle and compression is gradually increased, heart-block displays itself, as Erlanger (143) showed, in successive stages ; these stages comprise : — (a) An increase in the interval between the onset of auricular and ventricular systoles ; an increase of the As-Vs interval as it is termed, which in the normal dog approaches O'l sec. (6) An occasional failure of the usual ventricular response to the regular auricular impulses. (c) A failure of ventricular response to each tenth, ninth, eighth, seventh, sixth, fifth, fourth or third auricular contraction. {d) A failure of the ventricle in its response to alternate auricular beats ; the establishment of-2 : 1 heart-block in which the auricle beats precisely twice as rapidly as the ventricle (Fig. 113). (e) The response of the ventricle to each third or fourth beat of the auricle ; so-called 3 : 1 or 4 : 1 heart-block, of which the latter is by far the commoner in experiment. The foregoing grades of block are spoken of as partial. (/) The onset of complete heart-block (or dissociation) between auricular and ventricular rhythm, i.e., entire failure of conduction (Fig. 114). Soon this is associated Mdth a new event, namely, the wakening of a rhythm inherent in the ventricle. Examples of heart-block produced by compressing the A-V bundle in dogs are shown in Fig. 113 and 114. EXPERIMENTAL H EAUT-B LOCK 163 Kig. I 13. (X -g-.) The figure exhibits 2 ; 1 heart -block, the result of lightly clamping the bundle in H, dog; the auricle is heating twice as rapidly as the ventricle Alternate auricular systoles, those to which the ventricle fails to respond, fall within the confines of ventricular systole ; the corresponding P summits appear immediately prior to the T summits. Time in fifths of a second. Fig. 114. (x ^.) Two curves from another animal : the first taken immediately after the clamp was applied ; the second after firm compression. The top curve exhibits a slight prolongation of the P-R interval. The bottom curve shows complete dissociation ; in this figure auricle and ventricle beat regularly but at independent rates and the relation of the ventricular cycles to adjacent or simultaneous auricular cycles is variable. These two curves are also published to illustrate the similarity of the ventricular complexes, before and after the disso- ciation. The right bundle division was also injured in this instance, whence the anomalous character of the ventricular electrocardiogram. Time in fifths of a second. Methods of producing heart-block in animals. Heart-block between the auricles and ventricles may be produced, in one of three chief ways. 1. By direct interference with the conducting tracts, namely, the auriculo-ventricular node, the bundle or both divisions of the latter : — The experi- ments in which this tract has been injured by pressure or section have been described in detail. Heart-block may be produced more delicately by cooling this region of the heart (41, 797). 164 CHAPTER XIII. 2. By stimulation of the vagus. — That vagal impulses are capable of producing heart-block has been recognised since Gaskell's observations {214). In studying conductivity changes in the frog's heart, many workers have adopted almost exclusively this method of producing heart-block. The vagal tone may be increased reflexly by stimulating the gastro-intestinal tract (127). Chauveau (47), one of the earliest observers to record heart- block in the mammal, produced it in the dog by vagus stimulation. Rothberger and Winterberg (665), working upon the dog's heart, suggested that the cardiac nerves are distributed in a special mannej', it being supposed that the right vagus is distributed mainly to the S-A- node, the left vagus to the A-V node. These writers state that weak stimulation of the right nerve usually produces slowing of the whole heart, while similar stimulation of the left nerve has a less conspicuous retarding action, but produces defects in A-V conduction. These effects of vagal stimulation have been amply confirmed (57, 60, 466). That the S-A rhythm is most readily inhibited by stimulation of the right vagus is undoubted ; the difference between the two nerves is profound in this respect. But that A-V conduction is affected chiefly by stimulation of the left nerve is more open to question. The apparent difference (Fig. 115 and 116) is a profound one, but it is chiefly conditioned by the associated rates of auricular contraction. If a constant auricular rate is deliberately maintained in these experiments, the difference in the effects of the two nerves is less conspicuous ; the left nerve appears, however, to be slightly, though not uniformly, preponderant in these circumstances (74). The effects of vagal stimulation are not confined to the main bundle, an influence is exerted also upon the divisions of the bundle (60, 100, 123), for not uncommonly when partial heart-block is produced by vagal stimulation, responses of the ventricle of aberrant type are to be seen* (Fig. 117). There is, however, no definite relation between the side on which the nerve is stimulated and the division of the bundle which is affected. 3. By introducing toxic bodies into the blood-stream. — A very large number of toxic substances will produce varying grades of heart-block when they are injected into the blood-stream of animals. Amongst those which may be named are digitaHs (85, 716), strophanthine (672), aconitine (87), muscarine (663), physostigmine (663), nicotine (594), glyoxyUic acid, morphia (123), adrenaHn (360), and potassium salts (546). Certain of these substances, for example morphia and adrenalin, are known to act through the vagus centrally, for the disturbances are abohshed by section of the vagi ; others such as the salts of potassium appear to act directly upon the musculature of the heart. * These observations are originally experimental ; Wilson has recently recorded instances in which aberrant curves were seen to follow vagal stimulation in the human subject (783), EXPERIMENTAL HEART BLOCK 165 p i? 1 — 1 ^^/ .= : v.; 1 = = 1 = 1 3 =: = i = = = = = P = = ' > ■l^ = ^ i^ :^ = = L y i 1 t.1 \ - g j nf 1 ■=: 5 1 " 1 ■ M M m i in — ./ \ * 1^ ^ ^ %% 1 — — - — , -f 1 % i i » f-:: '\^. 1 1 •1 ^ i 1 ;; 1 = 1 ■I M -j - m 1^ i -- v-l 1 1 ^ I % H=^ 1 1 i I 1^ 1 -7' i I 1 i 1 -i"- 1 1 := 1 1 1 1 1 1 i f p- \ 1 \z i i i \ 1 ? 1 ! 1 = K -3 ~ "^-Ts 1 e r? 1 '^ 1 ^z: BB r^i :^ i lU KW r= li — T~ ■M ^ - - _-_ -3 ■ f: = = = zz Fig. 115T~ (X \x') Myooardiograms from auricle (4) and ventricle (F) and electrocardiogram, showing the custonxary effect of stimulating the right vagus in the dog. The P-R interval widens a little at first and then the whole heart ceases to beat. The signal of stimulation is seen below. Time in fifths of a second- Fig. 116. (X 1&.) Similar curves from another animal showing the customary effect of left vagal .stimulation. The P-B, intervals widen out, and the auricular rate being only a little diminished, the ventricle fails to respond, In this instance, only a single ventricular beat is missed. Time in fifths of a second. M 16G CHA PT E R XIII. Fig. 117. (x y"^.) Similar curves, showing an iinusual effect of stimulating the right vagus in the dog. A-V heart-block results and the character of the complexes of ventricular responses alters in a manner indicating hindered conduction in the left division of the bundle. Time in fifths of a second. Heart-block may be very readily induced in cats, less readily in dogs, by asphyxiating them {481, 699). Such heart-block is independent of the vagus, for it occurs equally after full saturation with atropine or after vagal section (481) ; it is also independent of blood-pressure changes, of dilatation of the heart and of excessive accumulation of carbon dioxide in the blood (545) ; it probably results from lack of oxygen and is perhaps due to an excess of acid products in the blood. The heart-block may be of any grade, partial or complete (Fig. 118-121) but, when ventilation is restored, recovery is rapid and perfect, providing that the experiment is not carried too far. The administration of fatal doses of diphtheria toxin is followed by A-V heart-block ; the same failure of conduction is witnessed in death from anaphylaxis (12, 651). Records of experimental heart-block.* Records of the heart beat, subsequent to damage of the A-V bundle, or subsequent to other interferences productive of heart-block, may be taken in many ways ; but we may confine ourselves at the present time to a brief description of the electric curves, for they are particularly legible. The curves obtained during asphyxia are used for illustrative purposes. * While the experimental records of mammalian heart-block are described before the clinical for purposes of convenience, it is interesting to note that the majority of the observations were first made upon patients. EXPERIMENTAL H E A RT BLOCK. 167 UMUUUUMkUUMiUiUMMMMUMrUMMlM MMMMUMMMMMMM Fig. 118. (x |.) An electrocardiogram from «, oat, after one and a half minvites of asphyxia. In the first portion of the curve the P-R interval widens. This widening increases at the seventh cycle and the next P summit falls with T of the ventricle. A single response is missed, and the P- R interval then shortens up, gradually to widen once more as sequential contraction is resumed. Fig. 119. (x |-.) A later stage of the same period of asphyxia ; 2 : 1 block is established. Fig. 120. (X 4.) After profound asphyxia in which dissociation developed, ventilation was re-established. The heart recovered completely. The curve shows the recovery from the stage of 2 : 1 heart-block to that in which there is at first some prolongation of the conduction interval. The simultaneous curve is aHiirthle manometer tracing of carotid pressure. Fig. 121. (X ^■) A curve of complete heart-block, in which the rhythms of auricle and ventricle are each regular but independent. Resulting from profound asphyxia. Time in all this series of curves in thirtieths of seconds. If a cat is asphyxiated for periods varying from 1-7 minutes, a regular succession of events is observed in the heart. Within 1-3 minutes of the onset of asphyxia the P-R interval (representative of the ^s-Fs interval) shows a notable prolongation. This phase may last for a shorter or longer time, and is succeeded by a condition in which single responses of the ventricle to auricle are missed (Fig. 118). At this stage the P-R interval shows changes of interest and importance. The P-R interval, which is primarily M 2 168 CHAPTER XIII. increased, shows a gradual and further increase up to the point where the response is missed (Fig. 118); the delay in the ventricular response may be so great that a given ventricular contraction coincides with the following auricular contraction (the P-B interval in experiment may lengthen to 0-3 sec. at least). Subsequent to a failure of transmission and following the resultant ventricular pause, the interval decreases abruptly and the whole process is repeated. The alteration of the P-R interval changes the time relations of the ventricular beats. The dropping of a single ventricular beat is not accompanied by a pause equal to two heart cycles, but to a pause of shorter duration. Thus the grade of the irregularity in the ventricle is diminished. Minor changes in position occur in the remainder of the ventricular cycles ; as the P-R interval increases, the rate of increase in P-B interval diminishes and the ventricular rate consequently quickens slightly. The phase of prolonged P-R interval with dropped beats is succeeded by one of 2 : 1 heart-block, in which only alternate systoles of the auricle awaken ventricular responses (Fig. 119). Finally, complete dissociation develops, in which independent rhythms are found in auricle and ventricle, each regular, but bearing no simple mathematical ratio to each other (Fig. 121). In the electric curves the ventricular systoles ( R and T) are represented at regular intervals in the curve, so also are the auricular systoles (P), but they fall in haphazard relation to the ventricular contractions. Where auricular and ventricular contractions coincide, accurate summation of the electric effects is noted. The susceptible region. When partial heart-block is produced by compression of the A-V bundle, the shape of the ventricular complexes does not change. This is no more than is to be expected, since the auricular impulses still flow along the accustomed channels and the distribution of the excitation wave to the ventricle is undisturbed. The curves are usually similar when partial heart-block results from vagal stimulation or from poisoning. It is almost inconceivable that the spread to the ventricle would remain unchanged as it does unless the vagus or the poison exerted its chief effect at a level of the conducting system where it is concentrated to form a single strand. For if the effect were produced more peripherally, and were not identical in its degree throughout the main stems of the conducting system, the spread to one part of the ventricle would be relatively quicker than to another, and a. noticeable alteration in the form of the resultant curve would be seen. This argument suggests that the influence, nervous or toxic as the case may be, is a local one. i To take the vagus as an example, heart-block due to its over-action might conceivably be brought about by the influence of this nerve upon the conducting tract, or it might be supposed that its action is exerted upon the EXPERIMENTAL HEART-BLOCK. 169 ventricle, rendering this organ irresponsive. In this instance we are in a position to define its mode of action. That the whole conducting tract is under some measure of vagal control seems clear, for under stimulation of one or other nerve, defects in conduction in the bundle divisions may be witnessed from time to time (Fig. 117, page 166) ; the complexes of the responding ventricle are anomalous, and the anomaly is recognised to result from aberration. But that the vagal influence does not produce A-V block by acting upon the bundle branches is suggested by the comparative rarity of branch conduction defects. Its influence is concentrated at a higher level of the system, and the complexes of the ventricle in the electrocardiograms remain unaltered in form. It has recently been shown by experiments specially devised for the purpose that the most susceptible point lies in the region of the junction between A-V node and auricular tissue {466). If heart-block is induced by stimulating the vagus, while the auricle and ventricle are responding simultaneously to impulses arising, not in the S-A node but in the A-V node, conduction to the auricle suffers in greater degree than does conduction to the ventricle. Like- wise, in the heart-block of asphyxia, it has been demonstrated {482) that the same region suffers in highest degree and that, when conduction Fig. 122. (Heart, 1913-14, V, 289, Fig. Id..) (x f.) An electrocardiogram showing reversed heart-block. A cat was asphyxiated until the several grades of heart-block, shown in Fig. 118-121 were seen. After recovery it was again asphyxiated and at the same time cold was applied to the sulcus terminalis. The last procedure depresses impulse formation in the S-A node, and the A-V node then usurps its function (see Chapter XV). While this ^4 -F rhythm is in progress, each impulse makes its way upwards to the auricle and downwards to the ventricle. It is the passage of the impulse upwaxds to the auricle which is affected by asphyxia, the response of the auricle at first lags by comparison with that of the ventricle. Eventually auricular response fails, while ventricular response continues. The present figure shows this failure of the auricular response at alternate beats. The two auricular complexes P which are seen in the figure are inverted (because the passage of the excitation wave through the auricle was reversed) and fall in the centres of corresponding ventricular responses. The mechanism is illvistrated by the diagram below the photograph ; the black rectangles represent the systoles of the auricle and ventricle and the impulse is represented as starting between auricle and ventricle and spreading simultaneously to each at alternate beats, and to the ventricle only at alternate beats. The region of block during the asphyxial period clearly lies immediately to the auricular side of the focus which creates the heart's impulses, Tirne in fifths of a second. 170 CHAPTER XIII. from auricle to ventricle is prolonged or ventricular responses are missed, the obstruction is at the^-F node (Fig. 122) or at its junction with the auricle. It is true of asphyxia, as it is true of vagus stimulation, that aberrant forms of ventricular complex appear from time to time ; but this does not affect the chief conclusion. The vagus and those poisons which have been sufficiently investigated produce heart-block, not by a generalised action on the ventricle, though the whole organ is exposed to their influence, but by a selective action upon the special tissues ; the intense action is upon the upper extremity of the A-V node. This conclusion is confirmed by studjdng complete heart-block resulting from the more intensive action of the same causes, for even when dis- sociation of auricle and ventricle is thus promoted, there are many reasons for believing that the peripheral sections of the distributing tracts are still competent to conduct and do conduct (see page 198). To sum up, heart-block may be produced experimentally in one of three ways. First, it may be produced by injuring the main stems of the conducting tract, or by depressing the function of these by cooling. Secondly, it arises when vagal tone is greatly increased, the auricular rate being maintained or not greatly slowed. Thirdly, it may be produced by the injection of poisons ; some of these act through the vagus, others act, as do the products of asphyxia, directly on the muscle — not on the muscle of the ventricle, but again upon the A-V node. The general conclusion is justified that experimental heart-block, whether produced by injury or by disturbed innervation or by poisons, is produced by a change in the conducting power of the main neuro-muscular tract which unites the auricle and ventricle, meaning by the main tract, the node and the bundle proper with its right and left divisions ; it appears to be produced in this manner only. Chapter XI V. CLINICAL HEART-BLOCK. Heart-block in man was first recorded in 1875, when Galabin (196) reported a case of slow ventricular action (25-30 per minute) and remarked, " we have here a heart, the auricle of which sometimes contracted twice in the interval between two ventricular pulsations, and sometimes singly in the midst of a long pause instead of just before the systole of the ventricle." He based his account upon the auscultatory phenomena, and upon curves taken from the heart's apex. Two excellent tracings which he published show beyond question that he was dealing with what we now recognise as complete heart-block. These observations are the more noteworthy because they were published the year before those of Romanes in the Philosophical Transactions, and several years before the work upon the cold-blooded heart had begun. In 1885, Chauveau (47) described a case of heart affection and was able to recognise, in tracings taken from the radial pulse, the apex beat and the neck, that auricle and ventricle were beating at different and entirely independent rates. He compared the phenomenon to the dissociation obtained upon stimulation of the vagus. In the year 1899, Wenckebach (756) and His {319) both described heart-block in the human subject and suggested a lesion of the auriculo- ventricular bundle as its cause. The former based his diagnosis entirely upon the arrhythmia produced in the ventricle ; the latter published clear polygraphic curves. The early publications of Mackenzie {502, 504, 505, 509) notably advanced our knowledge, and during the last few years very many cases have been placed on record. (For collected cases, see 10, 108, 605, etc..) The first post-mortem examination showing changes in the A-V bundle in a case from which a clear record of heart-block had been obtained during fife was published by Hay {241) in 1905. The signs and grades of heart-block in man. Heart-block in the human subject may be demonstrated by any of those clinical methods which record the activities of auricle and ventricle. It is elegantly displayed by the galvanometric method, with almost equal clearness by the polygraphic method and less distinctly by subsidiary methods. 172 CHAPTER XIV Prolongation of the " As-Vs" interval,. — Where the defect is sHght it is shown by prolongation of the As-Vs interval, the a-c interval in the venous curve and the P-R interval in the electrocardiogram being of increased duration {444, 462). The natural a-c interval varies between 01 and 0-2 second, the natural P-R interval between 0-12 and 0-17 second. In heart-block these intervals may be much longer. An example of continuous prolongation of these intervals to approximately 0-5 seconds is shown in Fig. 123. In Thayer's case (723), the a-c interval was increased for long Fig. 123. Simultaneous electrocardiographic, venous and radial curves from a patient exhibiting an early grade of heart-block. The a-c intervals and the P-R intervals are approximately 0-45 seconds in duration. Time in fifths of a second. periods to almost the full second ; the P-R interval to 0-7 second and more. I was fortunate in seeing Dr. Thayer's original curves, and no doubt remains in my mind that in his patient the response of the ventricle to the auricle was delayed to this extent. When the heart rate is rapid, or prolongation of tlie As- Vs interval is great and the heart rate normal, the auricular systoles may coincide with ventricular systoles of the preceding heart cycles. In such cases P falls upon the preceding T in the electrocardiogram (Fig. 131, page 176) and a of the venous curve falls during the confines of ventricular systole and is exaggerated in height. Interesting examples of this kind have been pubhshed by Wardrop Griffith (234). Sometimes when the As-Vs interval is prolonged, the beat of the auricle becomes audible and produces a form of reduplicated first heart sound or, falhng near the end of the preceding ventricular systole, produces a reduplicated second heart sound (196, 461). Prolongation of the interval is also to be seen sometimes in curves from the apex beat, an observation first made by Galabin (196) ; in these curves and in those taken by a sound introduced into the oesophagus (see 349), the left auricle is responsible for the auricular summits, CLINICAL H EART-BLO CK. 173 Drop'ped heats. — When the grade of block is higher, the ventricle fails to respond to occasional auricular impulses. This form of heart-block is rarely a simple phenomenon ; it is usually associated with variations in the lengths of As-Vs intervals over the period of the disturbance. The relation of chamber contractions may be studied in Fig 124 Fig. 124. The stage of heart-block, to which the term " dropped beats " is applied. Up to the point where the chief disturbance occurs, the gaps between the auricular and corresponding ventricular contractions widen out. The impulses travel to the ventricle with increasing diflSculty. The fourth auricular contraction stands isolated, it yields no response ; a ventricular contraction is " dropped." Following the ventricular pause, the As-Vs interval is short, for the tissues have rested, but it again widens as successive cycles follow. Fig. 125. Simultaneous venous, radial, and electrocardiographic curves from a patient showing the condition described as " tlropped beats." After each long pause in the action of the ventricle, the a-c and P- R intervals are short ; in the succeeding cycles they lengthen, while the pulse rate increases ; eventually response to the auricle fails again and the events are repeated. Time in fifths of a second. A " dropped beat " produces a pause of exceptional length and this pause breaks the natural rhythm of the ventricle. Where there is no associated variation in the As-Vs intervals, the length of the pause is necessarily equal to that of two regular pulse beats. But this is rarely the ease ; the " dropped beat " is foreshadowed by a progressive increase of the preceding As-Vs intervals (cp. Fig. 124 and 125). Moreover, the As- Vs interval which follows the dropping of the beat is generally curtailed (Fig. 124 and 125). These two events shorten the long pause and conse- quently diminish the disturbance of the ventricular rhythm. The exact 174 CHAPTER XIV. manner in which, the changes happen requires closer study. Consider the first three As-Vs intervals of Fig. 124 ; as illustrated by the obliquity of the lines, the interval gradually widens, but it widens in a peculiar manner. The increase of the second interval over the first is greater than the increase of the third over the second. The result is a decrease of the interventricular period directly preceding the ventricular silence. The ventricle quickens to the -point of disturbance. The shortening of the As-Vs interval following the pause, and the subsequent prolongation of it, produces a similar quickening of the ventricle after the disturbance. These relations are well displayed in the accompanying electrocardiograms (Fig. 125 and 133); they are equally well shown in many polygraphia curves (Fig. 127 and 129). In venous curves, where a falls further back upon the preceding ventricular waves, a characteristic and gradual heightening of the a wave is seen (Fig. 129).* In some patients ventricular beats are dropped without there being an accompanying variation or prolongation of the As-Vs interval {242, 761) ; or there may be a suddenly developed complete block without the usual preliminary and slighter defects of conduction (see Chapter XXXI). The suggestion that block in these circumstances is due to depressed condition of the ventricle, and that that organ refuses to respond to normal impulses conveyed to it, is one that cannot be accepted unreservedly. | Frequent failure to respond. — When there are repeated failures in the ventricular response, a simple ratio between auricular and ventricular rate becomes established. By far the commonest of these ratios is 2 : 1 block, where the auricle beats twice as frequently as the ventricle (Fig. 126, 128 and 134). In human curves of this kind, the auricular contraction immediately following the ventricular is often a little or even conspicuously premature J, a disturbance of the dominant rhythm for which there is no present explanation, for it is the rule in partial heart-block to see no disturbance (see Chapter X). The 3 : 1 ratio may also be seen, but it is rare ; so also is the 4 : 1 ratio. More commonly the ratio is one in which 1 : 1 and 2 : 1 (Fig. 129), or 2 : 1 and 3 : 1 cycles alternate. Complete heart-block. — The final grade of heart-block is reached when no impulses are transmitted to the ventricle. When this happens, the ventricle, having lost the controlling influence of the auricle, beats in response to a slow and regular series of impulses which it forms in its own substance. In this * In instances of this kind the a-c interval may increase to as much as 0-8 seconds {800). The P-R interval to 0-6 seconds {462, 723) (Fig. 133), or even more. t Especially so as well defined changes in the bundle have been found in such cases (see case reported by Cohn and Lewis (6.9)), and in others complete block has subsequently developed. J For examples in venous curves see 236 ; in electrocardiograms the premature auricular complexes are natural, the beats therefore are probably not extrasystolic. CLINICAL HEART-BLOCK. 175 3ra£hia2 Fig. 126. Venous and arterial curves showing 2 : 1 heart-block in a patient. All the auricular contractions fall in ventricular diastole ; consequently the a waves do not materially vary in height. Fig. 127. Venous and arterial curves, showing prolongation of the a-c interval and several dropped beats in a patient. With three exceptions (the 3rd, 7th and 13th of the curve) the auricular contractions coincide with preceding ventricular contractions ; a falls back until it coincides with t) or c and is consequently exaggerated in height. Fig. 128. Venous and arterial curves showing 2 : 1 heart-block in a young woman. The auricular contractions to which there are no responses fall with the preceding ventricular contractions and produce very tall " waves. From the same case as Fig. 129. f V 1 Fig. 129. Venous and radial curves from the same patient, showing frequent dropped beats, and 1:1, 2:1, heart-block. Notice the increase in the amplitude of the a wave as the auricular systole falls further back into the preceding ventricular systole. Fig. 130. Venous and arterial curves exhibiting complete heart-block in a patient. The rhythms of auricle" and ventricle are regular but independent of each other. Where a and c coincide, a conspicuous summit appears. 176 CHAPTER XIV. .Tiiir.inmiLTUTrri Fig. 131. An electrocardiogram showing prolongation of the P-R interval in a girl. P falls before the termination of T of the preceding ventricular cycle. This is known by comparing this curve with the following, which is from the same patient. Time in thirtieths of a second. Fig. l.'Jii. Curve showing prolongation of the P-R interval and a failure of response to each third or fourth auricular impulse. Note the decreasing length of the ventricle cyc'les as the pause is approached ; and the relation of P' to T'' and compare the latter with the relation of P and T in Fig. 1.31. Time in thirtieths of a second. As 3£ Fig. 133. (X |.) Clinical curve showing very gradual prolongation of the P-R interval and two dropped beats. P falls back further and further upon T until it is actually buried in the preceding R (P"). The preceding PR interval (P*) measures OoT seconds. Time in thirtieths of a second. Fig. 134. Clinical electrocardiogram showing a period of 1 ; 1, 2:1 block, passing into 2 : I block Time in thirtieths of a second. CLINICAL HEART-BLOCK. 177 Fig. 135. Curves from the three leads in a case of complete heart-block. Time in fifths of a, second. condition, also termed "dissociation," two entirely separate rhythms are generated in the heart ; the one starts in the natural pacemaker and controls the auricle, the other starts in and controls the ventricle ; the former has a usual rate of 72 to the minute or thereabouts, the latter an approximate rate of 30 to the minute. The rhythms are each regular* and quite independent ; the systoles of auricle and ventricle fall with varying time-relations to each other (Figs. 130 * Not infrequently, in clinical examples, a disturbance of the auricular rhythm is manifested. As Wilson and Robinson (784) express it, the inter- auricular period during which the ventricular systole falls is shorter than those which follow it. This irregularity of the auricle is shown slightly, but quite definitely, in Fig. 136. The similar phenomenon referred to in describing 2 • 1 block is often more striking in degree. It is certain that the systole of the ventricle may influence impulse formation in the auricle, though the nature of the influence is unknown. To explain variation in the length of the inter-auricular periods in complete heart-block Wilson and Robinson cite the view adopted by Erlanger and Blackman (148), namely, that vagal tone is increased with each arterial pulse. But they point out that in rare cases the first auricular systole to follow the beginning of ventricular systole (the systole of the auricle which is super- imposed on that of the ventricle) is ectopic. That is so in the case which illustrates their paper and a similar event has been reported by Cohn and Fraser (65) in a case of complete and incomplete block. In such instances a direct mechanical stimulation of the auricle by the ventricular systole has been suggested as the cause. j j 4. j These curious influences of ventricular systole upon the auricle deserve more extended study : sometimes, in cases of actual complete block, the control of one auricular systole by each ventricular cycle influences the time-relation of all the auricular systoles to the ventricular systoles (for the second auricular systole is controlled by the first) and gives rise to a fictitious appearance of partial block. , . , j: i- 4.v,„ In considering the influences of ventricular systoles upon auricular impulse formation, the reader sho.ild also refer to some curious examples published by White and myself (490) (especially to Fig. 18, 19 and 20 of our paper). 178 CHAPTER XIV. and 135). Isolated auricular systoles are seen in the long diastoles. In electrocardiograms of complete heart-block the super-imposition of auricular and ventricular complexes is peculiarly clear (Fig. 109, page 159 and Fig. 135). The same events are evident in venous curves, except that when the auricular contraction falls within the confines of ventricular systole the veins of the neck expand to an exceptional extent (Fig. 130). In dissociation the auricular sounds are often audible (172, 196, 227, etc.) during the long diastoles (Fig. 137), and when auricular contractions coincide with ventricular ones, intensification of the first heart sound (234), or reduplication of the same sound or of the second sound (474), is frequently to be heard. In mitral stenosis a murmur may appear at each beat of the auricle which falls during ventricular diastole, an observation (54, 256) which has been made both in 2 : 1 and in complete block. The independent movements of the auricle may be seen in some cases upon a fluorescent screen (43, 641) ; they may be recorded at the apex beat. The pulsations of the auricles are also conducted along the arteries (102, 180, 504), and not infrequently appear upon radial tracings (Fig. 136). The mode of transmission to so distant a point is not clearly understood, but it is probable that the base of the aorta acts in much the same way as does an oesophageal sound, such as is used for recording the movements of the left auricle, and that the contractions of the auricular part of the heart which is wrapped around the aorta produce changes in systemic arterial pressure which are transmitted to a distance. The causes of clinical heart-block. When we inquire into the causes of clinical heart-block, two conclusions should stand out prominently in our minds ; (1) conduction between auricle and ventricle depends, as experiment has clearly taught us, upon the auriculo-ventricular bundle and upon this alone ; (2) all the forms of experimental heart-block which have been sufficiently investigated in the mammal have been shown to result from defects in this tract of tissue. These two conclusions are based upon relatively simple though exacting methods of investigation ; disease undertakes few simple experiments, its lesions are rarely localised, usually the damage is widespread. Our first conclusions apply to the lower orders of mammalia ; we may expect morbid anatomy to demonstrate in man, what has been discovered in the lower animals; yet morbid anatomy is still relatively an inexact science and we cannot expect its evidence to be so clear or so convincing as is experimental evidence. There is a further consideration ; the facts harvested in the laboratory are gathered by those methodically trained as investigators ; the opportunity of recording the effects of the disease, that is the relation of lesions of the A-V system to heart-block in the human heart, is offered to many who are inexperienced in precise methods both of recording the heart beat and of examining the heart microscopically. Thus, when we reflect, we shall be CLINICAL HEART-BLOCK 179 B'ig. 136. (X -|-) Simultaneous venous, radial and electrocardiographic curves in a case of dissociation. In the venous curve, frequent a waves and occasional c and v waves register the movements of auricle and ventricle; the last is a composite a and c wave. 'I'lie radial curve is of large amplitude and minute a waves are quite clearly inscribed upon it. The constant time-relation between these little waves, the a waves in the jugular and the P summits of the electrocardiogram, should be observed Time in fifths of a second. F^fsr- Fig. 137. (x ^■) Simultaneous electrocardiogram and heart sound record, the latter taken from the apex beat. From a case of complete heart-block. The sounds of the auricle are double (a} and a^) and occur a little after the beginning of each P summit. The reason of the double sound, which is unusual, is discussed in the original description {474). Time in thirtieths of a second. surprised, not that reports upon the human subject are sometimes in conflict with the testimony of experiment, but that, viewed generally, they are in such remarkable harmony with them and with each other. A mass of human material has been studied, some of it has been suitable for this purpose, much of it has been quite unsuitable ; it can scarcely be questioned that a large proportion of the reports are inadmissible, marred as they are by unavoidable 180 CHAPTER XIV. or avoidable imperfections. The task of the present generation, and of those who follow, would have been lighter had those reports been suppressed which did not contain simple and less questionable data. I emphasise these points of view, because a general survey must take account of them. Heart-block arises in the human subject from the same causes as in experiment. 1. From lesions of the conducting tract. — When persistent heart-block of a high grade has been demonstrated during life, lesions are usually found affecting either the A-V node or the bundle ; scores of cases* might be cited in support 6f this statement. As a general rule the lesions are prominent {11, 37). On the other hand, no instance of complete destruction of the bundle has been satisfactorily demonstrated, where impulses were known to have passed from auricle to ventricle a reasonably short time before death. The most exquisite example of a lesion created by disease is that of the child fully reported by Armstrong and Monckeberg (8) ; complete heart-block resulted in this patient from an endothelioma arising in, confined to, but totally destroying the hinder part of the A-V node. A single case, often cited as an exception, is that of Heineke, Miiller and Hosslein (248). Partial heart-block was recorded and six iveeks later complete destruction of the bundle by fibrosis, haemorrhage and calcification is said to have been found ; the haemorrhage certainly, the calcification very possibly, was of terminal origin. •)• But it is clear from reliable records that, excepting instances of complete transection, it is not possible to estimate microscopically the antecedent grade of functional impairment. There are cases on record in which gross lesions have been described, but where only the slighter grades of heart-block or merely the temporary defects of conduction have been observed (69, 220, 228). There are cases in which, with far slighter grades of apparent damage, dissociation has been complete and permanent (398). Further, slight changes in the conducting system are common in elderly subjects (37), though heart-block is comparatively rare ; they were found in 70 per cent, of all cases of heart affection examined by Sternberg (707) in a search through 72 hearts. When in instances of gross damage, a few fibres appear to be intact, it is impossible by examining them to estimate their previous functional efficiency ; and we know that undamaged fibres in small numbers are sufficient to serve as normal conductors. In examples of diffuse but slighter change, the func- tional capacity of the bundle as a whole is equally difficult to gauge when it is * A complete list and analysis of cases in which cases of heart-block have come to post- mortem examination up to the year 1910 was ]niblished in my book (447, page 100). ■j- Holtz and Krohn's case (,332) should not be cited, for neither the account of the microscopic anatomy, nor the interpretation of the curves is satisfactory. In Martin and Klotz's case (541), no curves were published, and there was no proof of conduction, neither can it be reasonably inferred. CLINICAL HEART-BLOCK. 181 submitted to microscopy^. The seeming discord between the grade of heart- block and the degree of tissue change is to be explained in some cases by the insufficiency of our histological methods,* in others by change in the lesion in the interval between the times when the functions of the system on the one hand and the structure of the tissues on the other were investigated. 2. The vagus and clinical heart-block. — That vagal impulses may produce heart-block in healthy human subjects, has been shown. If the nerve be pressed upon in the neck, prolongation of the P-B interval is witnessed (653) ; the reaction is abolished by atropinisation. Higher grades up to dissociation have been produced in heart cases by the same procedure (654). -f Heart-block can be induced through the vagus reflexly also by pressure on the eyeball ; excellent examples (controlled by atropine injections) have been published by Petzetakis (592). The effect of vagal compression is greater when a tendency to heart-block is pre-existent. Thus, when prolonged As-Vs intervals have been induced by the administration of digitalis, vagal compression may prevent the response of the ventricle for many auricular cycles (625). One of the most notable examples of heart-block arising from vagal stimulation is that published by Mackenzie (509) ; he observed that swallowing, an act which provokes an inhibitory cardiac reflex, produced a failure of ventricular responses in a patient, who previously exhibited a prolonged a-c interval. J A curve from this patient is reproduced in Fig. 138. A? id. Fig. 138. (Mackenzie, Brit. med. Journ., 1906, II, 1110.) A polygraphic tracing from a. case of partial heart-block (prolonged a-c interval), showing the effect of swallowing upon the heart's mechanism. After an interval of three heart-cycles, temporary 2 : 1 heart-block is produced. A clinical example of an increase in the degree of heart-block as a result of vagal inhibition. Heart-block conditioned in man by these interferences is necessarily temporary ; it is not possible to maintain heart-block of uniform grade for very long periods in experiment by vagal stimulation. It is open to question, therefore, if persistent heart-block in the human subject can ever * Before the discovery of Wallerian degeneration, the reason for many defects of the nervous system was unrecognised. t The conclusion that the left nerve has a more marked effect on conduction does not seem to me justified as yet. J A similar instance has been reported by Laslett (412). N 182 CHAPTER XIV. be attributed to nervous influence. Unquestionably, in most patients, the persistent heart-block of high grade is not of vagal origin for it is unaffected by full doses of atropine. Atropine usually has no effect on the ventricular rate in cases of dissociation. So far as I am aware this drug has never been known completely* to abolish heart-block unless such heart-block has been induced by the administration of drugs. That it may reduce the grade of a digitalis block or abolish it altogether is known {517, 625, 743). It may be that the vagus is accountable for transient heart-block in man. It may infrequently be responsible for temporary increases in the degree of block through reflex channels. It may be responsible for continued block when the vagal mechanism is stimulated by drugs. Nevertheless, it has not been shown that persistent heart-block of high grade is due to this cause ; our evidence is decidedly opposed to this conclusion, which has been formed with more persistence thai) wisdom. 3. Heart-hloch as a result of poisoning. — It has been shown (646) that heart-block occurs in children dying from anterior-poliomyelitis. The clinical details of the cases, and the curves, seem to establish this form of heart-block as asphyxial. The most remarkable instances of toxic heart-block in man are those following the administration of drugs of the digitalis group. Until recently it has been thought that conduction changes cannot be induced in the heart of man by therapeutic doses of digitalis except in cases where there is already a predisposition to this defect. But it has now been shown that normal subjects also exhibit the reaction {62, 66, 708, 769). We are indebted to Mackenzie, [504, 505, 517), for the first observations, and for the greater part of our knowledge of digitalis effects. He has shown that where there is a slight defect of conduction, that defect can readily be increased and rendered more prominent — by full doses of digitalis. This conclusion has been abundantly confirmed by other workers {95, 306). In such patients, high grades of block, even dissociation, may develop. Similar effects are observed with strophanthus and squills {788). Digitalis heart-block seems in some patients to be of vagal origin, for it may be abolished by full doses of atropine ; in other cases its action appears to be directly upon the junctional tissues {106, 517). * An observation of Barringer (22) upon a case of temporary dissociation seems to demonstrate that the heart-block was in a measure vagal in origin, but his curves do not demonstrate that it arose entirely from this cause. CLINICAL HEART-BLOCK. NOTE ON DISSOC CATION OP THE TWO AURICLES. In 1906 Wenckebach (760) figured and described a series of curious curves which he interpreted as resulting from dissociation of the two auricles. Wenckebach's explanation is inaceeptable to-day in that such dissociation is unlmown in experiment and in that it is opposed to the now widely supported view, that physiologically, the muscle of the two auricles is to be regarded as one undivided mass. But an alternative explanation of Wenckebach's curves still fails us ; no similar mechanism has since been described to my knowledge. The importance of the curves lies perhaps in their strangeness. NOTE OS DISSOCIATION OF THE TWO VENTEICLES (HBMISrSTOLE). From time to time interpretations of clinical curves have been put forward in which independent contractions of the ventricles are supposed to be evidenced. For the most part these curves have been polygraphic. Notable examples of such curves are those published by Mackenzie (501, Fig. 309, 310, 313 and 314 of that book), and those published by Hewlett (308). Mackenzie's curve, Fig. 309, almost certainly represents successive extrasystoles, his Fig. 310 and the curves published by Hewlett would now be regarded generally, I think, as exhibiting unusually large " stasis " and b waves in the long diastoles of the venous curves. Mackenzie's Figs. 313 and 314 still lack an explanation. The same idea of hemisystole was used in explaining the electrocardiographic curves of ventricular extrasystole by Kraus and Nicolai (391), but this explanation is now recognised as a, mistaken one. There is, in fact, no clinical evidence of " hemisystole " ; there is nothing to make us believe that one human ventricle may beat without the other, continuously or occasionally. A continuous beating of one ventricle while the other stands quiescent would be a condition clearly incompatible with a continued circulation. Even in the dying heart it is most difficult to conceive of an occasional beat confined strictly to one ventricle, for the two chambers have a continuous musculature and in large measure a common blood supply. The question of hemisystole has been contested on the basis of experimental work. Opinion is decidedly against its occurrence. In the dying heart one ventricle may beat while on superficial inspection the other chamber may appear still, but closer inspection or a finer method of recording demonstrates movement in the second chamber also. While it cannot be said at present that there is any real foundation for the hypothesis, it can be said that our knowledge of the anatomy and physiology of the ventricles is strongly opposed to it. The following additional papers, which deal with this question at greater length, may be consulted (251, 252, 604, 606 and 623). N 2 Chapter XY. NEW RHYTHM CENTRES (ATRIO-VENTRICULAR RHYTHM). In 1903, Engelmann (135), while recording the movements of auricle and ventricle in the frog's heart, noticed in certain animals that when the region of the sinus is separated from the rest of the heart by tying (1st Stannius ligature) the auricles and ventricles subsequently beat simultaneously. He attributed this newly developed rhythm to impulses arising in the A-V ring. A year later a similar mechanism following vagal stimulation in the tortoise and rabbit was described (492), and attributed to a similar cause. Since the discovery of the A-V node, the impulses yielding simultaneous systole of auricle and ventricle have by general consent been supposed to arise in this structure ; the suggestion first came from the papers of Hering and Rihl {303), and Mackenzie {512, 513), describing certain isolated and premature beats discovered in patients.* It has been shown that atrio-ventricular rhythm may be induced in the mammalian heart by a number of different procedures. The most constant and striking method is by destroying the S-A node or by cooling the same structure {71, 204, 493). It is also to be produced by warming the A-V node. Atrio-ventricular rhythm may also follow interference with the nerves of the heart, especially combined stimulation of the right vagus and left sympathetic nerves {665). The change in the heart's mechanism is well illustrated by experiments in which the region of the 8-A node is cooled. A leaden tube is laid along the sulcus terminalis of the auricle, and iced water is passed through it. The heart rate is retarded almost immediately and its action changes (Fig. 139). The P summits of the natural electrocardiogram disappear and simultaneous action of the auricle and ventricle becomes established. This is seen in •the electro-cardiogram, but is more clearly deciphered in the muscle curves. The last three ventricular complexes of Fig. 139 are similar to the first complexes of the curve ; the corresponding beats are still supraventricular in origin ; the auricular complexes are of altered form and * Mackenzie, in the same series of papers (511 513), also supposed that what we now Icnow to be auricular fibrillation arises in this node and provisionally termed this condition " nodal rhythm." This hypothesis was discarded in 1910, when I published a paper proving the nature pf auricular fibrillation and recording an actual instance of nodal rhythm (434). AT RIO-VEN T RI GV LAR RHYTHM 185 Fig. 139. (Heart, 1914, V, 247, Fig. 1.) Myocardiographic curves (A = auricle; F = ventricle) and electrocardiogram from lead II showing the effect of applying cold to the sulcus terminalis in » dog. With the reduction of temperature (see signal) the S-A rhythm slows ; at the fifth cycle there is escape of the A-V node, though the auricle, being already in contrac- tion, does not respond. Subsequent cycles show the fully established A-V rhythm; the auricle is represented in the electrocardiogram by a minute dip preceding the upstroke of R. The A-V and P-R intervals are given in decimals of a second. Time in fifths of a second. Fig. 140. (Heart, 1914, V, 247, Fig. 3.) Similar curves from another animal, showing the awakening of A-V rhythm on the application of cold. The uppermost curve is from the veins of the neck and shows the exaggerated^ wave accompanying the simultaneous contractions of auricle and ventricle. The P-R intervals are given in decimals of a second. (The intervals have been entered incorrectly on this curve, they should all be moved one cycle to the right.) Time in twenty-fifths of a second. 186 CHAPTER XV. lie almost completely buried in those of the ventricle ; the initial descent of P, preceding R by an interval of 0027 of a second, is alone visible. An example of a similar change is shewn in Fig. 140. In this figure a volume curve of the veins of the neck is added : when the auricle and ventricle beat synchronously, the normal a and c waves are replaced by a tall wave 1 ; this exaggerated wave is due to the auricle forcing its contents into the veins of the neck, since during the systole the tricuspid valve is closed. When A-V rhythm develops in the dog's heart, there may be, as in the present examples, a conspicuous reduction of the As-Vs interval from the normal of 0-08-0-lOof a secondto002or 0-03 of a second. The reduction may be greater, and the interval may vanish altogether or become of opposite sign ; that is to say, the ventricle may beat a little before the auricle. On the other hand, the reduction may be less (Fig. 141). Be the reduction sUght or great, the new rhythm arises in the A-V node according to present day conceptions. The meaning of the variations in As-Vs intervals, which occur not only from animal to animal, but even in the same animal from time to time (Fig. 140 and 141 are from the same animal), will be discussed presently. Of peculiar interest is the shape of the auricular representative in A-V rhythm. Though inconstant in form it usually presents the inversion of Fig. 139 ; where it is buried in the corresponding ventricular complex, its form is not clearly seen, but it may be unveiled by breaking the A-V bundle and producing A-V dissociation {547). The auricle then responds to A^V nodal impulses, the ventricle responds more slowly to its inherent impulses. The usual form of the auricular complex, when A-V rhythm prevails, is that of Fig. 142 ; it starts as a rapid downward deflection of considerable extent and has a slower upstroke.* Less commonly, the complex is not so extensive and consists of a few minute deflections. These variations in the form of the auricular complex are scarcely to be explained as resulting from variation in the architecture of the auricles in different animals, for they are more distinct than are variations in the normal auricular complex in those of the same species. Possibly the excitation wave, in travelling backwards in the auricle, follows a less constant and ordered route than in its forward propagation. For one or more cycles at the transition from S-A to ^-F rhythm, the auricle may respond to an 8-A impulse, while the contractions of the ventricle are hastened by the development of A-V nodal impulses to which the last named chamber responds. Evidently the transitions of Fig. 139 to 141 are of this kind. Thus in the fifth cycle of Fig. 139 the auricle is represented by a complex which is like those preceding it, but the P-i? interval is reduced because the first A-V impulse culminates and excites the ventricle before response to the natural impulse is due. The auricle does not respond to the * According to Ganter and Zahn (206), this form of complex is always associated with an origin of the beat from the upper part of the A-V node ; this conclusion is, I think, questionable. AT RIO-VEN T RI CU LAR RHYTHM 187 Fig. Ul. (Heart, 1914, V, 247, Fig. 2.) Similar curves from the same animal as in Fig. 140, showing slowing of the heart and alteration of the pacemaker subsequent to the application of cold. P becomes inverted and the P-R interval is slightly reduced. Note the change in the shape of P at the transition. Intervals in decimals of a second. Time in twenty-fifths of a second. Fig. 142. (Meakins. Heart, 1914-15, V, 281, Fig. 9.) (X f.) An electrocardiogram from lead // in a dog. A - V rhythm was induced by cooling the S-A node, and during the maintenance of this rhythm the ^4-1' bundle was broken by clamping. In the curve dissociation is shown between the auricular and ventricvilar contractions, and the unusual form of the auricular complexes is plainly to be seen. Time in fifths of a second. A-V nodal impulse because when this impulse arrives the auricle is already in contraction and consequently refractory. A transition of a somewhat different and more complex type is shown in Fig. 141. Here the first altered auricular complex (the sixth in the curve) is intermediate in type between that which precedes and succeeds it, In instances of this kind it is probable that the S-A and A-V impulses entered the auricular substance at the same moment and that for one cycle the response was partially to one impulse and partially to the other. 188 CHAPTER XV . Evidence that these new rhythms arise in the A-V node. That the A-V node is responsible for certain rhythms in which the auricle and ventricle contract simultaneously is now generally accepted. The node is an integral part of the system of tissue joining the two chambers, and impulses arising in the node may be supposed to spread simultaneously to these chambers. The two nodes {8- A and A-V) are alike structurally and are probably morphological homologues ; the S-A node is known to possess a rhythmic power- in high degree, the A-V node is expected from its structure to possess a similar power. The electrocardiograph tells us that the ventricular beats, while A-V rhythm is in progress, arise above the division of the A-V bundle (Fig. 139- 141), for the impulses are distributed to the ventricle in normal fashion. That they are supraventricular is established by the shape of the ventricular complexes and by the following experiment. If during A-V rhythm the septum is cooled on the ventricular side of the node (797), or if the bundle is clamped [547), the ventricle fails in its responses while the auricle continues to beat. The inversion of the auricular complex in A-V rhythm indicates an upward passage of the excitation wave as opposed to its usual downward course ; this reversal of propagation is also indicated if we lead from two direct contacts placed upon the sulcus or parallel and near to it.* When S-A rhythm prevails it is the upper contact which first becomes negative, during A-V rhythm it is the lower one, as Wybauw {795) first demonstrated, and as I have myself often found. But perhaps the most convincing evidence is that A-V rhythm is retarded by coohng,-)- and accelerated by heating, the region of the node [797). It has also been stated (550) that negativity is first developed in this region of the heart while A-V rhythm is present. J The conclusion that the rhythm arises in the A-V node is not only supported by these observations, but it is in harmony with so many other conclusions, that it is scarcely to be questioned. Variation in "A-V " rhythm intervals. The conclusion that impulses which arise in the A-V node may be responsible for heart beats in which auricle and ventricle beat exactly simultaneously, or for heart beats in which auricle or ventricle has a short * This method, though its results are not often misleading when the two contacts are in the immediate neighbourhood of the rhythm -producing centre, or in the direct line along which the excitation wave is propagated, is nevertheless open to some -criticism. The curves are confused by extrinsic deflections. t Zahn's statement that he cools particular and known regions of the node is. scarcely convincing, but there seems no reason to doubt the more general conclusion. t This evidence is open to criticism, for it is not possible to lay an electrode directly upon the A-V node, and even though it can be placed inside the heart near the node, the precise region examined is difficult to ascertain subsequently. AT RIO-V E N T RI CV LAR RHYTHM. 189 lead, was first reached by considering the time intervals which separate the contractions of auricle and ventricle. We know that conduction in the auricle is rapid, we know that conduction in the Purkinje substance of the ventricle is still more rapid ; considering a normal As-Vs interval of 0-10 of a second, some 0-03 of a second may be allowed for conduction through the auricle, a smaller interval may be allowed for conduction from bundle to ventricle. Somewhere in the path of conduction there is an unexplained delay of from 0-04 or 0-05 of a second. It is generally accepted that this delay occurs in the fine fibres of the A-V node ; in Chapter VII the reasons for believing that slowness of conduction is associated with smallness of fibre have been given. One worker believes that he has demonstrated delay in this structure by direct experiment (286), but the reasons for his conclusion cannot be accepted unreservedly. We have no proof, yet we may assume, that the chief delay is in this structure. Thus it has been concluded (Chapter XIII) that it is the depressed function of this node which exaggerates the delay in A-V block, caused by asphyxia or vagal stimulation. It is to be inferred that the point at which the natural delay occurs is also that at which the delay is most susceptible to exaggeration. The strongest ground for our hypothesis is that it harmonises with many distinct observa- tions, and is, so far as I know, discordant with none. It is further argued that the variations in length of the As- Vs or Vs-As interval as the case may be, is due to variations in the level of the node from which the impulses arise ; and for this last assumption there are a number of arguments and some direct evidence. If the As-Vs interval is but slightly reduced, the portion of the node towards the auricle is considered the rhythm centre (Fig. 141) ; if the beat of the auricle and ventricle is simultaneous (Fig. 140) or a Vs-As interval is developed, then the portion of the node towards the ventricle is said to be active. A powerful argument in favour of this thesis is supphed by those agencies which produce disturbances of A-V conduction by an action concentrated upon the region of the node. Thus if A-V rhythm prevails and the reduction of the As-Vs interval is not great, vagal stimulation- sufficient in degree to produce heart-block while the heart is responding to the S-A node — produces the natural form of heart-block ; that is to say, the As-Vs interval increases and the ventricle occasionally fails to respond. But if the As-Vs interval is conspicuously reduced, or if a Vs-As prevails, then a similar stimulation of the vagus produces a reversed block ; the As-Vs interval decreases (a Vs-As interval increases) and auricular contractions are missed {466). To explain these phenomena I use the accompanying diagrams. In Fig. 143 the impulse is supposed to arise in the upper, in Fig. 144 in the lower, part of the node {N) and to spread to the auricle {A) and simultaneously to the ventricle (7) through the bundle (B). Vagal impulses acting on the A-V node cause a lowering of conductivity in this structure, and according as the obstruction is more on the auricular, or more on the ventricular, side of the rhythm centre, so the beat of the auricle or of the ventricle is delayed. As the vagal action 190 CHAPTER XV Fig 143. A diagram illustrating the disturbances produced when, during A - V rhythm originating in the upper part of the node, the vagus is stimulated. The increased vagal tone produces heart-block of the usual kind ; the As-Vs intervals widen and eventually the ventricle fails to respond. A ~| ^ / / ^ .^ .^ .. / B 1 1 1 V Fig. 144. A diagram illustrating the disturbances produced when, during A-V rhythm originating in the lower part of the node, the vagus is stimulated. The increased vagal tone produces a reversed heart-block; the As-Vs interval shortens (this is not, but should be, shown), a Vs-As interval develops, and eventually the auricle fails to respond. in a single animal is constant, so it must be assumed that it is the position of the rhythm centre which has changed. This modified action of the vagus is not an isolated example ; similar phenomena are witnessed when the changes in conduction intervals are conditioned by premature contractions (490). I hesitate to call up the observations of Zahn (797, 798), and those of Meek and Eyster (550), in support of the hypothesis that rhythms are developed at different levels of the node. The former states that, when the ^5- Fs interval is slightly reduced, temperature changes over the auricular portion of the node affect the rate of the rhythm, and that, when the reduction of the As-Vs interval is great, temperature changes affect the. ventricular portion of the node. The latter assert that the point of primary negativity hes nearer to the coronary sinus, or nearer the tricuspid valve, when the longer or shorter intervals, respectively, prevail . I hesitate because this exact location of events in a small and buried structure appears to me precarious, and also because the method of leading in Meek and Eyster's experiments is open to certain objections. For the reasons given I beheve, nevertheless, that the conclusions at which they arrive are the right ones. AT RIO-VEN T RICU LAR RHYTHM. 191 Influence of nerve stimulation upon " A-V " rhythm. Many observations have been undertaken with a view to ascertain the innervation of the A- V node. The effects of left and right vagal stimulation upon conduction from auricle to ventricle has been dealt with in a previous chapter (page 164), but these should be distinguished from those now discussed. It is generally accepted that the right vagus has a greater retarding influence upon the S-A rhythm than has the left. It is by no means so clear that the left vagus has a greater retarding influence upon A-V rhythm than has the right. Quite exceptionally in dogs it is sufficient to stimulate the right vagus to transfer the pacemaker from the S-A to the A-V node ; much more often a similar stimulation during A-V rhythm causes an opposite transference. The influence of both vagi over rhythm production in the A-V node is powerful ; but the right nerve, in my experience, is predominant in this respect in most dogs (466). The left vagus seems to act more powerfully upon the A-V node than upon the S-A node in most experiments, but so does the right nerve ;* both the degree of vagal control and its distribution is variable from animal to animal. The left sympathetic nerve seems to accelerate rhythm production in the A-V node powerfully and to a greater extent than the right nerve (665) ; so when the right vagus, which holds the S-A rhythm in abeyance, is stimulated simultaneously with the left accelerator, A-V rhythm displays itself; and this is the more explicable since the left accelerator is said to have relatively little influence upon S-A rhythm {403). Sometimes, indeed, isolated stimulation of the left sympathetic suffices to induce an A-V rhythm (403). The reactions to stimulation are so complex and the experiences of different observers, or of the same observer in different experiments, are so variable that a further analysis of the relative influences of the right and left nerves upon the two nodes cannot here be attempted in detail. Suffice it to say that the A-V node is richly supplied by inhibitors and accelerators ; and that while the right vagus has an unquestionable predominance over the left nerve so far as the right or S-A node is concerned, the left sympathetic has a predominance over the right nerve so far as the left or ^-F node is concerned. To this extent the nerve supply appears to be homolateral. But the crossed effects, the action of left inhibitor and left accelerator upon the S-A node, and of the right inhibitor and right accelerator upon the * One of my original reasons for stating that the vagus has a greater influence upon the A - V node than upon the S-A node in certain dogs was the observation that in these vagal stimulation would convert an A-V rhythm (induced by cooling the iS-^ node) to a S-A rhythm. As Schlomovitz and his fellow workers (689), have justly remarked, this reasoning is inadmis.sible, in that the cooling of the S-A node depresses the influence of the vagus upon it. Nevertheless, and in view of the profound slowing effect of the vagus upon A-V rhythm itself, I retain my former conclusion, being unable to accept the view which these authors express that the vagus has a lessened effect on nodal tissues, as these are traced downwards through the heart in the dog. Were that the case simple vagal stimulation, especially right vagal stimulation, would usually produce ^-F rhythm ; that is not so. 192 CHAPTER XV. A-V node are also to be obtained, and certain of these crossed effects seem to be as powerful as those which are uncrossed. The original papers should be consulted by those who desire fuller and more exact detail of the numerous observations which have been made upon the nerve supply of the two nodes {6, 207, 302, 403, 466, 550, 665, 667). Clinical examples of A-V rhythm. Atrio-ventricular rhythm was first described and figured in man by Belski (30) in 1909. Shortly afterwards I published {434) an example of a rapid rhythm* in which auricle and ventricle contracted simultaneously, 7.7-^—^ go 1 y-o ^ /•» ^ fa —J ys Fig. 145. An example of A-V rhythm, determined by polygraph in the human subject. The A-V rhythm prevails for five cycles, and tall waves ^'. due to simultaneous contraction of auricle and ventricle, are seen in the curve. At the sixth cycle the S-A rhythm escapes and controls the movements of the heart to the end of the curve. At first the heart rate is slow; rate 35 per minute ; the escape of the S-A node has been due to a quickening of its rhythm. The lengths of the cycles are given in fifths of a second. From the same case as Fie. 146 to 148. Fig. 146. A curve from the same patient showing the onset of A-V rhythm as the heart slows, and the re-establishment of iS-^ rhythm accompanying an increase of its rate. and was able to present both polygraphia and electrocardiographic curves. At later dates very many examples have been placed on record (201, 460, 544, 636, 766, 780, 782). The appearances of this rhythm in polygraphic curves and in electrocardiograms are identical with those described in the experimental section of this chapter. In the venous curves the striking chai-acteristic is the high wave produced by simultaneous contraction of auricle and ventricle. The waves are most exaggerated when the systoles of auricle and ventricle begin exactly together. This A-V rhythm in patients discloses itself when the 8-A rhythm is depressed. In * A rhythm which should not now be termed atrio-ventricular rhythm, but atrio-ventricular tachycardia. AT RIO-VENT R re U LAR RHYTHM 193 WMiiiMMiHMiillilMiiMiiiiiiiiiliiiii Fig. 147 and 148. ( x f ■) Two electrocardiograms (lead //) from the same patient as Figs. 145 and 146, showing the onset of A-V rhythm, its continuation and the first auricular response to the returning S-A rhythm. During the phase of established A-V rhythm, the auricular complex P is buried in the corresponding ventricular systole and is not visible. Time in thirtieths of a second. MriiMMriaAM*ailiii ■ --.-. .^ .^^^^-^^^^ Fig. 149. (x ^. ) An electrocardiogram from lead III, showing three responses of the heart to the >S-^ rhythm, an extrasystole, and the first three responses of an established A - V rhythm. When the A-V rhythm appears, the P-R interval shortens and P becomes inverted. In this example the A-V rhythm is from a high level of the node, higher than in the case in Fig. 148, and the systoles of the auricle and ventricle do not begin simultaneously ; the reduction of the P-R interval is slight. Time in thirtieths of a second. ^iri^^^UMMdOIMaiB^^ia^ttiM^BHiM img^gtt^mtmttmmimiaBium^ Fig. 150. { X A.) This figure is from the same patient, and displays the same A-V rhythm as does Fig. 149. Time in thirtieths of a second. most examples there is a waxing and a waning in the rate of the S-A rhythm, and as the heart slows the S-A node loses control and the ^-Fnode captures it ; while with its quickening the S-A rhythm once more asserts itself (Fig. 145 and 146). Interesting examples of ^-F rhythm occurring during the vagal slowing accompanying expiration have been published by Wilson (781) and others (193). The former writer has shown also that normal subjects often develop A-V rhythm as they are passing under the influence of atropine, the new rhythm appearing spontaneously or under vagal pressure. He suggests that in such cases the vagal endings in the A-V node are released first by a selective action of the drug. Escape of the node is not uncommon as a 194 CHAPTER XV. result of simple vagal stimulation in man. It seems clear that such temporary dislocations of the pacemaker are due to reduced activity of the 8-A node ; but all instances are not of this kind, it would seem, for full atropinisation will not always re-establish the control of the S-A node {193). In these the new rhythm is probably due to heightened activity of the A-V node. Clinical electrocardiograms of A-V rhythm are shown in Fig. 147 and 148. Fig. 147 shows the onset of A-V rhythm ; the auricular responses to the 8-A rhythm are undisturbed, but the last two ventricular beats of the curve are responses to the A-V node. Later (Fig. 148) simultaneous contraction of, auricle and ventricle becomes established ; this is not clear in the electrocardiogram, for the auricular complex is hidden by the ventricular. At the end of the same curve the 8-A rhythm is appearing once again. Another example of clinical A-V rhythm, taken from a patient who displayed it almost constantly, is seen in Fig. 149 and 150. In the former of these curves the S-A rhythm is shown for three cycles ; an extrastole of auricular origin interrupts the regular action of the heart and this disturbance is followed by a new rhythm in which the P-R interval is shortened and P is inverted in the electrocardiogram. A similar rhythm is shown in Fig. 150, which is representative of the curves usually taken from this patient. In this case the ^-F rhythm comes from a high level of the node. Admirable examples of A-V rhythm in which ventricular contraction preceded the auricular have recently been published by Mathewson [544). Fig. 151. (Heart, 1914, V, 247, Fig. 5.) A clinical curve, for comparison with Fig. 141. The S- A rhythm is slowing and the A-V node gains control. The transition in the shape of auricular complexes is a gradual one ; this is due to the response of the auricular muscle to both centres for two or three cycles. Time in fifths of a second. Curious transitions from 8-A to ^-F rhythm in patients are sometimes recorded electrocardiographically. An example is shown in Fig. 151, and it should be compared with Fig. 141, page 187. Three cycles of S-A rhythm are shown at the beginning of the curve, and a single cycle of A-V rhythm (the P-R interval is shortened and P is inverted) is seen at the end of the curve. Between these are three cycles in each of which the ventricle responds to the ^-F node, but in which the auricle responds simultaneously to impulses entering it from both nodes. During the first of the three cycles the main mass of the auricle responds to the 8-A node and P is but little AT RI 0-VEN T RICU LA R RHYTHM. 195 changed, during the next cycle a greater portion of the auricle responds to the A-V node and in the third cycle a greater portion still. These events are reflected by the shape of P as it is traced from one cycle to the next. That clinical A-V rhythm is under vagal control has been shown by White {766). In his case of established A-V rhythm the rate responded to exercise, forced respiration, vagal pressure and atropine very readily. The same worker noticed that under vagal pressure the Vs-As interval lengthened, which conforms to the experimental findings. Chapter XV J. NEW RHYTHM CENTRES (IDIO-VENTRICULAR RHYTHM, ETC.). When the auricles and ventricles are dissociated, as by direct injuries of the A- V bundle, the ventricle assumes a rhythm of its own. This has been termed the idio-ventricular rhythm. It is dormant when the heart is beating naturally, in certain respects resembling, and in certain respects differing from, the atrio -ventricular rhythm. It may be stated as a general law that if several heart centres are simultaneously active, the heart, as a whole, will be dominated by the centre which develops its rhythmic impulses most rapidly. It is for this reason that the S-A node controls the naturally beating heart, for the rate at which impulses are built up in this node is greater than in any other centre in the organ. If a rhythm, more rapid than the natural, is excited by stimulating a heart, the natural rhythm becomes submerged, to reappear immediately stimulation ceases. This submerging of the natural rhythm is explained by supposing that a contraction wave which passes over the heart discharges any partially developed impulses in the musculature. So, if rapid rhythmic contractions are forced by stimulation, natural impulses developing at a slower rate never reach maturity, and therefore fail to provoke contractions while the artificial rhythm is in progress. Similarly, an A-V rhythm is not displayed by a normal heart because it is of a lower order than the 8-A rhythm ; its rate is less. This fact sufficiently explains the normal subjection of the A-V node. The still lower rate at which the idio-ventricular centre produces its rhythm would be sufficient to account for the usual subserviency of this centre both to the S-A and A-V nodes. Destroy the S-A node and the A- V node promptly secures control ; but the heart beats at a slower rate. Destroy the junction between A-V node and ventricle and the ventricle beats more slowly still and in response to intrinsic impulses (idio-ventricular rhythm). But in the case of this idio-ventricular rhythm there appears to be an additional factor. The rate of this rhythm while it is developing is influenced by the rate at which the ventricle was beating previously. Sever the A-V bundle abruptly and the ventricle remains quiescent for a considerable period — it may be a minute or more — and the new rhythm develops slowly, increasing in rate as it proceeds. It was for this reason that Gaskell (215), observing this phenomenon in the frog's heart, termed the rhythm a IDIO-VENT RICULAR RHYTHM, ETC. 197 "rhythm of development." Erlanger {150) noted that if the bundle is compressed gradually, thereby reducing the number of ventricular responses gradually, the idio-ventricular rhythm develops its full rate more speedily. The same feature was displayed by these workers in another fashion. If the rhythm has fully developed, and the ventricle is stimulated and made to beat rapidly, then, at the cessation of stimulation, the ventricle becomes quiescent and the development of the returning rhythm is gradual. But the length of the pause, and the subsequent rate of quickening, is controlled by the rate of preceding stimulation and by its duration ; the centre being less active after rapid or long stimulation of the muscle. Cushny {90), who has confirmed these facts, explains them by supposing that the new centre is fatigued by the receipt of extraneous contraction waves. A similar fatigue is not observed in the case of the S-A node or in the case of the A-V node providing these centres are well nourished.* E j 1 =E = =?^ i eSi \ 1 1 1 = M 1 1 3 1 ^ ^ 3 1 1 i^ i i 1 g 1 1 ^ i 1 -(* p[ I ! 1 1 1 = i i = = 4= :z/im i Cl^ iiS 1 S ^ ^ = ^ iSfe = |K = = 1 1 r=rt ^ = i = a= i Fig. 152. (Heart, 1913-14, V, 335, Fig. 16.) An electrocardiogram from a dog in which the sulcus was cooled and an ^-F rhythm developed. The auricle was stimulated by successive induction shocks at a rate sufficiently rapid to submerge the A-V rhythm. At the cessation of stimulation the A-V rhythm at once took control (4th cycle of the curve), but the rate of the A-V rhythm declined until it reached the rate shown before stimulation. The figures on the curve are the intervals in decimal points of a second ; inter-auricular intervals above, inter-ventricular intervals below, and As-Vs intervals written vertically. Time in fifths of a second. Origin of the idio-ventricular rhythm. The new rhythm which controls the movements of the dissociated ventricle appears to have its origin in the uninjured portion of the bundle directly below the seat of injury. That the bundle below the injury continues its function is suggested by the maintenance of its structure ; it does not degenerate {374), an observation which has been made both * On the other hand, the S-A node and the ^ -F node (Fig. 152) often appear to be stimulated rather than fatigued [490), and in this respect therefore, there is a contrast between these centres and that responsible for the idio-ventricular rhythm which develops after section of the bundle. Exceptionally, and when there is reason to believe that the heart is poorly nourished and in a hypodynamic condition, the S-A node or the A-V node may exhibit similar fatigue to that of the idio-ventricular centre (see Fig. 210, page 244), O 198 CHAPTER XVI. clinically and experimentally {148). If complete heart-block is obtained experimentally by cooling the region of the A-V node, and if subsequent to the block being established cooling is continued, standstill of the ventricle is said to occur {41). It is to be assumed that in this experiment, the new impulse centre is so near to the region cooled as eventually to be involved in it. Satisfactory evidence for the belief that the new rhythm arises in the bundle is derived from electrocardiography. The natural form of ventricular complex is preserved (see Fig. 114, page 163, and Fig. 135, page 177) ; the complex is the same before and after the injury, indicating that in the latter circum- stance the beat is still of supraventricular origin. It arises between the lesion and the division of the bundle into its two branches, and the excitation wave is distributed to the ventricle along the accustomed channels. If, in addition to the crush of the bundle, a main bundle branch is cut, the resultant ventricular curve is similar in every respect to that obtaining where the second lesion alone has been produced {475). Thus the lesion in the division of the bundle interferes with the spread of the excitation wave in precisely the same fashion, whether the wave originates in the auricle, or is derived from the idio-ventricular impulse centre. Clearly, therefore, the latter arises above the bundle division. The ventricular rhythm in dissociation, produced, not by lesions, but by the action of other agencies (for example, adrenalin, asphyxia, digitalis, anaphylaxis, etc.), is also known, from the form of the electrocardiogram, to have a supraventricular origin. But in these instances, it is not so clear that the bundle is responsible. Thus in the complete heart-block of asphyxia, a higher origin in the A-V node has been determined. If a cat is asphyxiated while A-V rhythm is in progress, the reversed form of heart-block appears ; the auricle, and not the ventricle, fails to respond. When the block becomes complete, the auricle no longer beats, but the ventricular rhythm continues undisturbed and at its previous rate {482). From this we may conclude that the centre forming the ventricular impulses remains unchanged and is situated, as at the beginning of the experiment, in the A-V node. The region of block in asphyxia is above this rhythm centre. To sum up, it may be said that the rhythm governing the ventricle probably arises immediately below that region of the junctional tissues upon which the injury falls, be the injury physical, chemical or nervous or be it situate in the upper or lower reaches of the tract. The influence of the vagi upon idio-ventricular rhythm is considered in Chapter XXXI. Escape of new centres ; ectopic impulses. It is convenient to describe the response of the heart, in part or in whole, to impulses discharged by a new centre as an " escape " of that centre, when the rhythm of the new ceatre is of the same genetic type as the physiological rhythm of the S-A node. Such escape is conditioned if the IDIO-VENT RI CULAR RHYTHM, ETC. 199 rate of impulse formation in the new centre exceeds the rate at which impulses are received in that centre from outside. Usually it is due to a change in the relative activities of the old centre and the new, that of the old centre being depressed or that of the new centre being exalted {284) ; and of these alternatives, the former is much more common than the latter.* Escape may happen also (as in vl-F dissociation) when the path of conduction from the old centre to new is interrupted. I i ^ig. X- S.P. y- 153. A diagram illustrating ventricular escape, when the sinus rhythm is sufficiently retarded and when the conditions predispose to such escape. It is supposed that stimulus production (S.P. ) occurs at a, constant rate in the ventricular centre, but that as a rule the pauses are of insufficient duration to allow the stimulus material to accumulate to the critical point (the line x, x') at which the growing impulses are discharged. At one point in the figure a spontaneous ventricular beat (marked with an asterisk) is represented, where the pause is longest. Fig. 154. (x §.) A polygraph curve illustrating single escape of the A.V. node in a patient who exhibited persistently slow heart rate. At the escape, auricle and ventricle contract together. The escape occurs at the end of the longest pulse beat. It is known to be an escape of this node because the tracing is from the same case as are Fig. 145 and 146. The simpler forms of escape are those in which the old heart rhythm is submerged by a rhythm emanating from a new centre. The examples cited so far have been of this kind. But escape may be a much more transient event ; there may be escape for a single heart cycle only. The * Examples of ventricular escape in which the rate of the escaping ventricle rises above that of the auricle are rare ; in such oases the ventricle and auricle beat for the most part independently, but the rate of the ventricle is the faster (see White and Heard and ColweU) (2ii, 767). This disorder has been seen in patients under the influence of digitalis, O 2 200 CHAPTER XVI. UP.Rt^ fl/? r ^1 I-? >7-^ >|P 7- "/| j 3£::^ ^~il ^ - — y pr^j?^^^^ -^- -— " g!_" iii»i^ 1 |H — ^ ^ ^^ _ ^ ^ - V .^^-^ .--^ .^^^ .^-^ ^^-^ Fig. 155. ( X Y(j.) A single escape of the ventricle, conditioned by slowing of the S-A rhythm and shown electrocardiographically. Each beat of the auricle is awakened by natural (S-A) im- pulses. The fourth ventricular contraction is not in response to the auricle but to a new centre. The diagram below the curve shows the time-relations and origins of the auricular and ventricular systoles. The zig-zag line (S-P) represents the building-up of the impulses which from time to time are responsible for ventricular escape. Fig. 156. (x j;'^.) A similar curve from the same patient, showing repeated escape of the ventricle. Time in both curves in thirtieths of a second. Fig. 157. (X 1^.) Venous, radial and electrocardiographic curves taken simultaneously from a patient and showing heart-block. The As-Ys interval is prolonged and at one point the ven- tricle fails to respond, a long pause ensues and during this pause the ventricle escapes. In this cycle auricular and ventricular systoles fall |5artly together ; the corresponding P-R interval is yery short aricj an exaggerated wave " appears in the neck. Time in fifths of a second. IDIO-VENT Rl CVLAR RHYTHM, ETC. 201 to-6 ,o-r ^■^ 'o-y 'o-y /o-y \ Of ^J Fig. 158. [Quart. Journ. Med., 1908-09. II, 361, Fig. Sandg.) A polygraph curve illustrating escape of the ventricle at alternate beats. The ventricle escapes whenever the opportunity is afforded to it. Partial heart-block is present and the ventricle is wavering between a 2 : 1 and 3 : 1 response. The blacls triangles represent impaired conduction in the bundle, the depth of the triangle at any point represents the measure of this impairment. After the first response of the ventricle, conduction is so far impaired that two auricular impulses (A^ and A^) fail to reach the ventricle ; the next (A'^) would produce a response but it is anticipated by the ventricular centre. This slightly premature contraction of the ventricle enables the bundle to recover its conductivity and to pass the sixth auricular impulse (A^) to the ventricle. The same events are then repeated. The result is an almost regular ventricular action, alternate cycles being a trifle short ; these are terminated by responses to the auricle ; the alternate and longer cycles are terminated by escaped beats of the ventricle. This is an example of escape of the idio-ventricular centre, for curves from the same patient frequently showed the usual picture of dissociation. centre which escapes is either the A-V node or the bundle below it, for the escaped beats are of supraventricular type. In the case of isolated escape a nearer localisation is not always possible, but it is clear that many in- stances hitherto regarded as " ventricular escape " — a term hitherto implying the activity of the idio-ventricular centre — are in reality due to ^- F nodal escape (see Fig. 154). I use the term " ventricular escape " in the present chapter in a more general sense, and without wishing to infer that the impulse necessarily arises in the ventricle. The mechanism may be illustrated diagrammatically. The auricular and ventricular contractions are indicated by the rectangles A and F respectively (Fig. 153). Stimulus production in a new centre is represented by the rising and falling line S.P. It is supposed that the auricle is beating in response to impulses produced in the 8-A node. At the same time the new centre is elaborating impulses at a constant rate. At the beginning and ending of the tracing, when the S-A rhythm is rapid, the ventricle has no chance to escape, 202 CHAPTER XVI. the impulses being discharged while they are still immature. In the middle of the figure, where the S-A rhythm slows sufficiently, a single impulse develops fully and, discharging, creates a ventricular response which anticipates the response of the ventricle to the oncoming auricular impulse. Solitary escape of this kind is frequent clinically (351, 427, 480, 631). It was first described by Mackenzie (502), and a notable example has been recorded by Wenckebach (762). Illustrative examples are seen in Fig. 154 to 157. It is to be noted that such solitary escape of the ventricle comes when the ventricle is starved of its natural impulses to contract, either by slowing of the 8-A rhythm (Fig. 155) or by A-V heart-block (Fig. 157) ; and in a given patient it is often a feature that the escaped contractions terminate all diastoles of more than a given length and terminate these only (414, 427, 480, 762). This is explained if we assume that the impulses which occasionally excite the ventricle are built up persistently and at a constant rate (Fig. 158) ; whenever a sufficient time interval elapses an impulse matures and, discharging, excites the ventricle. Escape of the ventricle accounts in part for the rarity of high grades of partial heart-block ; 2 : 1 block is common ; 3 : 1 block is rare.* Curious disorders of conduction in which 2 : 1 and 3 : 1 cycles are mixed with spontaneous ventricular beats are on record (480) ; and in instances where there are successive escapes of the ventricle the picture which Erlanger terms " relative complete heart-block " is produced. These instances are certainly instances of idio-ventricular escape ; they pass insensibly into dissociation. Instances are on record also of alternate responses of the ventricle to auricular and to new impulses in which the pulse may be almost if not quite regular. An example of this kind is shown in Fig. 158. A not dissimilar instance, associated with factitious alternation of the pulse, has recently been recorded (713). The reason for escape of a new centre is not always clear, or rather the reason for the change in the relative activities of the centres, the one emerging-, the other submerging, is not always easy to define. Thus, if a given heart rhythm is interrupted by a single premature beat or by a number of such beats following each other successively, another rhythm centre may be stimulated to an unusual, though transient, wakefulness (490). In Fig. 159 a rhythm of A-V nodal origin is interrupted by a beat forced from the ventricle ; after the disturbance, the S-A rhythm pre- dominates for a short while, its rate being temporarily enhanced through unknown channels. A clinical example of escape of the A- V node, consequent upon a premature contraction originating in the auricle, is to be found in Fig. 149, page 193. Another clinical instance of escape, though of a slightly different kind, is to be seen in Fig. 160. In this patient (451), premature contractions, arising in an abnormal auricular focus, were frequent, and it * There is another cause for the rarity of -3 : 1 bloei;, though we are not aware of its nature. In experiment 4: 1 block is much more frequent than 3 ; 1 block. ID I 0-VENT RI GIJLAR RHYTHM, ETC. 203 Fig. 159. (Heart, 1913-14, v, 335, Fig. 9.) An electrocardiogram from a dog. A-V rhythm is interrupted by a beat forced from the right ventricle. An escape of the S-A node succeeds this disturbance and is associated with accelerated action of the heart. As the heart rate slows again, A - V rhythm is resumed. Time in fifths of a second. Fig. 160. A single premature contraction of auricular origin in a patient. The first beat after the pause originates at an abnormal auricular focvis, as does the premature contraction. Time in thirtieths and fifths of a second. nearly always happened, as in the present instance, that the first beat following the disturbance was generated from the same or a neighbouring and unnatural focus. Centres of rhythm production. The natural pacemaker, the 8- A node, is endowed with higher rhythmicity than any other portion of the cardiac musculature. Of other centres capable of usurping this function in pathological conditions there are several. Of these ectopic* centres, the A-V node, of similar structure, stands first. There are many reasons for believing that all intact portions of either of these it nodes are possessed of rhythmic activity. In destructive experiments has been shown necessary to eradicate the whole S-A node to ensure the in- activity of this region of the auricle (77) ; in my own experience, I have found cooling of the head of the node insufficient to induce A-V rhythm in dogs, which develops whenever the temperature of the node is reduced in its whole * Hering uses the term heterotopic in the same sense. By the adjective ectopic I wish to express in this and subsequent chapters the origin of beats, whatever their nature, from a focus other than the S-A node, and to convey no more than that in respect of their qualities. 204 CHAPTER XVI. length ; stimulation of the vagus often depresses rhythmicity in the head of the S-A node and permits the escape of slower rhythmic impulses from its tail (483). It is probable that the same universality of rhythmic function is possessed by various parts of the A-V node ; many suggestive reasons for this belief have been discussed. It is not proved that any parts of the auricular muscle other than those mentioned have enough inherent rhythmic power to control and maintain the heart beat. Erlanger's experiments (145, 147), it is true, profess to show that many regions of the right chamber and septum may produce rhythmic impulses under given conditions ; but it is not always clear that portions of one or other node were not included unintentionally in the areas tested.* The left chamber is considered by these writers to be wanting in rhythmicity, though Fredericq (186) appears to hold that exceptionally it may be rhythmic. It is certain that if rhythmic power is present in the portions of the mammalian auricle which do not contain nodal tissue, the power is but poorly developed ; and it is probable, even if rhythms may so arise, that they are incapable of maintaining the heart beat for any considerable length of time under natural conditions. In the ventricle, the bundle is certainly endowed with rhythmic power ; its continuation in branches of similar elemental constitution suggests that these possess similar properties. It is said that the main divisions are readily excited by heat (205). and we know that when both divisions of the bundle are cut the ventricle continues to respond to impulses generated within itself. Automatioity in the arborisation has received little or no study ; if it be present universally it can only be so in relatively low degree. The trend of conclusion, as it concerns the complete mammalian heart, is to proclaim the special structures as the chief agents of rhythm ; quite possibly they are the sole agents. Those special structures which are composed of minute muscular elements are particularly, but not exclusively, endowed with the rhythmic function. * Moorhouse {562), writing more recently, states that strips are equally rhythmic whether they contain nodal tissue or not. If that is so, it is difficult to understand why the auricle in situ ceases to beat when iS-^ and A-V nodes are thrown out of action (482, 689). Hering [284) stated that after isolating the two nodes from the mass of right aviricular tissue, the latter continued to beat, and, on the strength of these experiments, concluded quite positively that there are other auricular centres capable of rhythmic action. The hearts, subsequently examined by Koch (388), showed the weakness of this conclusion, for nodal tissue was found attached to the mass of auricular muscle. The discussion over these hearts has been continued by Hering (295), who still maintains that in one instance his conclusion has obtained justification ; it serves to empha- sise the importance of strict histological controls in all experiments of this type. Chapter XVII. VENTRICULAR EXTRAS YSTOLES. When the ventricle of an animal is stimulated in diastole by a mechanical or electric shock, it responds to the shock and contracts. This power of response to artificial stimuU has been termed excitability* The ventricle is excitable, as Marey (537) showed, during the whole of its diastole, but is inexcitable or refractory during the period of its systole. The strength of a response of a ventricle to stimulation obeys the law stated by Bowditch (38) ; it is independent of the strength of stimulation. But the strength of the beat is controlled by the rest which the muscle has enjoyed before a fresh contraction is forced. If the ventricle is excited to contract during its natural diastole by a single induction shock, a disturbance is produced which may be illustrated diagrammatically (Fig. 161). For the first two cycles lo Ic Fo Re Rsc Rsc Fig. 161. A diagram illustrating the disturbances of the heart's mechanism when a systole is forced by exciting the ventricle during diastole. Ic = initial cycle ; Fc = forced or extrasystolic cycle ; Be = returning cycle ; and Rsc = restored cycles. P is the premature or forced beat. The auricular rhythm remains undisturbed. The forced and returning cycles are together equal in length to two initial cycles. the heart is represented as beating naturally. During the next ventricular diastole a premature beat P is forced by stimulation of the ventricle. This forced beat is followed by a diastole of unusual length and then the normal heart beats return. It will be convenient for descriptive purposes to name the several cycles of this figure ; calling the first cycles "initial cycles" (/c) ; the short cycle, the "extrasystolic cycle" or " forced cycle " (Fc), according as the beat which ends it is * The term is also used very generaUy, but I think inad^sedly, to cover the power of response to natural impulses. Of the last we have no measure, consequently I use the term in its more restricted sense. 206 CHAPTER XVII. spontaneous or forced by stimulation ; the long cycle, the " returning cycle " (Re), and the remainder the " restored cycles " (Rsc). The systoles of the figure may be conveniently quahfied by the, adjectives applied to cycles which precede them. It was first pointed out by Knoll (381), that the forced and returning cycles are together equal to two initial cycles ; the explanation of this fact we owe to Engelmann (128), who worked with the frog's heart. It is that the premature or forced beat has arisen independently Fig. 162. Myocardiographic curves ( F = ventricle, -4 = auricle) and Htirthle carotid pressure curve (C), from a dog in which the right coronary artery had been tied. Showing a single spontaneous and premature contraction of the ventricle. The pause following the weak beat in the arterial curve is compensatory ; the auricular beats maintain their rhythm. The prematurity of the carotid beat is barely perceptible on account of the temporary increase of presphygmic interval. Time in seconds. of an auricular contraction, and that the succeeding auricular impulse (represented in the diagram by the dotted oblique line) falls upon the ventricle when it is in a refractory state [refractory feriod). As a consequence, the • ventricle fails to respond, and awaits the call of the next auricular impulse. The cycle which follows the premature contraction is therefore prolonged, and VENTRICULAR E X T R AS Y S T L E 8 . 207 inasmuch as it makes amends, by its length, for the shortness of the preceding cycle, its diastole is termed the compensatory pause. The same phenomena were observed in the. mammalian heart by Cushny and Matthews (96). An experimental example of a premature ventricular contraction is shown in Fig. 162. In this figure the rhythm of the auricle is undisturbed ; the rhythm of the ventricle is disturbed temporarily. The returning and restored contractions of the ventricle fall where they would fall had there been no disturbance. Such is the rule. But not infrequently they may be a little displaced ; the systoles of the ventricle, which succeed the disturbance, appear a little earlier than is anticipated, when conduction from auricle to ventricle is quickened, as it may be, after the unusual rest of the long pause. This change in the conduction interval (As-Vs) is illustrated in an exaggerated form in Fig. 163. Fig. 163. Similar events to those shown in Fig. 161, but showing some displacement of the beats which follow the forced contraction (P) as a result of change of conduction intervals {As-Vs). The returning cycle is long, and the pavise rests the conducting tissues ; when the heart resvimes its harmonious beating the conduction intervals are at first shortened by reason of this rest. Another variation of the events may be seen from time to time. Of the rhythmic auricular beats, one customarily fails to produce a corresponding beat of the ventricle. It fails because the impulse conveyed from auricle to ventricle falls during the refractory period created by the premature systole. But if the heart's action is slow and the forced beat occurs at an early period of diastole, the forced systole may terminate before the rhythmic impulse from the auricle becomes due (584) (see diagram below, Fig. 164). In such a case the ventricle responds to each auricular impulse, and also contracts in response to the unusual excitation. This form of disturbance is termed an interpolated extrasystole (104, 411, 704).* The As-Vs interval * The reason why many of these extrasystoles fail to propagate themselves to the auricle, producing a retrograde beat, remains mysterious. In some instances the retrograde impulse is calculated to coincide with the next natural auricular impulse, but this is not always the case. 208 OH APT E It XVII. ^ , : . 1 , \ : : . ' inU^ : -iU^ : ' 6i 1 * r — 1 p — 1 — 1 — r n i -t-'H ^ =-EE5 p — M- ^ n n i TJ-r* — B ^M 1 i 1^ ■ ^ y i P N ^ to ^ Fig. 164 Venous, radial and electrocardiographic curves from a patient, exhibiting an interpolated extrasystole of the ventricle. The first two cycles are natural ; the diastole of the third is disturbed by a premature beat of the ventricle ; the succeeding auricular impulse yields a ventricular response, but, as the rest has been short, the corresponding As-Vs interval is lengthened somewhat. The P summit of the returning cycle is obscured because it falls on the end of the anomalous ventricular complex. Time in fifths of a second. UiUiU'lUiltiiUlJlj^ Fig. 165. Human curve. Premature contractions arising in the right or basal portions of the ventricle and interpolated between normal heart cycles. Time in thirtieths of a second. VENTRICULAR E XT R A S Y 8 T L E 8 . 209 following the interpolation is usually, though it is not always, materially increased.* The disturbances of the heart's mechanism which are seen when the ventricle is stimulated artificially are faithfully reproduced when, in man, premature contractions of the ventricle come spontaneously and when, in animals, they appear in response to the introduction of toxic substances into the circulation. Because there is this likeness in detail, spontaneous extrasystoles are often ascribed to an unusually excitable condition of the ventricle. This conclusion or inference is not justified ; it is not proved that extrasystoles as they occur spontaneously are due to disturbed excitability of the heart (see Chapter XXIX). Records of ventricular extrasystoles. It was Cushny {86) and Wenckebach [754, 755) who first suggested that a certain form of pulse intermission seen in the human subject is similar, in so far as the disturbed order of chamber sequence is concerned, to the intermission of the pulse produced experimentally by stimulating the ventricle. The proof that the two disturbances are aUke came when Mackenzie published his early venous curves {500, 501). Arterial records. — The rhythm of the pulse is disturbed by an intermission. In most instances, this long pulse cycle has exactly twice the duration of the usual pulse cycle. The pulse shows, on measurement, that the dominant rhythm is undisturbed. The long pulse pause may or may not show a trace of the premature ventricular contraction. The premature beat of the ventricle is weak since the contractile power of the heart has had insufficient rest fully to recuperate. The weak expression of the contraction in the arteriogram is in part the outcome of this unrestored power ; it is in part due to relative emptiness of the ventricle when the latter is called upon to contract. If the premature beat occurs sufficiently early, it may be so weak, or the blood content of the ventricle may be so small, that the aortic valves are not raised ; in that case a second heart sound is not produced and the pulse is not affected by the extrasystole (Fig. 166). Occurring later in diastole the extrasystole affects the pulse and, according to the power of the contraction and the amount of blood evacuated, the extrasystole expresses itself as a minute wave or as an almost fully developed pulse (Fig. 167). The returning contraction of the ventricle, i.e., that which follows the long pause, is powerful and the pulse prominent, for reasons the reverse of those which have been considered {626). Premature beats are always accompanied by a movement of the heart's apex and by a first heart sound, though the latter is often modified. * The reason of this prolongation is also obscure seeing that the extra beat has not been propagated through the A-V bundle. A premature beat, forced from the ventricle in experiment during partial A-V block, often heightens the degree of block notably though the beat is not propagated to the auricle {484, 570). 210 CHAPTER XVII. The presphygmio interval of the weak beat is usually longer than those of normal beats in the same case, and so the degree of prematurity is not always fully displayed in the pulse (see Fig. 162). Interpolated extrasystoles usually fail to affect the arterial pulse (see Fig. 164) ; when an arterial pulsation does occur, the next pulse wave is diminished in size. The larger the pulse of the interpolated beat, the smaller is the succeeding pulse. When, as sometimes happens, the interpolated arterial pulse is equal to the succeeding pulse beat, the two have the appear- ance of a pair of extrasystoles, from which oftentimes they may only be distinguished by electrocardiography. T»r»»ti»»»»t»»ir»rr»»»»»»»»»»»»»r»rr>»»» B^m-oTuL Fig. 166. Venous and femoral curves from a dog, showing the effects of a single premature ventricular contraction induced by electrical stimulation of the ventricle. There is no sign of the early beat in the arterial curve ; it falls at the same time as the anticipated a wave and gives rise to an exaggerated wave ° in the phlebogram. Fig. 167. (x §.) Extrasystoles of ventricular origin in a patient. The last part of the tracing displays a normal heart action. The first part shows extrasystoles, alternating with normal cycles. A rhythmic auricular beat falls with each extrasystole and produces an exaggerated wave ? in the venous curve. Venous curves . — These show the undisturbed sequence of the auricular contractions (Fig. 166 and 167). Where the premature ventricular beat falls synchronously with an auricular systole (see Fig. 175), an exaggerated wave " is produced. Electrocardiograms. — When a premature beat arises spontaneously in the ventricle or is forced in an experiment, the electrocardiogram usually displays every contraction of auricle and of ventricle, and the full analysis of chamber contraction may be read in these curves alone. The auricular complexes VENTRICULAR E XT R A S Y 8 T L E 8 . 211 are of constant form throughout (Fig. 168) ; the ventricular complex of the forced beat is anomalous and varies in form according to the area of muscle in which it has arisen. In these electric curves the anomalous ventricular complex falls simultaneously with the rhythmic auricular complex, and the two are superimposed (see page 160, Fig. 110 and 111, and explanations). :»/-f m _^ uMmmmmtmmmm AMMJ^AAMki Fig. 168. Myocardiographic curves (^4) from the auricle and (F) from the ventricle of a dog, accompanied by an electrocardiogram (lead II). A single beat has been forced prematurely by stimulation of the apex of the left ventricle and has yielded an anomalous complex. The rhythmic auricular complex falls with this and is superimposed upon it. Time in thirtieths of a second. Ventricular curves of heats arising from the ventricle. Forced heats. — In curves taken from an axial lead (such as lead II), stimulation of the apex of the dog's left ventricle produces a diphasic curve, of which the first phase is dowriward and the second phase upward (Type L, page 154, Fig. 103). On the other hand, upon stimulating the basal region of the right ventricle, the phases are reversed (Type R, page 154, Fig. 104). Kraus and Nicolai {390, 391), who first recognised these distinct types, regarded them as expressing contractions of the corresponding ventricle, and contractions more or less completely confined to them {hemisy stole). Their hypothesis as stated in its original form is untenable and, indeed, has been modified by the same writers (392). It is true that stimulation of the greater part of the ventral surface of the right ventricle yields a curve of the second type (Type R) and of the ventral surface of the left ventricle a curve 212 CHAPTER XV II. of the first type (Type L). But, as subsequent writers* have pointed out, the question of type in relation to region stimulated is not so simple as Kraus and Nicolai imagined [361, 362, 475, 670, 673). The chief facts may be summarised. No two points of stimulation yield precisely the same resultant curve (Fig. 170), and an infinite variety of forms may be obtained from one and the same heart ; but if two points stimulated lie close together, the corresponding curves resemble each other, and the resemblance is the greater the nearer the points of stimulation approach each other. Stimulation of a given point always yields the same ventricular curve, and this is so whether the stimulus falls in early or late diastole {447). Now when the ventricle is directly stimulated, the spread of the excitation wave must clearly be abnormal, and to this abnormality of spread, the dissimilarity of the complexes of natural and excited beats is due. Clearly the distribution in the case of two points stimulated will not be very different, providing these points are close to one another. The excitation wave spreads from the point stimulated and travels in every direction radially 4?^^ ■ ■ ' '^"" • Fig. ] 69. Premature contraction arising in the left ventricle. A human curve for comparison with Fig. 168. through the muscle ; it also pierces the thickness of the wall and reaches the Purkinje substance ; reaching this substance, it is propagated with great rapidity over the whole lining of the corresponding chamber, and ultimately spreads to the opposite ventricle. | Thus the ventricle being stimulated, the wave is at first confined to the adjacent muscle, and, until the muscle area involved is considerable, the electrocardiograph, arranged at its usual sensitivity, shows little or no movement. But later the spread is controlled by Purkinje paths, the remainder of the muscle being stimulated through these, and the excitation wave then travels simultaneously from within outwards over a large part of the wall {475). This event is signalled by the first conspicuous movement of the string. The ultimate direction of spread therefore is similar to that in the corresponding ventricle -when this is activated through normal channels. To this fact, in large part, is due the * A full summary of the observations up to 1914 will be found in Kahn's monograph (367). ■j- Hemisystole, or a limitation of contraction to one ventricle, has not been shown to occur either experimentally or clinically, see page 183. VENTRICULAR E XT R A8 Y ST L E S . 213 close resemblance between the curves corresponding to defects in the divisions of the bundle (dextrocardiogram or levocardiogram) and those obtained by stimulating the opposite ventricle (right or left, respectively). The influence of spread and the forms of curves obtained may be illustrated by the details of the following experiment. The ventral surface of a dog's heart having been exposed (Fig. 170), the muscle was excited at a number of points (1-14), using at each point successive threshold stimuli. The corresponding responses of the heart were recorded, using lead II, and single ventricular cycles are shown to the left in the figure. The signal of stimulation was also recorded (see curve 13), and in each instance the time interval between the signal of stimulation and the first prominent deflection was measured. These intervals are tabulated. Subsequently the heart was hardened in a natural condition of diastole and an oblique section was cut passing through the points of stimulation along the line 3-13. A diagram of this section is also represented in the same flgure. The distance of the ventricular surface to the Purkinje network was measured in millimetres at all stimulated points. If the series of curves is examined Point stimulated. Distance to Purkinje system. Signal to 1st prominent deflection. mm. sec. 1 R. 4 00363 2 R. 3 00416 3 R. 3 0-0424 4 R. 4 0-0405 5 R. 3 00407 6 R. 3-5 00473 7 R. 1-5 00293 8 R. 2 00245 9 R. 2 00343 10 R. 3 00497 11 R. 7 L. 8 0-0686 12 R. 12 L. 10 0-0920 13 L. 9-5 0-0920 14 L. 8 0-0699 it win be noticed that the ventricular responses obtained over the whole ventral surface of the right ventricle yield very similar outUnes in the electrocardiogram ; these differ from each other in detail and magnitude.* As the point of stimulation is moved along the A-V groove from the right towards the left margin and on to the conus (from 1 to 5 in Fig. 170), the excursion of the string, in its response to the heart beat, increases somewhat * The curves are complicated by the presence of auricular complexes ; these are of the type accompanying natural auricular beats in the case of curves 3, 4, and 5 ; the remaining curves show auricular complexes of the retrograde type. 214 CHAPTER ZVIl. in magnitude. As the stimulating electrodes are moved from the A-V groove (at 3) downwards towards the apex of the heart, the change in form and in magnitude is slight* until the region of the coronary vessel is passed. As the descending branch of the left coronary artery is crossed there is an abrupt change in type ; point 1 1 yields a complex of intermediate type ; the com- plexes from points 12 and 13 have their two chief phases inverted as compared to those obtaining over the right heart. Briefly, the responses from the right ventricle produce electrocardiograms presenting the chief features of dextro- cardiograms. while the responses from the left ventricle produce electrocardio- grams showing the chief features of levocardiograms. Now there is a relation between the length of the interval (signal of stimulation to first chief phase) and the thickness of the underlying muscle (see Table). In curves 12 and 13. and preceding the chief or downwardly directed movement, there are very evident preliminary phases of a diphasic character. In curves 4, 7, 9 and 10 the complex opens with a steep upstroke, and there is no preliminary phase, but small preliminary phases are distinct in all the remaining curves. These preliminary phases are due to the initial spread in the muscle : for, as they are more or less prominent according as the muscle path is long or short, we may conclude that one of the chief factors governing their appearance is the amount of muscle activated before the Purkinje network begins to convey the impulse. The first chief phase, be it upward or downward, appears immediately after the Purkinje system is involved, and its magnitude is due to quick spread to a relatively large and corresponding muscle area.f Its appearance is delayed according as the muscle layer penetrated is thick. A little consideration will suffice to show that, when the surface of a ventricle is stimulated, only a limited mass of muscle responds to the ingoing wave of excitation ; the chief part of the ventricular wall responds to the outgoing wave due to involvement and spread through the Purkinje tissue. The mass responding to the ingoing wave will depend upon the muscle thickness, but it will never be large, because the velocity of conduction in Purkinje tissue is relatively very great. Thus it happens that the direction of travel in the ventricular tissue cannot be controlled by altering the point of stimulation. Stimulation at the base or apex of the right ventricle produces the same end result, an excitation wave travelling from within outwards over the greater part of the ventricular substance. This accounts for the general similarity of curves obtained when stimulating a large area of the right, or when stimulating a large area of the left ventricular surface. The second point to emphasise is the relatively abrupt transition from the type shown in curve 10 to that shown in curve 12. Clearly such transitions are due to * Sometimes the change is greater than here shown (666), the greater or lesser change depending to some extent npon the hne along which points of stimulation are chosen. t In extrasj-stolic curves, as they are seen clinically, preliminary phases are scarcely ever to be distinguished. Their absence suggests that the extrasystoles arise in the Piirkinje tissue and not in the ventricular muscje, V EI^ T RICU LAR E XT R AS Y ST L E S . 215 Hfe^ /-? (, Fig. 170. (Phil. Trans., roy. Soc, 1916, GCVlI, 279, Part IV, Fig. 10.) A series of curves (^f^ nat. size) takea from lead II in a dog and resulting from stimuli applied at the corresponding numbered points in the upper outline drawing of the dog's heart. The lower outline is a section of the same heart through points' 3 to 13 (J^ nat. size). The vertical lines cutting the electrocardiograms are the fifth spcgrid tinio lines of the original curves. P 2 216 CHAPTER XVII. spread into the right Purkinje system on the one hand and into the left Purkinje system on the other. Finally, when the point stimulated is immediately to the left of the artery (point 11) curves of intermediate type are obtained, and these curves often resemble natural bicardiograms in the same animal very closely, as Rothberger and Winterberg {666) and others have pointed out. The explanation is that the two Purkinje systems are involved almost simultaneously. Thus point 11 lay 7 mm. from the Purkinje network of the left ventricle and 8 mm. from that of the right. The electrocardiogram (curve 11) is of dual origin ; it is an algebraic summa- tion of right and left curves and closely resembles the natural bicardiogram in outline. The type of curve yielded by stimulating the surface of the ventricle seems to depend upon two chief factors, the relation of the point stimulated to the two networks of Purkinje and its relation to the mass of ventricular muscle as a whole. Of these two factors the first exerts the dominant influence. Fig. 171. (Xy^-) Simultaneous curves from a dog's heart. .4 = auricular and F = ventricular myocardiogram. After three natural cycles the heart responds to rhythmic induction shocks thrown into the right ventricle. The fourth Ijeat of the ventricle in this 6gure is in part a response to the auricular impulse and in part to the induction shock ; the remaining ventricular beats are pure responses to the rhythmic induction shocks. Time in fifths of a second. The type of curve naturally changes profoundly with change of lead. We shall understand more clearly the manner in which the excitation wave is distributed in forced contractions when in the future the change of the electrical axis of these contractions has been studied. Sometimes, when a stimulus enters the ventricle in very late diastole the whole ventricle does not respond to it, some parts of the chamber being VENTRICULAR E XT R A S Y ST L E 8 . 217 simultaneously excited by the natural impulse descending from the auricle {451). In these circumstances the ventricular complex has a transitional outUne. Fig. 171 is an experimental curve ; after three natural cycles the heart responds to rhythmic stimuli thrown into the right ventricle (see signal). There are six responses of the ventricle in which the form of curve is fully controlled by the artificial impulse ; these are the last six cycles of the curve. But between these two series stands a ventricular complex of transitional form ; its shape is intermediate between that of the natural complex and the complex corresponding to the excitation wave propagated solely from the point stimulated. The ventricle in this instance has responded partly to the auricular impulse (the P-R interval is slightly shortened) and partly to the artificial stimulus. A clinical curve in which a similar inter- ference between two excitation waves is seen in Fig. 171a. The first curve of this kind to be pubhshed will be found in my book (see 447\ Fig. 120). In that example a perfect transition from the normal ventricular complex to that fully representing the extrasystoHc form of spread is to be seen. A transition due to similar interference, though it is produced in a somewhat different fashion, is to be seen in the chnical example published by Christian {50, Fig. 10) ; in this instance the two centres from which the impulses spread were the auricle on the one hand and the ventricle (as an idio -ventricular rhythm) on the other.* Fig. 171a. (" Clinical Electrocardiography," London, 1913, 1st edit.. Fig. 48.) Premature contractions are shown which arise in the right ventricle late in the ventricular diastole. There are three such contractions (P-B) in the curve ; in the second and third instance the excitation wave has spread through the whole ventricle from the point at which the new beats arose. But the first of the abnormal beats has fallen later in diastole, so late in fact that the ventricle responds in part to the natural auricular impulse and in part to the new impulse ; the ventricular complex becomes therefore transitional in type. Time in thirtieths of a second. The complete length of an anomalous ventricular complex is equal to that of a natural ventricular complex, within small errors of measurement ; this rule holds for all such beats, forced or spontaneous, and is sometimes serviceable in defining the limits of a ventricular complex of odd outline {447). Spontaneous beats. — Clinical examples of ventricular extrasystoles give curves of varied form from case to case, but fall for the most part into two chief categories, the right and the left types (Fig. 172 and 173). * The events in this curve are quite clear, though they appear not to have been recognised by the writer. The shortening of the intervals between ventricular beats, while the ventricle responds to the auricle alone, should be noticed. 218 CHAPTER XVII TTmrtTniLinjijiiilijiHiiiiMiiinMiitt niimnnHiinTiTii.TTmTnmniimTiin Fig. 172. i-tt: -——«-— ■^»s-'»ssa!««!«r-a a^::r-T?-- •^— r-: W%i - -f r'li"" -M-n*rt I rri II I ■ > ■ I I ■ > "jr iR ;4___; hUMMUMMUM ""■'■"■■■ J.I.....M Fig. 173. Fig. 172 and 173. Figures illustrating the two chief types of premature contractions of ventricular origin as they are portrayed in the separate leads in man. Fig. 172 shows a, ]5remature beat which arises in the right ventricle and Fig. 173 in the left ventricle. Figure 173 is atypical in that the deflections in lead / are usually reversed. Time in thirtieths of a second. It is not difficult to choose from collections of curves, clinical and experimental examples showing close resemblances (Fig. 168 and 169 and Fig. 176 and 177). Nevertheless, anything more than approximate localisation of the spontaneous beats is impossible at the present time, for we possess experimental observations upon the dog only, and we know that the lie of the heart and the arrangement of the Purkinje strands in this animal differ materially from that in man. The constant form of extrasystole in the electrocardiograms of a given clinical subject, from day to day, month to month, or even year to year, is verv remarkable (489), and indicates the constancy of the focus of irritation, VENTRICULAR E XT E A S Y ST L E 8 . 219 1 ^ ^ -^?^ S — ^ -^ ^ ^ 1 — ^ 1 1 — ^ ^ r^ ^ ?r N H- /^ •t— r^ ^ 4 ^ ^ 7^-r a^ * *^ ^ N ^ ^ ^ |w«i ^ *¥^ — 1 — — 1 ^ B*5 1 ^ 1 — iJ -V— H ^ i i -t! m 1 — * /= i n ■ Mi Pli -H^: l:z=L ^£^^3 li*^(i| m ^ "^^ P F P W W \ ~ ~ ~ ^ HH 1^1 —— I^H pii 1 Fig. 174. Simultaneous venous, radial and electrocardiographic curves from a pal,ient, showing an extrasystole arising in the right ventricle. A diagram placed below illustrates the mechanism of the heart over the period of the disturbance. Time in fifths of a second. iBMi^ 'fdo' Fig. 175. Simultaneous electrocardiogram, venous and radial curve from a patient exhibiting a bigeminal action of the ventricle, resulting from ventricular extrasystoles. A diagram placed below illustrates the relation of chamber contractions. Time in thirtieths of a second. 220 CHAPTER XVII. Fig. 176. (X y.) Simultaneous curves from the auricular and ventricular muscle (^ and V) and an electrocardiogram from a dog. The heart is responding from time to time to induction shocks thrown into the right ventricle. Time in fifths and twenty -fifths of a second. Fig. 177. (X 1^.) A human curve for comparison with the Fig. 176. The disturbance of the heart's regular action is due to extrasystoles arising in the right ventricle. Time in thirtieths of a second. It is relatively infrequent to find anomalous beats of more than one kind in a patient but, if found, both types are usually maintained. It has been suggested that these facts are to be explained by supposing ■ the Purkinje system to be the chief source of such extrasystoles {489). Chapter XVlll. AURICULAR EXTRAS YSTOLES. The premature contraction, when forced by stimulating the auricle, is followed by a similar premature contraction of the ventricle (Fig. 178). The disturbance affects both chambers, but the disorder is a little less in the veritricular than in the auricular movements. The conduction interval Ic Ic Pea Rca Rsca Fcv Rov Rscv Fig. 178. Diagram of the events when the heart is disturbed by a premature auricular contraction. Ic = initial cycle ; Fca = extrasystolio or forced auricular cycle ; Fcv = extrasystolic or forced ventricular cycle ; Rca and Rev = returning auricular and returning ventricular cycles ; Rsca and Rscv = restored auricular and restored ventricular cycles, respectively. varies according to the degree of preceding rest, the As-Vs interval of the forced beat being prolonged and that of the returning beat shortened. These changes are illustrated in an exaggerated form in the diagram (Fig. 178); they may be conspicuous in clinical instances of premature contractions where there is primarily impaired conduction, but when the heart is normal they are inconspicuous and may require very precise measurement fully to display them (Fig. 181). The returning cycle in the case of a premature auricular contraction is variable in length ; it may be equal to an initial cycle, it may be much longer and compensatory, that is to say its duration taken with that of the forced cycle may be equal to two initial cycles or, finally and usually, its length may have an intermediate time value (96, 252, 759). 222 CHAPTER XVIII. Fig. 179. Myocardiographic curves {A = auricle, V = ventricle) and Hiirthle carotid pressure curve (C) from a dog. A single and spontaneous contraction of the auricle is shown ; the ventricle follows suit and contracts early. The pause in the carotid curve is not fully compensatory. A curve taken from an experiment in which the right coronarj' artery was tied. Time in seconds. Engelmann (130), in his researches upon the amphibian heart, found that when a premature beat is excited from the sinus, the returning cycle has exactly the same length as an initial cycle. It is supposed that the forced beat discharges the immature impulse and that this is built up again from the instant of its premature discharge and at the accustomed rate (Fig. 182). Stimulation of the mammalian pacemaker produces similar elf ects (Figs. 180 and 183) according to my own observations [445, 490). and the recent experiences of Sansum (687). When this region is excited, the returning cycle does not exceed the initial cycle by greater time intervals than one or two hundredths of a second,* and often the two cycles may be of equal * It is not easy to stimulate the actual pacemaker, for the node is a long structure and the rhythm springs from a point. Where there is this difference in the lengths of the cycles, it is probably due to inaccuracy in placing the stimulus. Hirschfelder and Eyster [316), in the days before the pacemaker was located, were unable to find differences in the length of the returning cycle according to the point stimulated. AURICULAR EXT RASYST L E8 . 223 Fig. 180. (Heart, 1913 14, V, 335, Fig. 1.) Electrocardiogram from a dog showing a singlo premature contraction forced by stimulating the right auricle near the top of the sulcus. The lettering is the same as in Fig. 178. F.B. = forced beat. The initial cycle (7.C.) and returning cycle [R.G. A.) are almost equal in length. Time in twenty-fifths of a second. Fig. 181. (Heart, 1913-14, V, 335, Fig. 2.) Electrocardiogram from a dog. A single premature beat forced from the inferior caval region. The figures are time measurements, in decimals of a second, of the auricular (above) and ventricular cycles (below), and of the PR intervals (up and down). Time in fifths of a second. S.P Fig. 182. A diagram illustrating the disturbance of the heart's mechanism when a premature contraction is excited in the neighbourhood of the pacemaker. Stimulus production in the tissue which originates the heart rhythm is indicated by the line S.P. ; the impulse is supposed to explode when it reaches the line x x' and to fall at each contraction of the heart to the level y y' . c and d are equal in length. 224 CHAPTER XVIII. length. It seems probable, therefore, that the time interval which elapses between the discharge of the pacemaker and the full growth of the next impulse is constant and independent of the manner in which the discharge is brought about ; it seems to be the same if a stimulus arises in or is directly apphed to the pacemaker, or if an excitation wave reaches the pacemaker from some other region of the auricle. The length of the returning cycle exceeds the length of an initial cycle when the stimulus is applied to some outlying region of the auricle. It exceeds the initial cycle by the time taken for the excitation wave to travel to the pacemaker from the point of stimulation, as long since suggested (96, 759), and recently demonstrated {483). The time interval between the forced beat and the returning beat — the returning cycle, as I term it — comprises the period during which, in the first place, the excitation wave is travelling from the point of stimulation to the pacemaker where it discharges the forming impulse, and during which, in the second place, the new impulse is being built.* Fig. 183. Electrocardiogram from a dog showing a single premature contraction of the auricle, forced from the region of the pacemaker. The returning cycle in the auricle has the same length as the initial cycle and the complex P of the premature beat is normal in outline. Time in fifths and twenty-fifths of a second. The signal of stimulation is shown below. When the returning cycle is compensatory, as happens rarely, then it is assumed that impulse formation in the S-A node has remained undisturbed,' that is to say, the excitation wave of the premature beat has failed to reach the pacemaker. This may happen [490) if the forced beat comes relatively late in diastole, and the 8- A node discharges its rhythmic impulse before the extraneous wave arrives ; the two waves, natural, and forced, then meet in the auricular wallsf (Fig. 185 and 186). * This method of forcing extrasystoles was at one time used in the attempt to isolate the mammalian pacemaker (252, 316), but was unsuccessful, although the results which it yields do actually conform to the position of the pacemaker as it is now recognised. f This explanation does not apply to all compensated disturbances of the human auricular rhythm ; some are, I believe, accidental and attributable to temporary slowing of the heart. AU B I CULAR EXT RASYST LE8 225 fn^-t^-i^-^-:i::!-:^i-i--;if Fig. 184. Fig. 18'), Fig. 186. Fig. 184-6. {Heart, 1913-14, V, 335, Fig. 5, 6, 7.) From a dog. Three examples of a forced contraction disturbing S-A rhythm. The point of stimulation was in each instance the inferior cava. The shape of the pure inferior caval complex is shown in Fig. 184. In Fig. 185 and 186 there is almost exact compensation, and this has resulted because the forced contraction waves have failed to reach the pacemalcer before its discharge. This fact is jyitnessed to by the transitional form of the electric complexes. In each case two excitation waves, one of S-A nodal and one of inferior caval origin have met in the walls of the auricle. Time in fifths and twenty-fifths of a second. 226 GHAP'I'ER XVI t I Records of auricular extr asystoles. Premature auricular contractions, or auricular extrasy stoles, were first believed to occur in the human subject by Wenckebach {754, 755) and Cushny {86) ; these writers had arterial curves alone to guide them to their conclusion. Mackenzie's polygraphic records {500), and later electrocardio- grams, have fully substantiated their view. Arterial curves. — The arteriograms are similar in appearance to those disturbed by ventricular extrasystole, the premature beat is weak (Fig. 187) and may fail to appear in the curve. But measurement of the curves demonstrates a disturbance of the dominant rhythm of the heart (see page 132 and Fig. 187) ; the returning cycle is not compensatory except in rare instances. Fig. 187. A polygraphic curve from a patient exhibiting premature auricular contractions. The normal cycles are accompanied by a, c and v waves. The premature radial beats are represented in the venous curve bj' waves c'. Preceding the latter are prominent waves due to auricular systoles, ",. The auricle contracts prematurely and during the last phase of the preceding ventricular systole. Venous curves. — These show the auricular wave a' preceding premature ventricular waves c', v' (Fig. 188). If the auricular contraction falls with the preceding ventricular beat, as frequently happens, then the a' wave is exaggerated (Fig. 187). Electrocardiograms. — If the mammalian auricle is stimulated by means of single induction shocks and the responses of the heart are studied in electrocardiograms, it is found that the curves vary as each new region of the muscle is selected for stimulation. The ventricular phases are of constant form and usually of the same form as those accompanying the natural heart beat (page 60, Fig. 32), for the beats are all of supraventricular origin. It is the auricular complex which changes. When stimulation is near the pace- maker its outUne is almost natural (Fig. 183). When stimulation is near the A-V node or inferior cava or the orifices of the pulmonary veins, the outline is often inverted or partially inverted {445). AU B I CU LAR EXT EASY ST LES. '121 h. ==r- ::^C ^El^g^^ /^ -;— 0.-£l :Ut:^ Fig. 188. (X J.) Simultaneous venous, radial and electrocardiographic curves from a patient who presented auricular extrasystoles. A diagram representing the disturbed mechanism is placed below the photograph. Time in fifths of a second. Auricular extrasystoles as they occur in man present the same features ; the ventricular complex is usually of natural form, the auricular complex is variable in outline. Most commonly it is inverted in leads //and ///(Fig. 188 to 190), clearly indicating the ectopic origin of the beat {433), and probably pointing to an origin from those portions of the auricle which are in the vicinity of the A-V groove. It may be upright, differing from the natural complex in minor detail ; it may be isoelectric, and therefore invisible, {451) in some leads (Fig. 190, lead /). A beat arising in the auricle may be regarded as ectopic if its auricular complex departs from the natural complex in the same patient. It often happens, when an auricular extrasystole disturbs the heart's rhythm, that the anomalous auricular complex falls with the T of the preceding systole. In such cases the deflections are superimposed, T being double or notched (Fig. 189, compare T'^ and T^ in the top curve), and the auricular systole is found by carefully comparing the outlines of the several T deflections of the curves. The ventricular complexes associated with some premature auricular contractions are anomalous {433, 445, 451, 658), and may be confused with the curves of ventricular extrasystoles. They should not be confused if the}'^ are preceded by clearly inscribed and premature auricular complexes, and the P-E interval is not shortened, for in these circumstances the auricle is indicated as the primary seat of irritation. The change in the form of the ventricular complex described is due, so it is maintained, to defects in conduction through some of the chief Purkinje strands {445), {aberration). The divergence of the ventricular complex from normality is most conspicuous when the extrasystole falls early in diastole ; iri some patients the degree 228 CHAPTER XVIII Fig. 189. Four curves from a single subject. Each shows a solitary premature auricular contraction. The premature auricular complex falls with the commencement of the preceding T and notches it. The corresponding ventricular complexes of the first three curves are of various forms ; the central curves illustrate " aberration." In the last curve the premature auricular contraction is blocked. Time in thirtieths of a second. "" ■■■■—■■■■■■■■.»■■ ~ .- - ■■■•.- ■"irimiiiiiiiiiiiiiiliiiiuiuii Fig. 190. Two curves (leads / and ///) from the same patient and showing each an auricular extrasystole. In lead I the premature auricular complex is isolectric and therefore uuisible. Time in fifths an thirtieths of a second AUniGULAR EXT RA8YST LES. 229 of aberration is closely connected with the degree of prematurity. It is notable that in these patients minor conduction defects in the main A-V bundle are the rule (see last curve of Fig. 189). In Fig. 189 two conspicuous forms of aberration, occurring in one and the same patient, are depicted. The last curve illustrates an auricular extrasystole which has failed to evoke a ventricular response, the blocked auricular extrasystole first described by Hewlett {307, 433, 640, 655, 658) ; the additional P deflection falls with T^ and notches it and a long diastole follows. Aberration has been recorded in experiments as an accompaniment of forced auricular contractions ; the form of the ventricular curve in these circumstances is independent of the site of auricular stimulation, but it is governed largely by the phase of diastole in which the stimulus falls ; the distortion is greater if the stimulus falls in early diastole {445).* Sinus extrasystoles . — As a rare phenomenon, extrasystoles occur in the human subject which are to be interpreted as arising in the S-A node ; they have been termed sinus extrasystoles {758). The returning cycle is no longer, it may be actually shorter, than the initial cycle (Fig. 191). In electrocardiograms the auricular summits are of constant outline, whether they belong to the rhythmic series or to the premature beats. Fig. 191. Human curve. A single premature contraction, probably arising in the immediate neighbourhood of the sino-auricular node. The rhythmic and premature contractions of the auricle yield similar electric curves ; the returning cycle is in this instance shorter than the initial cycle. * Auricular extrasystoles, which yield ventricular complexes of aberrant types, may be forced experimentally by stimulating the auricle in an animal in which slight conduction defects have been induced already, for example in the cat by asphyxiation (unpublished observations). Chapter XIX. PREMATURE BEATS ARISING IN THE JUNCTIONAL TISSUES; RETROGRADE BEATS; PREMATURE BEATS DISTURBING NEW RHYTHMS, ETC. Premature heats arising in the junctional tissues. Examples of experimental extrasystoles [269) and of clinical extrasystoles [303, 512), have been described* in which both the auricle and ventricle contract prematurely, and in which the contractions of auricle and ventricle begin almost simultaneously. They are explained by supposing that the new impulse is formed in the junctional tissues, and that the excitation wave spreads at one and the same time to auricle and ventricle (Pig. 192). Contraction of the auricle in response to one impulse and of the ventricle in response to a distinct impulse cannot be allowed, for in patients who present these singular beats, the same events are exactly repeated. Fig. 192, A diagram illustrating the mechanism of the heart when the sequence is disturbed by a premature beat arising in the junctional tissues. A clear example of this disturbance of the heart's action is to be found in the accompanying figure, which illustrates the features of these beats in simultaneous records (Fig. 193). In the arterial curve of this example, the disturbance resembles the premature ventricular contraction in every respect, for the returning cycle is compensatory ; in other examples the cycle * The first clinical examples are those published as retrograde contractions by Pan (585), and by Volhard (741). PREMATURE AND RETROGRADE BEATS. 231 IS less than compensatory in its length. The venous curve shows a tall wave ^, resulting from contraction of the auricle while the tricuspid valve was closed ; a prominent wave of similar origin may be found as an accompaniment of ventricular extrasystoles, but the junctional extrasystole is distinguished by the position of the composite wave ; it falls before the time when the rhythmic auricular wave is expected. In the case of the ventricular extrasystole, the exaggerated a wave belongs to the rhythmic series ; its coincidence with ventricular systole is in this circumstance accidental ; in the case of the junctional extrasystole the auricular rhythm is disturbed and coincidence is forced. In the electrocardiogram a premature ventricular complex is seen and this is of sufi&ciently normal outline to show that the ventricle is stimulated from a supraventricular focus ; the auricular complex is buried and hidden in the ventricular complex and in this instance is not cletfrly seen; the positioii of the auricular systole may be judged by comparing the venous and electrocardiographic curves. Fig. 193. ( X^Tj.) Venous, radial and electrocardiographic curves from a patient who exhibited extrasystoles arising in the region of the 4 - F node. Time in fifths and twenty -fifths of a second, ' In other instances of premature beats, whose origin is ascribed to the junctional tissues, the auricle contracts before the ventricle ; the P-R interval, however, is slightly or conspicuously shortened ; in these curves a greater or lesser part of the auricular complex is visible and the origin of the beat from the junctional tissues is confirmed by the inversion of P. In yet other instances, there is a distinct Vs-As interval. An example of this class is shown in Fig. 194. Somewhat less than one-fifth second after the beginning of the first premature ventricular systole of this figure, the auricle also contracts, and the inverted auricular complex, (P) marked below the curve, notches the beginning of T. The origin of this extrasystole is clear. The Q 2 232 CHAPTER XIX. chambers have contracted in response to a common impulse, for in both chambers contraction is premature. The contraction of the ventricle precedes that of the auricle by an interval which is rather less than the natural P-R interval in the same subject ; the impulse has arisen at such a point that the spread to the ventricle takes less time than the spread to the auricle. The impulse has not arisen in the ventricle itself, for the shape of the ventricular complex indicates the supraventricular origin of this beat. The inversion of P corresponds to the abnormal course which the excitation wave takes in the auricle. The shape of the auricular complex is consistent with the spread through the auricle from the A-V node ; and the length of the Ys-As interval indicates the origin of the impulse from the lower reaches of the A-V node or the A-V bundle (477). When the extrasystole is less premature, it may happen, as in the second premature beat of the same figure, that, while the impulse is passing towards the auricle, the auricle responds to the rhythmic impulse of the 8- A node [448). The nodal impulse finds the auricle in the refractory state and the natural outline of the auricular contraction (the seventh P of the curve) is consequently maintained. Retrograde heats. In some rare instances where auricle and ventricle both contract prematurely and where a Vs-As interval is found, it is probable that the extrasystole has arisen in the ventricle and has spread in retrograde fashion to the auricle. That the heart's mechanism may be reversed, i.e., that the auricle may beat in response to the ventricle, is readily shown by stimulating the ventricle with successive induction shocks sent in at a rate exceeding the previous heart rate. The ventricle responds to each stimulus and, after a variable number of cycles, each ventricular response is followed by an auricular contraction. The impulses are conveyed from the beating ventricle to the auricle through the A-V bundle. But it is generally acknowledged (26, 523), that in roost animals conduction from ventricle to auricle takes place less readily than from, auricle to ventricle ; and it is usually considered that the interval Vs-As of retrograde beats is longer than is the .4s- Fs interval of the normal heart beats in the same animal.* * This is, I think, open to question ; the difference is in any case inconsiderable and the measurements of the intervals are beset with sources of small error, whether they are calculated in myocardiograms or electrocardiograms. The difficulty of exact calculation comes in estimating the beginning of contraction in a heart chamber ; tthe myocardiogram records contraction of the particular muscle area to which it is attached ; the electrocardiogram employed at ordinary sensitivities and taken from a limb lead does not signal the earliest involvement of the muscle. Although this doubt as to the actual intervals exists, yet there is little doubt that retrograde conduction is more easily hindered or disturbed than is forward conduction. In the presence of a slight defect of conduction in the A-V tissues, a retrograde action cannot be induced by rhythmic stimulation of the ventricle [484). PRE MATURE AND RETROGRADE BEATS. 233 Fig. 194. (Xf.) An electrocardiogram from a patient who exhibited extrasystoles arising in the junctional tissue. The first extrasystole involves both ventricle and auricle ; the ventricular complex is of the supra-ventricular type ; the auricular complex is inverted and notches T. The second extrasystole involves the ventricle only ; it falls later in diastole and the A-V nodal impulse finds the auricle refractory, as the latter is already contracting in response to a S-A nodal impulsey Time in fifths of a second. ^EESEE^ ^~v^n7/i^ " '^ g ■ > ^ <^^^^:^^m^^^ ^^ Fig 195 Simultaneous venous, electrocardiographic and radial curves, showmg an extrasystole arising in the junctional tissues. The ventricle responds to the new impulse, the auricle does not, for it is already contracting in response to the S-A nodal impiUse. Timem fifths of a second. 234 CHAPTER XIX But the relation of these two intervals and the factors governing reversed conduction are still by no means clear. That an impulse from a premature ventricular contraction is not conveyed to the auricle as a rule is easily understood, for at the time when the reversed impulse should reach the auricle that chamber is in contraction in response to the S-A node and is therefore refractory to further stimulation. But a similar explanation does not hold good for all premature contractions arising in the ventricle ; it is not always clear for example why some ventricular extrasystoles are interpolated and yet fail to reach the auricle (see footnote page 207). Fig. 196. (x^^Q.) Myocardiograms from auricle (A) and ventricle (V) and electrocardiogram from a dog. The normal rhythm is disturbed by four rhythmic beats forced from the conus region of the right ventricle by rhythmic stimulation (the signal marks of stimiilation are white dashes at the bottom of the photograph). For two cycles the auricle fails to respond to the forced ventricular beats, the impulses conveyed from the latter falling during the refractory period of the auricle. The last two cycles show response of the auricle (see myocardiogram), and the corresponding auricular complexes are ininute notches (P) on the ventricular complexes ; these little notches are more distinct in Fig. 171 (page -16), which is from the same animal. Time in fifths of a second. Pan (585) and Volhard (741) explained certain clinical curves on the assumption that the extrasystoles arose in the ventricle and were retrograde to the auricle, but it seems more probable that the examples with which they dealt were junctional extrasystoles. I have seen but a few examples of what may be regarded as isolated retrograde extrasystoles in man.* * The reversed rhythm described by Williams (780) is an undoubted example of a rhythm arising in the junctional tissues, for the ventricular complexes are of supraventricular type. A similar explanation might be applied to the case recorded by Norrie and.Bastedo (575'), though in this instance the mechanism is uncertain ; their published curves might be interpreted in either way or might be attributed to an unusually prolonged a-c interval. PRE MATURE AND RETROGRADE BEATS. 235 >>. ij__ ,,__,, '^ . ' ' ^ .' As 5 Vs Fig. 197. (>i S Ja tf S t4_, B O ca fi >> m ,o tS rl ^ -M o, fcl o -a "^ c C8 o O CO fl S HI TS r-J a i ^ c3 fe ^ s trt >- C (J « a e -a a '5b — ' •c -1 rfl Hi f^ (N bf) 14 244 CHAPTER XX. In clinical examples, the rate of the natural rhythm, before and after the paroxysm is the same in most instances. In exceptional cases the full rate of the natural rhythm is developed slowly at its resumption ; in these the post-paroxysmal pause (or returning cycle) is of unusual length and the heart rate accelerates for six or more beats until the original rate is restored. The slow rate of the natural rhythm when this is resumed is comparable to the slow rate of the idio-ventricular rhythm when the latter has been inter- rupted by a forced tachycardia ; but in the case of an 8-A rhythm the retardation appears to be exhibited by pathological hearts only. Fig. 210. (Xi-) An arterial curve showing the termination of a paroxysm of tachycardia The post-paroxysmal pause is unusually long and the natural rhythm at its resumption is slow but accelerates as it proceeds. The action of the heart during the paroxysm is remarkable for its regularity, the beats follow each other at equal intervals ; the regular sequence of the beats is conditioned by their origin from a single focus (as ascertained electrocardiographically). The rapid rhythm is also remarkable for the relative constancy of its rate for long periods and under various conditions ; thus the rate is uninfluenced by posture and by exercise ; it cannot be retarded by pressure on the vagi, though the new rhythm is sometimes abolished by this procedure (64). In short, these paroxysmal rhythms are not under the adjusting control of the central nervous system, as is the normal rhythm {464). Reaction of rapidly beating hearts to postural changes of the body. Simple Tachycardia. Standing. Lying. Standing. Lying. 2. Exophthalmic goitre 180 145 1. Pulmonarv tuberculosis 1. Cardiac case 142 133 2. Exo 144 130 143 129 Paroxysmal Tachycardia. Standing. Lying. 205 213 2. Care 209 204 204 210 212 212 176 147 Sitting. Lying. 195 190 200 199; 197 203 199 187 PAROXYSMAL TAGHYGARDIA. 245 The duration of paroxysms in a given patient is fairly constant ; from subject to subject the duration is very variable. A single paroxysm may consist of six or more beats ; it may continue for an hour, a day, a week or more, without interruption. Exceptionally in one and the same patient short and long paroxysms may be witnessed. That which constitutes a paroxysm of tachycardia is a matter of terminology. The soUtary extrasystole has been studied in a previous chapter, but extrasystoles may be grouped (585), the groups comprising two (Fig. 211), three or more beats (Fig. 214), and no dividing line can be drawn between these disturbances and such short paroxysms as are shown in Fig. 212. 1 9 v<^w ¥ f t / ^■■»> J V > ^ y rr » * w yy^v * » y » JihJiKJVittAAJUUUU^ ^ad/aX ex ex E« Fig. 210. (x J,) An irregularity of the heart, resulting from single extrasystoles of auricular origin, and pairs of extrasystoles. Apical and radial curves are shown. It is indeed notable that short paroxysms often immediately precede the onset, or immediately succeed the ending (Fig. 214) of a long continued paroxysm, and that in most patients who exhibit long or short paroxysms isolated premature contractions interrupt the slow periods of normal heart action. The isolated extrasystole and the shorter preliminary or terminal paroxysms have for the most part precisely the same origin as the long continued paroxysm in the same case (433) ; this fact has been determined repeatedly both by polygraphic and electrocardiographic analysis. The point of origin. — In localising the seat of disturbance in patients who suffer from paroxysms of tachycardia, the same methods are adopted as for the extrasystole. The paroxysm may arise in auricle, A-V node or ventricle. The origin in some patients is readily determined by means of venous curves ; it may be localised more exactly in most cases by thr employment of the electrocardiograph. Auricular origin. — A number of paroxysms are frankly auricular in origin (55, 80, 171, 243, 431, 433, 655), each paroxysmal beat being represented in the venous curves by a and c waves, though the former often falls with the preceding v wave and is of increased amplitude (Fig. 213). The same origin may be ascertained electrocardiographically, for the beats of such paroxysms present ventricular complexes of supraventricular type (433), and these are preceded by auricular complexes of anomalous outline, indicating the ectopic origin of the corresponding auricular impulses (433). 246 CHAPTER XX PAROXYSMAL T AGHYGARDIA. 247 These features are clearly seen in the accompanying figures (Fig. 215 and 216) in which beats from the normal rhythm and from the paroxysm are placed side by side ; curves from the three leads are shown. The constancy of the ventricular deflections, even in their smallest detail, in corresponding leads of the slow and fast rhythm, is excellently displayed in this figure ; the auricular complexes alone change ; during the paroxysms they are inverted in leads // and ///. Fig, 215. Fig. 216. Fig. 215 and 21G. Two sets of curves from a case of simple paroxysmal tachycardia. Fig. 215 was taken while the heart was beating slowly ; Fig. 216 while it was beating rapidly. The curves demonstrate the supraventricular origin of the paroxysm. The inversion of P in leads II and III oi Fig. 216 indicates that it arose in an ectopic auricular focus. Note the precise similarity of the ventricular elements in corresponding leads of the two series. Time in thirtieths of a second. Two other examples of auricular paroxysms are illustrated in Fig. 217 and 218. Fig. 217 was taken from a continued paroxysm in a patient who suffered from mitral stenosis ; the auricular complex is upright and falls with the preceding T deflection. The similar origin of paroxysmal beats and the extrasystoUc beats, interrupting the slow natural rhythm is illustrated by Fig. 218. The figure opens by showing the last five beats of a paroxysm : the post- paroxysmal pause follows and is terminated by a single normal cycle ; this in turn is succeeded by a pair of extrasystoles of auricular origin, the first ventricular response being aberrant ; two normal cycles follow. In this curve the aviricular complexes of the paroxysmal stage are of smaller amphtude than those of the normal cycles and rise more steeply : the auricular complexes of paroxysm and extrasystoles are alike, it may be noted, in parenthesis, that isolated aberrant ventricular contractions, such as are shown in Fig. 218, are commonly seen in patients who suffer from R 2 248 CHAPTER XX . paroxysmal tachycardia. I have pubUshed another striking example of the same kind (433). Fig. 217. Electrocardiogram from a continued paroxysm of tachycardia in a case of advanced mitral stenosis- The paroxysm arose in the auricle. Time in thirtieths of a second. Fig. 218 The end of a paroxysm of tachycardia of auricular origin and the commencement of a slow normal rhythm, interrupted by premature contractions. These are auricular and a pair of them is shown, the former of the two yielding an aberrant ventricular response. The timie-marlcer in this curve rules vertical lines ; a pair of lines occurs at each thirtieth of (I second. A-V nodal origin. — Paroxysms in which simultaneous contraction of the auricle and ventricle occur were first published by Pan (585) and Rihl (630). Their curves were polygraphic. Of paroxysms arising in the A-V node and characterised by a slight reduction of the As-Vs interval, I have recorded two examples electrocardiographically (434, 487). In the venous curve, the customary a and c waves give place to conspicuous combined waves and the calculated a-c interval is reduced. In Fig. 219 the normal interval is 0-2 of a second, while during the paroxysm it falls to 0-06 of a second. An electrocardiogram, accompanied by a radial curve from the same patient is shown in Fig. 220. It shows the onset of a similar paroxysm, which opens with two extrasystoles, presumably of auricular origin, and continues as a regular paroxysm of beats of A-V nodal origin. The last beats present an inverted auricular complex and a P-R interval shortened from 0-14 to 0-08 of a second. Another example is shown in Fig. 221 and 222 ; the lower curve illustrates the normal rhythm, the upper curve illustrates a paroxysm in which the P-R interval is reduced and P inverted. Both the normal rhythm and the paroxysm are interrupted by extrasystoles from a distinct fiuricular focus, PAROXYSMAL TACHYCARDIA 249^ 250 CHAPTER XX. The beginning and end of a brief paroxysm of A-V nodal origin is shown eleetrocardiographically in Fig. 223. The inversion of P and the shortening of the P-B interval is excellently displayed in this curve. Fig. 221 and 222 (x J^.) Two curves from a. case of simple paroxysmal tachycardia. Fig. 222 is from the period of slow and Fig. 221 is from the end of a period of rapid heart action. The curves show the auricular but ectopic origin of the paroxysm. Both slow and fast rhythms are interrupted by occasional premature beats having a common and auricular focus of origin. Time in fifths of a second. Fig. 223. (x ^-(j.) A short paroxysm of six beats shown eleetrocardiographically to be of A-V nodal origin. Pis inverted during the progress of the paroxysm and the P-i? intervals are shortened. I am indebted for this curve to Dr. John Parkinson. Time in fifths of a second. Paroxysms of supraventricular origin in which the ventricular form of venous pulse is seen. — In many examples of paroxysmal tachycardia the venous curve is of the ventricular form (Fig. 214) ; in these it is not always possible to locate the seat of disturbance, though in all we probably have to deal with simultaneous contraction of auricle and ventricle. The difficulty is that there may be no visible trace of auricular complex in the electrocardio- gram {324, 446, 488). In describing the varieties of electrocardiogram accompanying the simple and slow form of ^-F rhythm (Chapter XV), it was stated that when auricle and ventricle begin their systoles simultaneously, the abnormal auricular complex is superimposed upon the ventricular complex. Consequently, it may be impossible to trace the PASOXYSMAL TACHYCARDIA 251 auricular complex in the composite curve ; it is especially difficult when the form of the auricular complex is unknown. Most paroxysms of tachycardia presenting similar electrocardiograms to those of Fig. 225 are probably due to simultaneous contraction of auricle and ventricle. Simultaneous contraction is occasionally demonstrable, and in some instances the mechanism can be elucidated completely. ■ if ^^i j K |A i 1 o E a": ■"o " (JJ J— n h S > rt a I % 2 m CD H ta ^ (D S£ B •5= S ^ S ^ 03 t3 cS ' ■■3 ■fl -^ 5 B i* lo cs ° S B be 5 -P S ^ ,JB ^ to be o -B : 2 .-J E- fC 'C CQ -p -P B © S > s -p 03 5 jB s en bp cS -c 's- . > "S br .3 o © -s .B ^ ^ £ o © 3 S d .2 o X P +2 o -u o o !h 'S c6 O iU -p i -p o -p CO B a CD ■S G 13 B B 'S CJ T3 1 1 -p CO a © c3 ,B B © > i a -p "5. a o o cS >i 3 f-l ^ -B c >-H -p ^ C i c3 p B B © " © 13 "^ p 'p cS a ©:§ p X •; CO -P P bb s 256 CHAPTER XXI. Fig. 241), and of the similar complexes in A-V rhythm, has been much discussed. A summary will be found in Kahh's monograph (367). When the heart is beating in retrograde fashion and either the ventricular action is very rapid or the A-V bundle is conducting inefficiently, the auricles may fail to respond (432) and a condition of reversed hearts block then manifests itself (Fig. 238) ; the degrees of reversed block are similar to the degrees of forward A-V heart-block. Fig. 237. {Heart, 1909-10, I, 104, Fig. J.) Myocardiographic curves ( F=ventricle, ^ = auricle) and Hiirthle carotid pressure curve ( G). A curve taken from a dog shortly after ligation of the right coronary artery, and showing premature ventricular contractions, single (e') and successive {e/, e"). In two instances a pair of premature contractions occurs, the second awakens an auricular response on each occasion ; the premature auricular beats are marked with asterisks. An instance of the mechanism which precedes tachycardia of ventricular origin. Time in seconds. Clinical examples. As in the case of auricular, so in the case of ventricular extrasystoles, the premature beats may not be isolated but may occur in pairs. Pairs ol ventricular extrasystoles (Fig. 239) are not very uncommon, groups of them QX short paroxysms are less frequent. Paroxysmal tachycardia 257 Fig. 238. {Heart, 1909-10, I, 112, Fig. 7.) Myocardiographlc curves (F= ventricle, yl = auricle) and Hiirthle carotid pressure curve (C). To the left of the index marks one normal and one premature ventricular contraction are shown. To the right (one minute later) the ventricle is in tachycardia and the auricle is responding to each second or to two in three ventricular beats (reversed heart-block). The ventricular rate is approximately 220. The disturbances were the result of obstructing the right coronary artery. From a dog ; time in seconds. Fig. 239. (X ^.) Electrocardiogram from » patient showing three pairs of extrasystoles arising in the ventricle. Time in thirtieths of a second. I have published {4i7, Fig. 132) an example in which a paroxysm of six beats of ventricular origin interrupts the normal rhythm ; it is seen in Fig. 240 ; later the curve shows a soUtary interruption arising in the same focus. Measuring the distance between the clearly defined P deflections in this curve, the positions of the buried auricular summits may be estimated, 268 CHAPTER XXI. f ^ ^ \ -h P R T -f^-^ W S/ V S/ sf-"-^^ sy— ^ ». i I i ( 3/ K 3 2. S s V i'ig. 240. An electrocardiogram from a patient, showing a paroxysm of six beats originating in the right ventricle and a solitary beat of the same kind. The auricular rhythm appears to be undisturbed throughout (see diagram below). The time is in fifths of a second ; it should be noted that the photographic paper was not travelling at a quite uniform rate. for in this example of a ventricular paroxysm the beats do not appear to have been retrograde and the auricular rhythm seems to have been undisturbed. Similar examples have been recorded by Hunt (45), by Cohn (63), and by Vaughan (738). Hart (240) has published very beautiful clinical curves of paroxysms lasting from a few seconds to several minutes. One of his curves, which shows a short paroxysm, is reproduced in Fig. 241. The curve _\j,v^V\yj^ T> -p p p: Fig. 241, {Ha,'L Heart. 1912-13, IV, 128, Fig.. 8.) A paroxysm of tachycardia of ventricular origin. There are eight beats in the paroxysm and an invert auricular complex is seen on the alternate downstrokes of the ventricular complexes. The paroxysm is retrograde but alternate impulses alone awaken auricular responses. Time in fifths of a second. presents a paroxysm of eight beats, arising in the right ventricle. Alternate impulses are retrograde to the auricle and the corresponding anomalous auricular deflections are seen where Dr. Hart has marked them on the alternate downstrokes ; alternate impulses are blocked. Now these examples are clear instances of paroxysms arising in the ventricle ; they are unmistakable because the first beat of the paroxysm PAROXYSMAL TACHtCARDlA 259 has the same relations to the preceding normal rhythm as has a ventricular extr asystole. When records are obtained during the progress of long paroxysms the analysis is far less certain. In Fig. 242 a set of curves is reproduced from a patient who exhibited tachycardial periods having an abrupt onset and ending ; side by side with these is a second series from the same leads (Fig. 243) showing the mechanism two days later and after the paroxysm had subsided. The paroxysm would seem at the first blush to be of ventricular origin, for the ventricular complexes are anomalous and the notch towards the end of the complex in leads // and III probably repre- sents an abnormal auricular systole. Yet this origin is not certain, for an alternative interpretation is equally plausible, namely, that the paroxysm is in reality auricular ; the view being that the ventricle is responding to the preceding auricular impulse after an increased conduction interval, and that the excitation wave takes an aberrant course in the ventricle ; aberration is known to be a frequent phenomenon in patients who are the subjects of paroxysmal tachycardia. It is impossible to decide the exact origin of a paroxysm of the kind illustrated in the present figure unless its first or last beat is recorded. Fig. 242. Fig. 24.3. Fi"^. 242 A series of curves from the three leads during a paroxysm of tachycardia of indeterminate origin. Fig. 24.3. The corresponding series after the resumption of the natural rhythm. Time in thirtieths of a second. An example of an auricular paroxysm, simulating a ventricular paroxysm, is shown in the next figures. Fig. 244, 245 and 246, were taken from a patient who exhibited frequent extrasystoles of auricular origin and short paroxysms of tachycardia. In Fig. 244 a single auricular extrasystole is 260 CHAPTER XXI Fig. 244. Radial and electrocardiographic curve ; a single extrasystole arises in the auricle and deforms the third T deflection ; its impulse fails to reach the ventricle. Time in fifths of a second. Fig. 245. Electrocardiogram from the same patient and showing a similar auricular extrasystole ; the excitation wave of the responding ventricular systole pursues an aberrant course. Time in thirtieths of a second. Fig. 246. Radial and electrocardiographic curves from the same patient, showing the termination of a paroxysm of tachycardia. Comparing the abnormal beats with that of Fig. 245, it is probable that the paroxysm arose in the auricle. Time in fifths of a second. PAROXYSMAL TACHYCARDIA 2(51 shown, but it yields no ventricular response ; the curve displays a conduction defect in the A-V system. In Fig. 245 a similar extrasystole is followed by an anomalous ventricular complex due to conduction of the impulse to the right ventricle only ; the curve displays a conduction defect in the left division of the A-V bundle. Fig. 246 shows the termination of a paroxysm of beats of the same type. Here there is strong presumptive evidence that the paroxysm arose in the auricle, and that the left bundle division failed to conduct while the heart's action was rapid. A still more convincing example of aberration during a continued paroxysm of accelerated heart action is provided by Figs. 247 and 248. As Vs s: As Vs :s Figs. 247 and 248. ( X -|.) Two electrocardiograms from a child, and corresponding explanatory diagrams. The auricle is beating at 290 per minute and the ventricle is responding : the P-R intervals are increased and occasionally a ventricular response is missed. From time to time also, certain of the Purkinje tracts fail to distribute the excitation process normally to the ventricle and the ventricular complexes become anomalous. Time in fifths of a second. These curves were taken from a child in whom the auricular rate approximated 290 per minute. Examining the lower curve first of all, we see a series of seven ventricular beats, each a response to a preceding auricular impulse (P) ; but the P-R interval is gradually increasing as the cycles succeed each other, up to the point where the 8th auricular impulse fails to yield a response. A single ventricular contraction is missed ; response to the s 262 CHAPTER XXI. 9th auricular impulse and to the 10th occurs, but this last-named impulse is followed by an aberrant ventricular complex, the first of a whole series. In the upper curve similar events are shown, but here the series of aberrant beats is fleeting ; when the customary form is resumed, the auricular complexes once more become recognisable. The point to be emphasised is that the last auricular summits fall exactly into place to continue in series with the first auricular summits of the curve. There has .been no disturbance of the auricular rhythm. The series is therefore drawn as complete in the explanatory diagram. The anohaalous ventricular beats in both curves hold the precise time relations to the auricular contractions, recognisable or calculated, which are to be anticipated if all the beats of the ventricle are responses to the auricle.* Thus it appears that in the human subject paroxysms presenting anomalous ventricular complexes may be produced in one of two ways ; these paroxysms either arise in the ventricle itself, or, arising in the auricle, the excitation wave pursues an abnormal ventricular course. * Almost identical curves have been published by White and Stevens (770). Chapter XXII. AURICULAR FLUTTER. In the present chapter a condition is considered which in many respects resembles the paroxysms of tachycardia of auricular origin discussed in Chapter XX. Yet in its fully developed form it differs to such an extent from these paroxysm_s that for the present it has to be considered as a separate affection. It is, however, to be acknowledged that the dividing line between the two mechanisms, the simple paroxysmal and the attack of flutter, cannot be drawn rigidly and that at a future date their separation will prove to be unnecessary if it is shown that the differences between them are purely differences of degree. The facts remain that when a new rhythm arises in the human auricle and its rate approaches and surpasses 300 per minute, the heart chambers respond in a curious fashion, and that certain qualities and associations of this new rhythm seem to be almost peculiar to it. We may divide the simple paroxysm from the attack of flutter arbitrarily by defining the latter as a new rhythm whose rate surpasses 200 and may reach 350 per minute. There is a group of patients in whom the auricles beat constantly or transiently at rates of about 300 per minute ; this is the group which supplies us with our present data, Experimental flutter. When the mammalian auricle is stimulated by successive induction shocks and is responding to each shock, the rapid heart action ends immediately the stimuli cease. But when the stimuli follow each other so rapidly that the auricle is only just capable of responding to each stimulus, or when the auricle is excited by means of a weak faradic current, the accelerated and regular action may continue for a httle while or for considerable periods of time after the stimulation ceases (Fig. 249).* It seems clear that a relation exists between the rate of the auricular responses and the tendency for the acceler- ated action to persist, but it must be confessed that the precise conditions * The end curves of fibrillation which Rothberger and Winterberg {674) describe in detail, and to which they have applied the term flutter, are for the most part not identical with clinical flutter. I say so because the action of the auricle in their curves is not sufiiciently regular. This matter is more fully discussed on page ,'{28, S 'i 264 CHAPTER XXII. '-- — -- _, 1— I---I- -1 -i| __|._--i-^ "" ,_|.-_-..|.^„|_^.|-_-„i-r^ A [~ M K ^ ys • VS f*\ M^ Si 1^ i^ fl/ r» «« a. J ^ H fJ % n^ -:• .:- :,- -. -~ ^ :t- - - - :zL- f^ 7 \ 7\ \ Vf N ? ^ N "-/ ^ b \ r> \ /• S - f N «M| \- ^ > > \ \ J V \ L ' \ A K >L EE ■T ~ w >r ^ ^ EE 1 " TL r — il F - /T "p ^ i 7" 47? 7 1/' /r 7' TR T — 7^ II 7" — =E p nP t, i> *L i L i w>U^ J^ iin j| 1 1^^ k ^4 M« i J rp 7 ■! ' "r ■ J:iy fJ -I _t! ■ ■J tE := J i r" -E: ^- e: Et - « '-" zE EE -~ = ~ — : — — ^ r- == --- -E --"- --- ~- -- — Fig. 249. A cur\'e taken from a dog and illustrating auricular flutter. The auricle was stimulated by means of a w eak faradie current and a greatly accelerated heart action was thereby induced. The rapid action continuing after the end of stimulation is shown. It soon terminates and the natural action of the heart is then resumed. During the period of tachycardia the auricle beats more rapidly than the ventricle ; a condition of partial heart-block is ])resent. ^1 =auricular, and F= ventricular myocardiogram. Time in fifths of a second. which condvice to persistence are still obscure, and that it cannot be brought about at will by stimulation or other known means. Occasionally an extremely rapid and persistent auricular beating appears spontaneously during the course of an experiment and remains unexplained. In a few animals I have been able to induce repeated exacerbations of auricular rate by injecting glyoxyllic acid into the blood-stream ; the disorder resembled flutter (Fig. 250), but, as the auricular waves were not precisely regular, cannot be declared positively to be of this kind. If spontaneous flutter appears in animals it may persist for a long while or may stojD shortly ; when it stops it does so without warning, and after a pause the natural rhythm is resumed. In this respect it resembles clinical paroxysms of simple tachycardia. It differs from these in that the rate is so great that the ventricle usually fails to respond to more than one impulse in two. The term flutter was first used by Mc William {522), in describing the response of the auricle to faradie stimulation. It has been adopted by Jolly and Ritchie {355, 642, 645) and, since the publication of their paper, by other writers to designate a special condition in patients.* Clinical flutter. In man, acceleration of the auricle to a rate of 300 beats per minute or thereabouts is not very uncommon {457) ; generally speaking, it is an * The nature of the response which McWilliam observed cannot positively be identified vvith flutter or fibrillation at the present time, but it was probably coarse fibrillation. AV R ICU LAB F LUTT E R. 265 established condition, thereby differing from simple paroxysms, though in some patients the attacks are transient and repeated {515, 3rd ed.). According to Hertz and Goodhart (304), who first drew attention to the condition in man,* the auricular movements may be visible upon a l SJ=^-'/$/!-r- Fig. 250. (X Y^.) The end of a period of auricular disorder, produced in a dog by injecting glyoxyllic acid into the blood-stream. In the first portion of the curve the auricular waves follow each other at 492 per minute, the ventricle is beating at a rate of approximately 246 per minute. Time in thirtieths of a second. Fig. 251. (Xy%) Acurvefromapatient, for comparison with Fig. 2.50. It shows an auricular rate of 226 per minute, and a ventricular rate of 113 per minute. Time in thirtieths of a second. ■ HM flllfl^ |l I I ■ I ■ J I I I I I I I I ^ I I 'I I I t I I ■ I I tf J ' l '" l /i- ^/V\ ^^^'^o. Vg/fvOxiA Fig. 252. {Heart, 1912-13, IV, 201, Fig. 18.) Venous and arterial curve in a case of auricular flutter. The regular and small auricular waves are distinct in the jugular curve, especially in the long diastoles. The arterial pulse is irregular, the numbers set against the pulse beats represent the numbers of auricular beats corresponding to individual ventricular cycles. fluorescent screen ; they may be audible with the stethoscope, though uncountable because of their rapidity. They are to be recorded from bhe * Ritchie, in 1905, published a record (611) of the auricles raising their rate tu 275 per mmute after an injection of atropine. The record is almost certainly one of flutter, though induced in this manner. 266 CHAPTER XXII jugular veins (304) in many patients, especially in those, in whom the ventric- ular rate is relatively slow and the diastoles consequently long (Fig. 252) ; but when the ventricular action is rapid, and this is the rule rather than the exception, it is usually impossible to disentangle them from the ventricular Fig. 2.53. Venous, electrocardiographic and brachial curves from a, case of auricular flutter in which the auricle beat four times as fast as the ventricle. The venous curve superficially resembles a normal curve, and is apt to be interpreted as consisting of three chief waves to each cycle, namely, a, c and v (as marked to the left). The true interpretation of the events is shown to the right ; all the venous waves are auricular in origin, those which fall in the period of ventricular systole being prominent, while those which fall in diastole are insignificant. Time in fifths of a second. Fig. 254 Electrocardiogram, venous and radial curves from a case of flutter. The auricle is beating at 250 per minute, the ventricles at 125. In the jugular curve one wave a is distinct ; the second falls and fuses with v ; most of the force of this auricular contraction is lost, for it comes at » time when the auriculo-ventricular valves are open. Time in fifths of a second. AURICULAR F LUTT E R. 267 jugular waves which are themselves close set. Even when the ventricular rate is slow, many or most of the auricular waves may be too feeble to distinguish. The waves which fall with ventricular systole and the last waves in the diastole are the most distinct (Fig. 252) ; sometimes the venous curve it««*fct*t*»trtt«jhrt i h>tt«it**xi.ri^ixw^Ttty:4-a;it-L4XixTriTxxixuT-i-ixi-uT44i-i Fig. 257. {Heart, 1911-12, III, 279, Fig. 24). Curves from the three leads in a case of auricular flutter. The auricular rate is 324, the ventricular 162 per minute. Time in thirtieths of a second. new rhythms are ectopic in origin ;* for the auricular complex does not resemble the normal complex in the same patient (Fig. 305, page 327). The rate of the auricular rhythm from minute to minute, from week to week, and even from year to year when long continued, is wonderfully uniform in any given patient (454). Thus in 30 records taken from one patient, at intervals over a period of six months, the auricular rate always lay between 260 and 278 per minute. None of those influences, which produce retardation or acceleration of the normal rhythm, seem potent while the auricles are fluttering. f Exercise, posture (451, 454), effective pressure on one or other * Uncertainty is due to the change which complexes are known to undergo when the heart rate is very greatly accelerated, although the point from which the impulses start remains constant. Of hypotheses which have occurred to me in attempting to explain flutter of the auricle, the origin of the new rhythm in the S-A node is one. The only argument against this hypothesis, so far as I am aware, is the frequent dissimilarity of the complexes while the heart is beating slowly and rapidly. There are several arguments in favour of the hypothesis, but none of these can be regarded as conclusive. They are : — (1.) The tissue of the S-A node is special tissue and on the analogy of the A-V node, new rhythms might be expected to arise from it. Yet none have been recorded, unless indeed flutter has this origin. (2.) When the auricular rate is less extreme and the auricular complexes are not contiguous (as in Fig. 249 and 251) the normal auricular complex is simulated. (3.) When the end of a paroxysm is recorded, the last paroxysmal cycle (returning cycle) is equal in length or shorter (Fig. 249) than cycles of the normal rhythm. In this respect there is a resemblance to the sinus extrasystole (see page 229). (4.) If flutter has its origin in the iS-^ node, the similarity of the auricular complexes from patient to patient would be explained. (See further discussion of the cause of flutter in Chapter XXVIII.) f See, however, the third footnote on the preceding page. 270 CHAPTER XXII. vagus {451, 457. 635) (Pig. 258), the administration of cardiac drugs, each and all appear powerless ; the rate is unperturbed in all these circumstances. In this respect, flutter resembles the simple paroxysm, though in flutter the readings are even more uniform. iiiiiittiiiHuiittttiiuiiiuuiuiiHittiiHiuiiiiitiuimiuttuiniiiiuiittiuiiiiiiiiiumuiiiituttiiiiuimiuuuunui^ Fig. 258. {Heart, 1911-12, 111,279 Fig. 18). (x |.) The effect of pressing upon the right vagus nerve , in a patient who demonstrated flutter and while the ventricle was responding to each fourth auricular impulse. The auricular rhythm remains unaltered, the ventricular rate is notably retarded. Time in thirtieths of a second. Responses of the ventricle. — In a child (Fig. 248, page 261), the ventricle has been found to respond to almost every auricular impulse.* Usually the picture is that of 2 : 1 heart-block, maintained for long periods of time ; but occasionally the ventricular rate rises in these patients to the full auricular rate, a phenomenon generally (515, 3rd ed.), but not always (34), associated with temporary loss of consciousness. On the other hand, the responses may be less frequent, or may become so as a consequence of deficient conduction. A 4 : 1 response is not unusual, neither is a mingled 2 : 1, 4 : 1 response, nor more complex ratios (Fig. 252). Complete heart- block has been recorded and then the rate has been very slow (355). Stimulation of the vagus (451, 457, 635, 644), by decreasing conduction, always slows the ventricle (Fig. 258), though the auricular rhythm remains unaffected ; this effect is obtained in equal degree by both nerves- Digitalis and the allied drugs act in a similar manner ; given in sufficient doses they can be relied upon to produce heart-block or to accentuate a pre-existing block. Exercise or excitement exert the reverse effect. The almost constant association of auricular flutter with some grade of heart-block, and the ease with which higher grades of block may be induced, is attributable almost solely to the extreme acceleration of the auricular rhythm. The rate of the auricular contractions is a prime factor in determining the rates of response. It has been determined experimentally, that if conduction is defective and the auricles are forced to beat more rapidly, the ventricular rate falls ; the ratio is increased not only because the bundle fails to transmit impulses at the faster rate, but because acceleration lowers the conduction power. * First recorded in Fig. 79 of my " Lectures on the Heart " (474). White and Stevens and others have also recorded response to each auricular beat {34, 676, 770), AURICULAR F LUTT E R. 271 The degree of block is always patent in electrocardiographic records, unless the ratio is as 2 : 1 ; in this circumstance, alternate auricular complexes may be so fused with those of the ventricle that they are difficult to decipher (see 736 and redescription of same case in 457) ; but, generally speaking, the continuous activity of the auricle, represented by a zig-zag or undulating line, can be clearly traced throughout the whole curve, though it is broken or deformed by ventricular deflections which are superimposed upon it. When the ventricular responses are irregular, and this is frequently the case, the relative lengths of the ventricular cycles in flutter and in simple A-V block, are governed by the same law. A long pause permits recovery of conduction and at its termination the As-Vs interval is consequently short ; a short pause is succeeded by a response in which the As-Va interval is widened. The arrangement of auricular and ventricular beats and their inter-relation is shown in Fig. 259 and 260. The first two responses in Fig. 259. Fig. 260. Fig. 259 and 260. (Heart, 1912-13, lY, 171, Fig. 38 and 37.) Each photograph shows an electrocardiogram and radial curve. Auricular flutter to which the ventricle is responding irregularly is present. The auricular systoles, responsible for ventricular contractions, are indicated diagrammatically. The numerals placed above the arterial beats affirm the number of auricular contractions to the corresponding ventricular cycles. Time in thirtieths of a second. Fig. 259 are from the auricular beats immediately preceding the ventricular contractions. Not so the third response ; the auricle completes a cycle and, while its impulse is on its way to the ventricle, begins and half accomplishes a second cycle. That two auricular impulses, one of which is to be effective while the other is to be ineffective, may be traveUing towards the ventricle at the same moment, is witnessed to by this figure and with 272 CHAPTER XXII. equal distinctness by Fig. 260. When the auricular rate is extreme and 2 : 1 heart block prevails, the response of the ventricle is not to the auricular systole which comes immediately before it, but to that which falls with the preceding ventricular contraction (457). Arterial curves. — The arterial curves of flutter are often distinctive of the condition when the responses of the ventricle are irregular {457). They illus- trate in a very beautiful manner the principles of what is termed " spacing." In flutter the heart is dominated by an auricular rhythm which is remarkable for its regularity. Whenever an arterial pulsation reaches the wrist, its pos^ition in time is controlled by a given auricular impulse standing in a perfectly regular series. The time interval between an auricular beat and the ventricular response (with its corresponding pulse) is strictly governed by the state of conduction at the instant and this in turn is governed by the period of preceding rest. // two pulse beats are preceded by long pauses or by pauses of exactly equal lengths, however close together or however far apart these beats may stand in a curve, the interval between them, and the interval separating the two auricular contractions responsible for them, are identical in length. And each interval of the kind represents the interval between two adjacent auricular contractions or a precise multiple of it. If the responses of the ventricle are at first to each second and later to each fourth auricular contraction, the later pulse rate is precisely half the original rate (Fig. 261d). Flutter curves may be subdivided into a number of lengths, subtending chosen pulse beats, all of which have simple arithmetic relations to each other. When a whole pulse curve may be sub-divided in this fashion, it is certain that the same dominant rhythm controls it throughout ; and in all such curves phases of irregularity are accurately repeated from time to time. It does not follow that individual pulse cycles will have a simple arithmetical relation to each other ; but those which are preceded and succeeded by equal conduction intervals, will bear this relation to each other and to stretches of the curve which begin and end in the same fashion. If a given pulse cycle is preceded by a shorter As-Vs interval and succeeded by a longer one, the cycle measures more than the corresponding number of auricular cycles ; if it is preceded by a longer and is succeeded by shorter As-Vs interval, the cycle measures less than the corresponding number of auricular cycles. The lengths of the conduction intervals vary inversely with the lengths of the preceding pauses. A long pulse cj'-cle which is preceded by a short one consequently measures less than the corresponding number of auricular cycles ; conversely, a short cycle which is preceded by a long one measures more than the corresponding number of auricular cycles. The analysis of the arterial curves is illustrated by the accompanpng curves, all of which were taken from patients in whom the flutter had been determined electrocardiographically. In practice, the individual cycles in a given curve are compared by measurement, and above each is written the number of auricular cycles to which it is supposed to correspond. The AURICULAR FLUTTER. 273 rules which govern this numbering are simple. Runs of short regular cycles are assumed to be at the same rate as the dominant rhythm, or at a simple multiple of it ; they are numbered alike. The longer pauses of the curve are then compared with these, and they fall into two categories. (a) Those which are of a length which is a precise multiple of the auricular cycles (calculated from the rapid beats) . They are numbered correspondingly (see Fig. 26ld). (b) Those which have not these precise lengths. It may be that a cycle is longer than four calculated auricular cycles and shorter than five. If from a consideration of the pauses and the corresponding conduction intervals, shortening is to be anticipated, the higher number is adopted (Fig. 261a), if lengthening is expected, the lower number is used. Each cycle being thus numbered, the curve is subdivided into sections each containing a numjber of pulse cycles. In a case of flutter many such sections will be found of equal length ; these will be bounded by beats which have equal pauses preceding them and will each contain the same calculated total of auricular cycles (Fig. 261a-«^). '/s Fig. 261, h. Catc 4' i/>^/"' Fig. 261, c. Cast 6' t/'/'*'A" X Z. X. z Fig. 261, rf. Fig. 26]o.-rf. (Heart, 1912-13, IV, 198, Fig. 1-4.) Four arterial curves from cases of auricular flutter. They are published to show the methods adopted in analyses of these curves. The number of auricular cycles to the ventricle cycle is marked above the respective pulse beat. The bracketed portions of any single curve are of equ al length ; the mimber of auricular systoles corresponding to the beats included in a bracket is marked on the bracket. These analyses were checked electrocardiographieally. 274 CHAPTER XXII. pm^mfm^^m^m^t^^m^^m^^it^^^'^^^^'i^^^m^mifmai^^^'i^^^mf'^^mfm^am^m^tr ■#'» » tf' V tfV » » y » ^^^'W * <:iiy/'. J" J.7/IV/X.I' A III I II I I 'I I II Fig. 262. (Bean, 1912-13, IV, 199, Fig. 7.) An arterial curve from a case of flutter. The pulse is very irregular and bears a superficial resemblance to that of auricular fibrillation. Note the xjresence of alternation. An explanatory diagram, in which the relations of auricular contractions and radial upstrokes are illustrated approximately, accompanies this figure. This analysis was checked electrocardrographically. vK/xf\J\r\j\j I iU Ca^e. S- //^'A' Fig. 263. {Heart, 1912-13, IV. 201, Fig. 15.) (x f.) An arterial curve showing the effect of right vagal compression during a period of 4 : 1 heart block. The auricular rate is known to have been unaffected because the ventricular responses, after the long pause, occur at anticipated points. From the same patient as Fig. 258. The control of the irregularity by an unchanging dominant rhythm is well illustrated by vagal stimulation, an example of which is shown in Fig. 263. Upon compressing the vagus a regular 4 : 1 response is disturbed by two pauses, corresponding together to 24 auricular cycles. This corre- spondence is precise ; the ventricular beats, returning after stimulation fall at intervals which continue the original series. The arterial curves in flutter are complicated by two circumstances. First, when the rate of ventricular response is high, alternation occurs (Fig. 262) and, secondly, responses to the auricle are so frequent that many of the pulse beats are weak and resemble extrasystoles. The strength of a pulse beat does not depend to any great extent upon the point of origin of the ventricular contraction which is responsible for it ; it is controlled mainly by the length of pause preceding it. A short pause is succeeded by a weak pulsation whether that pulsation is a response to the auricle, or is premature because it is extrasystoHc ; the two types are indistinguishable in arterial curves ; the relation of the precise lengths of the cycles to one another alone enables their natures to be determined. The relation between flutter and fibrillation, upon which the present therapy of flutter is based, is described and discussed in Chapter XXVII ; the nature of flutter is further discussed in Chapter XXVIII. Chapter xxiii AURICULAR FIBRILLATION (THE CLINICAL CONDITION). In this and the succeeding chapter is an account of a specific clinical condition, which has come to be called " auricular fibrillation." In describing it I purpose to depart from the arrangement hitherto adopted, in which a description of experimental work precedes that of clinical observation. My chief reason for this course is that it permits me to present the subject, as it originally presented itself, namely, as a clinical problem, and to show how the solution gradually unfolded itself and how by persistent enquiry con- vincing evidences were at length obtained as to the true nature of this important disorder of the human heart. As we now recognise it in man it is characterised by a single chief quality, namely, the absence of all signs of normal auricular contraction ; further, it is responsible in the great majority of patients in whom it is found for a complete irregularity of the ventricular action and of the arterial pulse. The irregularity, which is one of the chief features of the condition, is the commonest persistent irregularity exhibited by the human heart, constituting as it does approximately 50 per cent, of all such cases. It will be demonstrated that this disturbance of ventricular rhythm is to be sought in the auricle and attributed to temporary or permanent fibrillation of this chamber. This conclusion and our detailed knowledge of the condition is due to the work of a very large body of men. Fully possessed of the facts, we may now trace the earlier work along two independent paths. Observations were undertaken upon the arterial pulse ; others were carried out upon the venous system ; each series being distinct and for very many years unassociated with the other. The two paths of investigation converged and finally met in modern times. On the one hand, a conspicuously irregular arterial pulse, especially associated with mitral disease in its later stages, was the subject of study by mechanical means from the time of the introduction of the sphygmograph. It is portrayed by Marey (536), Riegel {621, 622), Sommerbrodt [702), and many other writers. It has been termed the " mitral pulse," and has been attributed to " delirium of the heart," amongst other causes. It has passed by the names pulsus arrhythmicus and pulsus irregularis {622), and has been identified, in a classic but obsolete nomenclature, with the adjectives irregularis, incequalis. deficiens and intermittens. On the other hand, a prominent systolic pulsation in the veins of the neck was described {16, 701), and was attributed to tricuspid incompetence. This timing of the venous pulsation was endorsed by Riegel, who obtained the first graphic records of the movement ; but the class of cases in which 276 CHAPTEttXXtlt. such pulsation is essentially found was not isolated, neither was its full significance grasped, until the more exact and more applicable technique of Mackenzie was introduced. It is since the year 1902, when " The Study of the Pulse " was published {501), that chief progress has been made. To Mackenzie we owe the correlation of two phenomena, gross irregularity of the heart and the systolic venous movement, which he has termed the " ventricular form of venous pulse." In the work referred to, this writer first demonstrated their frequent association and ascribed them both to a single underlying condition, namely paralysis of the auricle. A year later, Hering {257, 274), describing the arterial pulse alone, laid more stress upon its characteristics and spoke of it under the title pulsus irregularis perpetuus. The conclusion that it is a condition sui generis came slowly and mainly through the efforts of Mackenzie; it was a conclusion of prime importance, being the first chief step which led to the discovery of the underlying mechanism. Mackenzie laid stress upon the frequent association between irregularity and the ventricular form of venous pulse in 1904 {503, see also 272), and the prominence given to this association in a later paper, based upon 500 patients {510), is to be emphasized. The final demonstration that gross irregularity of the ventricular action is of a specific nature was obtained electrocardiographically (434). In his papers of 1904-5, Mackenzie brought forward several new and important facts and most striking amongst them, in the light of our present knowledge, was evidence that the auricle is active. Formerly regarding the auricle as paralysed, because no sign of its activity could be found, and considering the rhythm to originate in the ventricle (508), he attempted to separate a special group of cases in which auricular activity might be considered probable. Auricular activity was assumed, because the auricle was found hypertrophied at autopsy (507) ; and because cases were observed in which the normal rhythm reasserted itself (503). It is to these papers that we are more especially indebted for the observation that in no case of complete irregularity of the heart are there signs of the normal auricular contraction during diastole and, further, for the first record of cases of this nature, in which it is probable that little dilatation of the right heart and little tricuspid regurgitation is present (see also 725). His earUer view that the condition resulted from auricular distension as a consequence of valve incompetence was at least partially abandoned, and the rhythm was ascribed as the cause rather than as the result of cardiac dilatation. In 1904, Mackenzie postulated the view that in some cases the seat of the rhythm may lie in the junctional fibres lying between auricle and ventricle, and, by conceiving the simultaneous contraction of auricle and ventricle in response to impulses from this single source, attempted to explain the absence of .all sign of normal auricular contraction which he had demonstrated to be one of the chief features of such cases. In 1907-8, Mackenzie (510, 511, 515) adopted the provisional hypothesis of the nodal origin of the rhythm more generally, holding the auriculo- AU RICU LAR FIBRILLATION. 277 ventricular node to be the seat of disturbance in all cases of irregularity found in combination with the systoHc form of venous pulse. He therefore included all such cases under the term "nodal rhythm." This conclusion was abandoned, when undoubted cases of " nodal rhythm " (see Chapter XX) were described, and when other evidences of clinical fibrillation of the auricle were placed on record.* The clinical condition. The radial pvlse curves. The character of the radial pulse curves in complete irregularity of the heart is so striking that it could not, and as we have seen did not, escape early attention. The irregularity is of the most varied description. The pulse may be slow or fast, and the variation in rate is great (30 to 200 per minute). All the beats may be of small excursion ; more commonly there is a haphazard interminghng of forcible and weak contractions, and the latter are often conspicuously dicrotic. The radial pulse is but an indifferent index of the rate of the ventricle ; many beats are not transmitted . The pulse rate may be considerably reduced, either as a result of these abortive contractions or as a consequence of the actual slow speed of the ventricle. The beats may show coupHng over short or long stretches of curve. The fast types are the commonest, and in these the usual rate of the ventricle is approximately double the normal rate (110 to 150). It is usually at these fast rates that the disorderly character of the pulsation is prominent. With the slower rates the irregularity is less evident. In arterial curves the disorder may be recognised by two criteria. First and most important is the absolute character of the arrhythmia. The heart action is never regular, and seldom do two beats of the same character or length occur in succession. In a long curve, it is rare to find any two short sections of tracing which possess even a superficial resemblance to each other. The pauses between the beats bear no relation to one another, and in this feature the irregularity stands in contrast with all other varieties. The second criterion consists in the lack of continued relation between the strength of a beat and the length of the cycle which precedes it. A strong beat may follow a short pause, and a weak beat may succeed a long pause. A few examples of the pictures presented by Dudgeon tracings are given in Fig. 264. They may serve, with the brief notes attached to them, as a guide in recognising the type of case with which we are dealing. They illustrate the main points referred to in the text, but the variety shown is so great that they can scarcely be held even as representative of the irregularities which may occur. Numerous and additional examples are scattered throughout the simultaneous tracings which illustrate this and the succeeding chapter. * An historical account of the discovery of auricular fibrillation is to be found in the British Medical Journal (452). T 278 CHAPTER XXIII The venous pulse curves. " The ventricular form of venous pulse " is a term which expresses the only fixed quaUty manifested by graphic records taken from the jugular veins in these cases. It impUes that all prominent and rapid changes of Fig. 264. ( X -^.-g.) ( Heart, 1909-10, I, 315. Fig. 1.) Radial pulse curves taken with a Dudgeon sphygmograph. The time tracing, which applies to all curves beneath it, is in fifths of a second. The figure illustrates the chief features of complete irregularity of the pvilse. 1 and 2. From a man aged 48, admitted to hospital suffering from mitral stenosis of rheumatic origin and general cardiac dilatation. Enlargement of liver, distended veins and dropsy were present. Irregularity complete and persistent ; apical murmurs early and mid-diastolic ; venous curves of ventricular form ; electro- cardiographic curves of usual type. 3. From a man aged 64, the subject of bronchitis, emphysema and arteriosclerosis. No history of rheumatism. Heart somewhat enlarged to right and left. Heart sounds normal ; S. B. P., 150 mm. Hg..* With the exception of shortness of breath on exertion no signs of heart failure were present. The irregularity disappeared on one occasion for a few days and the pulse regularity was then inter- rupted by auricular extrasystoles. The a-c interval was normal. With the complete irregularity the venous pulse was ventricular in outline, the electrocardiogram was typical. 4. From a man aged 37, suffering from mitral stenosis of rheumatic origin. Heart enlarged to right and left, dyspnoea and slight liver enlargement, no dropsy. Pulse persistently irregular ; ventricular form of venous pulse ; electrocardiograms typical. 5. From a man aged 65, suffering from aneurismal dilatation of the whole thoracic aorta, pulmonary oedema, associated with arterial sclerosis, emphysema and signs of sclerotic kidney. Dropsy and liver enlargement present. Pulse persistently irregular ; ventricular form of venous pulse. Died unexpectedly. * This blood-pressure reading is a measure ot the obliteration pressure of the most forcible beats. Blood-pressure estimations in cases of complete irregularity are extremely unsatis- factory ; the beats force their way through the armlet at widely varying pressures. AVniCULAR FIBRILLATION. 279 volume in the venous cistern fall within the Kmits of ventricular systole. The curves corresponding to the individual heart beats vary in their positions relatively to each other just as do the radial beats. There may be considerable variation in the amphtude of the separate curves in a given case. This variation is certainly not instrumental in origin, for close examination reveals the recurrence of a particular type of curve after a given interval of rest, or during a given phase of respiration. As a general rule, and in a single case, a large venous beat accompanies a large radial beat, but the difference in size from one beat to the next is less in the former than in the latter. A family resemblance between the separate venous beats of a single curve is generally if not always present. The venous curve corresponding to a single heart cycle is generally composed of two or three peaks, and a similar number of dips. The upstroke of the first peak is synchronous with the beginning of the carotid pulsation at the same level of the neck (though it may precede or succeed it slightly). The downstroke of the last peak starts at a point corresponding in the neck to the opening of the tricuspid valves. It is synchronous with the bottom of the downstroke of the cardiogram, or with a point a little later than the bottom of the dicrotic notch on the carotid tracing. The chief depressions follow the first and last peaks and are very variable in degree from case to case. As a general rule it may be said that the shorter the duration of the abnormal mechanism the deeper is the first as compared with the second depression, and that in old-standing cases the dip in the centre of systole is replaced by a larger and fuller complex of systolic peaks. There is a relation between the mean distension of the veins and the swelling of those veins in systole. Thus, in cases of long duration, in which the veins are much dilated, the venous curve is in the form of a prominent systolic plateau. The older conception, that the prominence of the venous pulsation is an index of the degree of tricuspid reflux, is probably not without some foundation. The curves obtained from patients soon after the onset of the disorder and the curves in cases in which the heart's engorgement is but slight are similar, in their systolic phases, to the curves from normal subjects (Fig. 2656). In cases where the heart is engorged and the veins overfilled, the first depression (corresponding to x in normal curves) becomes more or less completely filled (Fig. 265a) until in advanced conditions of venous stasis the curve assumes the plateau form (Fig. 267a and 269) ; in these circumstances it has a flat top and resembles the curve of intraventricular pressure. The transition from one type to the other may be followed from case to case, or in a single case. The flat topped curve often accompanies rapid ventricular action for then overfllling of the veins is the rule.* * The plateau form of venous pulse found in many cases of fibrillation is attributed by Niles and Wiggers (575) to defects in the method of recording and to impact waves from the carotid, To this opinion I cannot consent, although I fully realise, the limitations of the method adopted. The plateau form can be recognised clinically by watching the sustained swellings of the vein in systole or by projecting the shadow of the vein directly on to a recording surface. This form is not at all uncommon. T 2 280 CHAPTER XXIII. Fig. 2G5. Polygraphic curves from cases of clinical fibrillation of the auricles. The arterial curve is in each case grossly irregular ; the venous curves are of the ventricular form. The relation of tlie normal type of curve to the various forms of ventricular venous pulse curve is diagrammatised in Fig. 266. The outlines have been drawn from actual curves. The dotted outline is that of the auricular form of venous pulse. The diagram displays transitions between various types of the ventricular form of venous pulse. But the distension of the auricle and the deformation of the venous curve may take place, while, during the transition, the presystolic auricular contraction is present, or while, during the transition, the co-ordinate systoles of the auricle are suspended. The ventricular form of venous pulse may be conspicuous even in its plateau form, while the normal heart sequence is maintained (see Fig. 100, page 151), and, as we have already seen, the usual or normal systolic portion of the venous pulse may be found (the type with the deep x' depression) and yet the signs of co-ordinate and presystolic auricular contraction may be entirely in abeyance. The two phenomena, the sudden disappearance of all signs of normal auricular activity on the one hand, and venous engorgement on the other, are not necessarily associated ; either may appear without the other. The virtual palsy of the auricle is not responsible for material venous stasis {456), AURICULAR F IBRILLAT ION 281 Fig. 260. [Heart, 1909-10, 1, 320, Fig. 3.) A diagram illustrating the relation of the auricular and ventricular forms of venous pulse, and the commoner variations in shape to which they are liable. The dotted line represents the physiological type, the auricular form of venous pulse. The continuous lines represent the ventricular portions of types of curve met with when either the auricular or ventricular forms of venous piilse are present in patients. though it may presumably aggravate it in a small measure*; conversely, venous stasis is not the cause of the auricular paralysis in these cases. The diagram illustrates the transition of the systohc portion of either auricular or ventricular form of venous pulse curve from the type with perfect to the type with imperfect venous flow. In brief, it may be said that there is but one constant distinguishing mark between auricular and ventricular forms of venous pulse, and it consists of the auriculo-systolic or a wave in the former. Individual cycles of a curve, which is an example of the ventricular form of venous pulse, can be identified as such only by noting the absence of this wave. Certain waves which occur in diastole. It has been said that the ventricular form of venous pulse is composed of waves of which the most prominent and abrupt occur in systole. Diastolic waves are also seen. The auriculo-systolic wave does not occur, but any of the remaining waves seen in the diastoles of normal heart cycles may appear. Thus, it is common to notice a gradual or even abrupt rise of the line as * Gesell (284) believes that the auricles play a greater part (by filling the ventricles) in maintaining arterial pressure than I feel prepared to allow, 282 CHAPTER XXIII. ^ ^J \ / 3 u ■g -o Ti t O ^ 5 C3 o o m ■3 a ■4 i >> O cS .< H ^ rt C8 6(1 fl ^H s ^' S rB (A M O I °J -21, CD f>-' -? .1.1 297 » o ^ -P © fi P CS c ? s ^"-^-l ■a 3 ; t« . ^ .2 ^ V 1: ^ g "5 a .s g 2 £ a ■D t3 :g ■= =s 298 CHAPTER XXIV. the whole cardiac cycle. That they are associated with the actual fibrillation is known, for they terminate the moment the normal heart beat is resumed. Simultaneous electrocardiographic and carotid blood pressure curves are shown in Pig. 279 and 280. In the former of these the dog's heart is acting naturally ; in the second the . auricles are fibrillating. During the period of fibrillation the ventricular action and the pulse are grossly irregular ; the ■ P summits are no longer seen but are replaced by oscillations which roughen the whole curve and distort and obscure the T summits. Another example of fibrillation in the dog is given in Fig. 281, in which the oscillations are particularly coarse and irregular. At times there appears to be unison between the electric oscillations and the irregular movements of a lever attached to the auricle or to the movements in the veins (Fig. 284) ; of curves of this kind I have published several examples, but the unison is not always maintained, and it is both rare and imperfect. If there ever is a perfect unison, it is limited to those instances in which the oscillations are very coarse and tend to be more regular, in fact to instances where there is reason to believe the curves are transitional and not those of fully developed fibrillation. Niles and Wiggers in a recent paper {575), in which they describe the venous curves as they are obtained with very delicate apparatus, find no correspondence between the venous undulations and the electrical oscillations ; in view of these and other observations it seems probable that many of the published examples are to be attributed largely to coincidence.* The oscillations constitute an outstanding feature of both clinical and experimental electrocardiograms. They vary considerably in form, in rate and in extent from case to case and from experiment to experiment. But when the material for selection is abundant it is possible to choose examples of curves which are pictorially ahke. For purposes of pictorial comparison two curves have been selected, and are shown in Fig. 282 and 283. The former is an experimental curve of fibrillation, the second is from a clinical case of complete irregularity of the heart. In experimental curves, as in the clinical curves, the character of the individual ventricular complex is normal. The type is repeated in all its detail whether the auricle is beating normally or is fibrillating. This is well seen in Fig. 279 and 280, in both of which R is notched on the upstroke. The lead adopted is a matter of indifference ; the same statement is always applicable. I have examined the surface of the exposed ventricle, using a number of distinct leads ; the ventricular curve remains unchanged when co-ordinate beating of the auricle gives place to fibrillation, or when the reverse change happens {434) . In the experimental as in the clinical condition the ventricular beats are of supra- ventricular origin. Further, in experimental as in clinical curves, there is the lack of relation between the amplitude of a * A clinioal example in -which approximate correspondence has been supposed to exist between coarse electric waves and vibrations on the venous curve will be found in my book [447, Fig. 170) ; the correspondence is probably never exact {311). AURICULAR FIBRILLATION. 299 Fig. 279. Fig. 280. Fig. 279 and 280. (x ■^■) Simultaneous electrocardiograms and carotid blood pressure curves from a dog. In Fig. 279 the heart is beating normally ; the heart's action is regular and each ventricular contraction is preceded by an auricular beat. In Fig. 280 the auricles are fibrillating ; the ventricular action is grossly irregular and rapid, the blood pressure has fallen away ; the P summits have vanished and are replaced by coarse oscillations which render the whole curve ragged. Note the similarity of the ventricular complexes in the two curves. Time in thirtieths of a second. MT^ U=Hr^ -R -R — -^i 1 \ ' \ •' \ '' '^ ll -J 1^ 1. Ar^ ^ tffe \ i ! > \\ 1, i kg — '. 1 — ^ «'.sp£9s>ras:ir^;3 -i. w. ', s-ii^s.?ri:s' T-BBSs-ssiBszrj' s'.m — ^ as Fig. 281. (x Y^.) Another example of a dog's electrocardiogram taken while the auricles are in a state of fibrillation. In this the chief oscillations are very coarse and irregular. Time in thirtieths of a second. 300 CHAPTER XXIV. Fig. 282. (X ^.) Simultaneous electrocardiogram and carotid blood pressure curve from a dog in which the aiuricles were fibrillating. Time in thirtieths of a second. Fig. 283. (X f.) An electrocardiogram from a patient in whom the pulse was completely irregular. For pictorial comparison with the last figure. Time in fifths of a second. Fig. 284. (x |-.)' Simultaneous venous, femoral and electrocardiograiphic curves from a dog in which the auricles were fibrillating. The signal marks stimulation of the vagus. The ventricular rate became slow during and after vagal stimulation, and during the long diastoles large undulations appeared in the veins and conspicuous oscillations in the electric curve. There is at first an apparent though imperfect unison between them, but at the end of the photograph the oscillations fade away temporarily from the electrocardiograph. Time in fifths of a second. AURICULAR' FIBRILLATION. 301 given R summit and the diastolic pause which precedes the corresponding ventricular beat, and a similar want of relation between the amplitude of R and the strength of the corresponding pulse beat. To some extent these phenomena are due to the clashing of R summits with large oscillations. If the crest of a wave coincides with R, it lifts it ; if the trough coincides, it depresses it ; the currents developed by auricle and ventricle are super- imposed. But differences in the amplitude of R from cycle to cycle are often too great to be explained in this manner ; a further factor is involved which will be discussed in Chapter XXXIII. Reviewing the electrical phenomena, it is manifest that in all their features, the clinical and experimental curves are identical. The close pictorial resemblance between selected curves from the two series, the presence of the same characteristics and of those same variations which are displayed by a detailed analysis, carries complete conviction that the clinical and experimental curves are to be ascribed to a similar disturbance of the heart's mechanism. The harmony between the records obtained in complete irregularity of the heart in man and in experimental auricular fibrillation. The comparison between complete irregularity of the heart in man, and auricular fibrillation in the dog is now ended. It has been demonstrated that the clinical and experimental conditions resemble each other in every respect in which they have been investigated. The observations are summarised in the following table, which consists of a systematic list of the features presented in common. The radial curves. 1. Rate increased as compared with the no mal. 2. Presence of absolute irregularity. 3. Absence of a constant relation between the strength of a beat and the preceding pause. 4. Presence of extreme variation in the force of the ventricular beats, m,any of which fail to reach the arteries. The venous curves. 1. Presence of the ventricular form of venous pulse {the absence of " a " waves). 2. Presence of rapid diastolic undulations of venous volume when the heart beats slowly. 302 CHAPTER XXIV. The electrocardiograms. 1. Presence of the supra-ventricular type of ventricular complex. 2. Absence of relation between the height of "i? " and the length of the preceding diastole. Absence of relation between the height of "72 " and of the corresponding pulse excursion. 3. Absence of "P" summits. 4. 'Presence of characteristic oscillations which are shown to be generated in the auricular portion of the heart. The persistence of these oscillations throughout the whole cardiac cycle. Although auricular fibrillation is now universally recognised as the cause of complete irregularity of the ventricle in man, and although no one who has read and understood the evidence now doubts this conclusion, there was a time not long ago when many hesitated to accept it. The disinclination to believe in what was at first hypothesis is attribut- able to the frequence with which the human heart is affected by this disorder, by the tradition, now exploded, that irregularity of the heart is the result of distension of its walls, and by the occasional persistence of the irregularity for many years. It seemed inconceivable to those imbued with the mechanical theory of heart disease that so profound a disturbance of the auricular function could persist for many years. A case had been recorded (349) in which the irregularity had lasted for five and a half years and Mackenzie had watched cases unchanged for even double this period. Recently indeed, a case in which the disorder had in all probability persisted for 32 years has been placed on record (249). At the time of which I am speaking (1909-1910) the hypothesis was received with scepticism and it was necessary to collect and record all the available evidence. The proof as I gave it was written much as I have written it in these two chapters and was to my mind conclusive. But there still remained those who doubted, and it was for this reason that I attempted to obtain a further evidence which would silence all criticism. I turned to the lower animals in the hope of discovering the fibrillation of the auricles as a manifestation of disease in them. After a while I was fortunate in finding what I desired. Auricular fibrillation in the horse. As a very rare affection, complete irregularity of the heart is manifested by horses, who are the subjects of cardiac disease (434, 450) . In these animals, as in man, it is often associated with breathlessness, rapid exhaustion and other signs of distress, and, as in man, it may be accompanied by enlarge- ment of the veins, by ascites and by general dropsy. The ventricular rate AURICULAR FIBRILLATION . 303 is increased from the normal of 30 or 40 per minute up to 60 or 70 or even to as much as 150 per minute. The disorderly action is characteristic in that the lengths of the cycles are quite lacking in uniformity. I have had the opportunity of examining six horses with this disorder of the heart. Curves obtained from the arteries and veins are indistinguishable from others obtained from human patients ; the venous curve shows ventricular waves, c and v; the diastoles are devoid of a waves, but from time to time display the small undulations which are so typical of the condition. I have also obtained electrocardiograms from one animal; and in these tho usual features were to be observed. The irregular ventricular complexes were all of the supra-ventricular type, P summits were absent, oscillations were present, though ill-defined. In one of these animals I was able through the kindness of General F. Smith and Colonel Blenkinsop to make the desired observations. The animal, which presented the characteristic symptomatology and a gross irregularity of the ventricular beat, was thrown and shot through the brain. The right side of the chest was quickly opened and an excellent view of the heart was obtained, the ventricle continuing to beat rapidly and irregularly. At first no intrinsic movement could be seen in the auricle, its walls seemed fixed in diastole, but close inspection revealed the fibrillary movements. The epicardium covering the right auricle is sufficiently thick to be somewhat opaque in the horse and fine movements in the underlying muscle are not easy to see. But the independent movements of the light reflections in the little surface ridges, especially over the appendix, were perfectly clear and were demonstrated to and recognised by Colonel Blenkinsop and others who were present during these observations. Thus came the final and ocular proof that experimental fibrillation as we know it, occurs as a manifestation of disease. Much of the evidence in these chapters may now seem redundant; sufficient evidence to carry conviction might be set forth in a few paragraphs. My decision to preserve these chapters in a form similar to that in which they appeared in " The Mechanism of the Heart Beat " has come not only from a belief that, where a conclusion of first consequence is at stake, the reasons for that conclusion should be set forth fully, buo also in the hope that these chapters may serve to illustrate how persistent study may build up and consolidate our knowledge of phenomena at first obscure — obscure because they manifest themselves in the deep seated and hidden tissues of the human bodv. Chapter XXV. AURICULAR FIBRILLATION, HEART-BLOCK AND EXTR ASYSTOLES. Fibrillation and heart-block. In the healthy heart fibrillation of the auricles drives the ventricles at a greatly enhanced rate. In experiments in which the heart is in a fresh condition the ventricular rate rises to double or even treble the normal. But under some experimental, and under many clinical, conditions the ventricular rate may be normal or may even be reduced notably. Thus, in patients the ventricular rate may lie anywhere between 30 and 200 per minute ; the rate of the ventricle is controlled exclusively by the state of the conducting tissues, so far as we know at the present time. That the impulses of fibrillation are transmitted through the A-V bundle, as are the normal impulses, was first shown by Fredericq (187). He found that, when the bundle is broken, the auricle though continuing to fibrillate no longer affects the ventricle. His experiment I have often repeated and substantiated. Damage of the A-V bundle, conduction being abolished completely or partially, has similar effects while the auricles are fibrillating and while they are beating at an accelerated rate. When the function of this bundle is suppressed altogether, the ventricle develops its own regular rhythm. When its capacity to conduct is merely impaired, the rate of ventricular response is lowered, and the degree to which the rate is lowered goes hand in hand with the degree of damage. But the rate at which successive impulses impinge upon the junctional tissues in fibrillation favours the display of block, as does a simple acceleration of the auricle. A slight impairment of conduction is sufficient materially to reduce the ventricular rate. If we examine all the known ways of reducing the ventricular rate while the auricles fibrillate, we shall find that heart-block may always be ascribed as the cause. The same cause is alone found to reduce the ventricular rate in (Jinioal cases. The chief facts are set forth under the succeeding subheadings. Complete dissociation and auricular fibrillation. — Transection of the A-V bundle while the auricles fibrillate is followed by standstill of the ventricle, and subsequently a slow and regular idio-ventricular rhythm is FIBRILLATION AND HEART-BLOCK. 305 developed. If complete heart-block is first induced, either by section or by asphyxiation, and the auricles are then caused to fibrillate, the regular ventricular rhythm persists unaltered. When I first suggested that slow ventricular action in clinical cases of fibrillation is the result of heart-block {431, 434, 479), I cited a notable patient. A S3'philitic subject, he suffered from the Adams-Stokes syndrome. The ventricular action ranged constantly about 30 beats per minute and was usually regular (Fig. 285). From time to time ventricular extrasystoles were seen (Fig. 286) and these cycles were of the same lengths as the regular cycles. During his fits the ventricle ceased to beat. The ventricle behaved in every respect as it does in simple and complete heart-block. Yet the picture Fig. 285. (X rh-) (Quart Journ. Med., 1909-10, III, 273, Fig. 1.) Polygraphic curve from a case of complete heart-block and auricular fibrillation. The venous pulse is of the ventricular form ; the radial pulse is slow and regular. Fig. 286. (X -f.) (Quart. Journ. Med., 1909-10, III, 273, Fig. 2.) Polygraphic curve from the same case as Fig. 285. The ventricle is beating irregularly because ventricular extra- systoles disturb its rhythm. The returning cycles are of the same length as the initial cycles. differed from simple dissociation in that the venous pulse (Fig. 285) was of the ventricular form, showing no sign of auricular contractions. In electro- cardiograms no P summits were to be discovered, but replacing them were the usual oscillations in characteristic form. The heart, obtained at a later date {73), exhibited a complete break in the A-V bundle by an old-standing syphilitic lesion. A similar example has more recently been reported with post-mortem evidence {192). Other cases* have been collected and my * Though lesions of the A-V system have naturally not always been discovered (338). 306 C H APT ER XXV . original conclusion that heart-block is responsible for a slow ventricular response, when fibrillation is associated with a slow action of the ventricle, has now received a wide measure of support and confirmation {169, 192, 221, 350, 368). In some patients complete heart-block is induced by heavy doses of digitalis (Chapter XIV). I have known the ventricle to become slow and perfectly regular under the influence of this drug in more than one case of fibrillation (106, 451, 516, 717). In such instances the ventricular rate is relatively high,* usually from 40 to 50 per minute, but occasionally as high as 90 per minute (Fig. 287). Fig. 287. - (Heart, 1911-12, 111,' 279, Fig. l2.) From a case of auricular fibrillation, under treatment with digitalis: The. fibrillation is evidenced by the oscillations /, / apd by the absence of P summits. The ventricular action is regular because complete heart- block is present. The rate is exceptional for a ventricle in complete heart-block, being approximately 90 per minute. Time in thirtieths of a second. Partial heart-block and auricular fibrillation. — A simple method of producing partial heart-block of different grades in experiment, is by asphyxiation. If a cat is asphyxiated, while the auricles fibrillate, the ventricular responses are soon reduced in number, and continue gradually to be reduced if asphyxia is maintained. The rate of the ventricle is governed by the degree of block {439). Amongst clinical cases there is an interesting group. The patients to whom I refer present heart-block and later develop fibrillation of the auricles. I have drawn attention {434) to a description under the term " nodal bradycardia " of such a case by Mackenzie {516) ; and other examples have since been pubhshed {169, 228). In all these patients fibrillation has been associated with a slow ventricular action. The post-mortem evidence in cases of slow ventricular action is not of itself convincing. Many hearts from patients with slow and irregular ventricular action have been examined, and lesions of the ^-F bundle of vary- ing severity have been found {56, 103, 170, 192) ; but in respect of these it cannot be said, any more than it can be said of straightforward partial block, that the extent of the visible lesion goes hand in hand with the degree of slowing ; neither is such a precise relation to be expected. * Presumably because the block occurs at a high level of the A-V node, and because the ventricle is governed by a rhythm formed in the A-V node. FIBRILLATION AND HEART-BLOCK. 307 The effect of vagal stimulation upon auricular fibrillation,. — Stimulation of the vagus, while the experimental auricle is fibrillating has one of two chief effects (397, 597, 647, 650). If the fibrillation is of short standing, it may suppress it. If it is of longer duration, it usually fails to do so. The inhibitory impulses are accompanied by a certain alteration in the auricular mechanism itself, the movements in the walls of the auricle appear to be more finely subdivided, suggesting the establishment of hues of block in the musculature. But the most conspicuous effect is slowing of the ventricle (Fig. 288 and 289). Perfect clinical examples of this retardation have been published by Wenckebach (763). He has shown that firm pressure upon the Fig. 288. (x J.) Myocardiographie curves (from auricle and ventricle) and carotid pressure curve from a dog. A.F., the auricular curve, while the auricle is fibrillating. V.I., the ventricular curve, showing the irregularity of the response to the fibrillating auricle, the standstill of the ventricle produced by vagal stimulation and the subsequent recovery. Time in seconds. carotid sheath may cause conspicuous slowing of the ventricle in man when the auricles are fibrillating (Fig. 290). This observation has been repeatedly confirmed (447 (Fig. 187) and 653). The manner in which the rate is reduced is not in doubt, for if in an experiment the normal auricular action is resumed before the vagal effect has subsided, simple heart-block is seen (434). The vagal effects illustrate the rule that a slow ventricular rate, when the auricle is fibrillating, may be ascribed to decreased conduction power of the A-V tissues. The final illustration is that provided by digitalis and its allies. 308 CHAPTER XX V . intttitittimitmimm Fig. 289 { X A.) Venous, femoral and electrocardiograjjhic curves from a dog, during the progress of auricular fibrillation. The vagal stimulation is indicated by the signal. The ventricular action, at first rapid, is conspicuously reduced ; the oscillations in the veins and in the electric curve are reduced in size. Time in fifths of a second. Fig. 290. (xf) Arterial and electrocardiographic curves froma patient. Fibrillation of the auricles is present. The ventricular action, at first rapid, has been consi^icuously reduced by pressing upon the carotid sheath. The vagus was compressed at about the period indicated by the third time line. Time in fifths of a second. Digitalis in auricular fibrillation. — The potency of digitalis in slowing the ventricular pulsations in the condition now termed auricular fibrillation was first demonstrated by Mackenzie {507, 508), and has since been confirmed abundantly (95, 106, 168, 517, 700, 771). The action of this drug in this respect, so I have concluded, is to be ascribed to heart-block (431, 439) ; this conclusion first suggested itself to me because I was aware that d'igitaHs induces heart-block when the rhythm is normal. It is in fibrillation that the pace of the ventricle may often be set with nicety by regulating the dose of the drug. In these circumstances the ventricle slows gradually but eventually profoundly (Fig. 291), while it continues to beat irregularly in most cases. Exceptionally, as has been pointed out, complete block may develop and the ventricle then beats regularly. FIBRILLATION AND HEART-BLOCK. 309 WWMiiiiiiiiUiiiHHiiBWiM Fig. 291. Fibrillation of the auricles from a patient under treatment with digitalis; the ventricular action is very slow. Time in thirtiotlis of a second. My conclusion that digitalis produces heart-block in fibrillation and that the slowing is due to this cause is now generally accepted. But the exact manner in which this slowing is brought about has been much discussed. In a few patients atropinisation of the heart, which has been retarded by digitahs, produces little or no acceleration ; in these cases there is presumptive evidence that the vagi are not responsible for the slow rate. In most patients the slow heart action is accelerated, often conspicuously, by atropine(9.5, 517, 700). Thus, in these patients there is httle doubt that the actual slow rate under digitalis is in part due to vagal tone ; but whether as an effect of digitalis this tone is greater than the normal vagal tone is debated. To be certain that there is such an effect, it has to be shown that the rise of ventricular rate following atropine is to the same level when atropine is administered before and after the heart is brought under digitalis ; in which case the whole of the digitalis slowing might be attributed to the vagus. This rise to a given level in the two circumstances has only been observed in exceptional cases ; as a prevailing rule, the rate reached under atropine in the non-digitalised heart is greater, usually much greater, than that reached by the atropinised and digitalised heart. In such cases, as Cushny {91, 95) has rightly pointed out, a conclusion that the slowing is vagal is not warranted. Cushny (91) emphasises the direct action of digitahs, which he believes to be peculiar to hearts affected by fibrillation ; but that is not strictly the case, for some cases of simple heart-block appear to be unrelieved by atropine. The direct action of digitahs seems to be absent when the heart is quite normal, as in experimental animals ; in these digitahs slowing is abolished by section of the vagi or by atropinisation.* Cushny shows that, if the perfused heart is used, digitahs exerts also a direct influence upon the bundle, and concludes that malnutrition is responsible for this effect in experiment and also in certain instances of disease. It is difficult to beheve that an action on the vagi is not at least in part responsible for the actual drop of the ventricular rate when digitahs is employed in auricular fibriUation ; for the * Halsey (239) has shown that in dogs under digitalis or strophanthus full doses of atropine or amyl nitrite will abolish the block. Curiously, small doses of either drug will produce a measure of block, whether the heart has been submitted to digitalis previously or not. This is but one example of the now well-known reversed action of atropine in small and large doses. X 3.]0 CHAPTER XXV. powerful slowing effect of the vagi in this disorder is well-recognised, and digitalis is known to increase vagal tone in other patients and in experiment. If digitalis does not in part slow the ventricle through the vagi, but only by a direct action on the bundle, then it is necessary to assume that the pharmacological effect of digitalis upon the vagus is absent in cases which present this form of disordered heart action. Provisionally, therefore, it seems safer to conclude that the digitalis effect is twofold, that it acts in part directly, in part indirectly, and that its relative power to act through one or other channel is variable in different patients. At the same time, a review of all observations emphasises the direct action. Similar slowing is obtained with the allied drugs, strophanthus (5) and squills {124, 735, 788) ; both drugs produce heart-block, but whether through the vagus or directly is still uncertain. The known causes of ventricular slowing may be summed up by repeating that they are always such as induce heart-block. Naturally there are clinical cases in which heart-block cannot be directly proved. For instance, in those in whom the auricles are always fibrillating and the ventricular action is constantly slow ; but in these also we are justified in at present assuming that heart-block is responsible for the low ventricular rate. Fibrillation and ventricular extrasy stoles . According to current hypothesis fibrillating auricles shower impulses upon the ventricle indiscriminately ; it is in this way that the complete irregularity of the ventricular response is explained. The ventricular cycles are of very variable lengths. If it could be shown that many cycles are of equal length — so many that their equality could not be ascribed to coincidence — then either the hypothesis would become untenable or some complicating factor would have to be proved. Now we have seen that while the auricles are fibrillating, the ventricular action may be perfectly regular; in this circumstance, the rate is slow because heart-block is present. But ventricular cycles of constant length sometimes appear in another form. The condition has been termed, some- what loosely, " digitahs coupling." When the auricles are fibrillating and the heart comes fully under the influence of digitalis, the ventricular action is retarded. It is as an accompaniment of this retardation that coupling is usually seen {507). The irregularity is singular and distinctive (Fig. 292). The beats of the ventricle occur in pairs and the short cycles are almost or quite constant in length. Variation is still seen in the long cycles ; this is to be expected and is consistent with our view of fibrillation, for the ventricular systoles which terminate the long cycles, and govern their lengths, may be attributed to haphazard auricular impulses. Not so the beats which terminate the short cycles {i.e., the second beats of the pairs), for the pauses which precede the§e are of uniform duration. If our hypothesis is sound, FIBMILLATfON AND HEART -BLOCK. 311 these beats must be of different origin. Electrocardiograms show them to be so as I was able to demonstrate (434, 436) some years ago (see also 107). Fibrillation of the auricle is shown in Fig. 293 ; the majority of the ventricular complexes conform to the supraventricular type. But the curve opens with a period of " coupling " ; the second beat of each couple has the shape of an extrasystole arising in the left ventricle. That such is the origin of these anomalous systoles is, I think, beyond doubt, for digitalis in full doses is known to provoke extrasystoles. The pairing of the beats is similar to the -JW^^/— J" Fig. 292. Veuous and radial curves showing so-callod '" digitalis coupling " The arterial beats are in pairs ; the interval between the first anfl second beat of the couple is uniforjn ; the pauses following the pairs of beats are variable in length. The venous curve is of the ventricular form. -iU TTT \-n hn 1- 7. b n^-n i ^ / ./ ' L — y Fig. 293. An electrocardiogram showing a period of ventricular coupling ; later the ventricular complexes are quite irregularly placed and are then all of the supraventricular type. Time in thirtieths of a second. pairing which manifests itself when extrasystoles disturb a normal rhythm. The reason why the short cycles during the period of coupKng are not variable in length is that the second beat of the pair is not a response to an auricular impulse.* Coupling of the form described may appear independently of digitahs, but whether it is the result of overdosage with this drug (Fig. 294) or whether it occurs under conditions less clearly defined, the electro- * Very occasionally the ventricular complex is constant in form throughout » period of " coupling " ; this fact is not out of harmony with my conclusion ; in such cases the second beats of the pairs may be regarded as extrasystoles arising in the junctional tissues. The cause of " coupling " is further discussed in Chapter XXVIII. X 2 312 CHAPTER XXV . Fig. 294. (x f.) From a case of auricular fibrillation on large doses of digitalis. It shows a coupling of ventricular beats which speaks of over-dosage. The first complex in each couple is of the supraventricular type ; the second is of different form ; these last contractions are premature and originate in the ventricle. Time in thirtieths of a second. Fig. 295. ( X j^Q.) Two clinical curves showing fibrillation of the auricles. The responses of the ventricle are for the most part to auricular impulses ; but from time to time an isolated extrasystole appears ; such extrasystoles arise in the ventricle ; it will be noticed that they fall early in the preceding diastole. Time in thirtieths of a second. cardiograms present the same pictures. Single extrasystoles frequently complicate auricular fibrillation [275, 434, 436) (Fig. 295) ; the same beats may appear in short runs. In some patients all the extrasystoles are from one focus ; in other patients they may be derived from several sources. The forms of extrasystoles in a given patient are constant from month to month. In patients who suffer from them, if fibrillation is subsequently acquired, the extrasystolic irregularity persists ; the anomalous beats may be shown in these circumstances to be of the same outKne before and after fibrillation is acquired (441), an observation which places their extrasystohc origin during the period of fibrillation beyond reasonable doubt. Chapter XXVI. VENTRICULAR FIBRILLATION. It was found by Hoffa and Ludwig (320), in 1850, that the normal beat of the ventricle is abolished by applying to this chamber strong constant or faradic currents. The ventricle so assaulted exhibits very rapid irregular movements of minute amplitude ; co-ordinate contraction of the fibres seems to be lost ; the heart at first shrinks, but soon swells and no longer expels its contents. The condition outlasts stimulation and usually persists until all movement in the heart ceases. The phenomena presented by the fibrillating ventricle have been described on many occasions and are familiar to numerous workers (520). During the final stage, when the ventricle is ballooned and the arterial pressure has for long sunk to zero, the whole wall of the chamber is convulsed by miniite quiverings ; over the whole superficies twitchings of the muscle are visible. If attention is concentrated upon a small area, little waves of contraction may be perceived to pass along the muscle bands for a short distance, the course of the wave altering from moment to moment, but always dying out abruptly and often at a given line. It seems from inspection alone that while one small area is contracting, the adjacent area may be in a state of relaxation. At the same time a more general movement is often perceived ; it is as though a ring of tissue surrounding the cavities contracted more firmly, and as if this ring of contraction travelled from apex towards the base or base towards the apex. Yet the inconstancy of the picture from moment to moment perplexes the observer and the intricacy of the contractions and relaxations defies analysis. This fully developed condition is properly referred to as one of ventricular fibrillation. Fibrillation may be induced, not only by the application of electric currents, but by other irritants, thermal, chemical and mechanical. Especially is this the case when the ventricle is in an irritable state ; according to MoWilliam (522), " weak faradic currents, a touch with a hot wire, a mere scratch with a pin ... or even slight pressure with the finger, are each sufficient at such times to excite the fibrillar contraction." Spontaneous fibrillation, by which I mean the onset of fibrillation with no apparent cause, 314 CHAPTER XXVI. not infrequently terminates experiments on the heart. Fibrillation usually follows the sudden occlusion of a coronary artery or of a main branch ; a fact known for a long while to many (78, 609). The injection of large doses of digitahs (85), potassium chloride (522, 546), and many other toxic bodies will cause it. (a) (6) Fig.296. {Levy. Heart, 1912-13, IV, 335, Fig. 7.) Carotid curves of the carotid pulse in a cat. The first curve shows the effect of stimulating the right stellate ganglion of the sympathetic while the animal was under the influence of 2 per cent, chloroform vapour. The normal rhythm is replaced by an irregular tachycardia. The second curve shows the effect of stimulation under 0-5 per cent, vapour. A similar tachycardia is produced but it terminates in fibrillation of the ventricles, the blood pressure falling to zero. The white bands are the signals of stimulation. Time in seconds. Still more notable, because of their direct practical bearing, are the experiments of Levy (421-423, 425, 426). Working upon cats, he has shown that the heart is extremely susceptible to low percentages of chloroform vapour inhaled, and that when it is under the influence of this drug, ventricular irritability is enhanced to a precarious degree. The injection of minute doses of adrenalin, stimulation of the sympathetic (Pig. 296),* section of * The relation of ventricular fibrillation to nervous impulses is a subject which it will be convenient to defer to Chapter XXIX. VENTRICULAR FIBRILLATION . 315 the vagi, and stimulation of a sensory nerve, are each capable of forcing the sensitised ventricle into fibrillation ; and when the conditions are favourable, interferences, so inconspicuous as to escape observation,* may precipitate the state. That the fibrillation of the ventricle is not due to destruction or paralysis of a co-ordinating centre, as Kronecker and Schmey {396) imaginedf has been forcibly argued by McWilHam (522) and, in the light of present knowledge, this hypothesis has been almost universally abandoned [407). The whole excised ventricle, or any considerable part of it which is cut a*ay, may be made to beat co-ordinately or in a fibrillating fashion ; fibrillation is primarily due, as McWilham states, to changes within the ventricles themselves and is not caused, when induced by faradisation, by abnormal impulses transmitted through the nerves from other parts or by direct stimulation of surface nerves ; the state of fibrillation is conveyed from one part of the ventricle to another and will cross the bridges between zig-zag incisions which penetrate the walls. Transmission from one part to another is determined solely by there being sufficient muscular union (211, 459). Out- standing features of the condition are the complexity and persistency of the movement. Mc William concludes that it is related to the complex arrange- ment of the muscular fibres in the ventricular walls, for the complexity of movement is greater in the heart of adults than it is in the young, it is fully developed in the mammal, whereas in cold-blooded animals the response to faradic stimulation is far less complex in nature (24, 407). Garrey (211) expresses the relation in more simple, and probably more correct, terms ; he believes that the proclivity to fibrillation is related to the mass of tissue involved, because he finds it cannot be induced in small pieces of ventricular tissue and because the smaller the mass of tissue the less persistent is the con- dition. All observers are agreed that recovery may take place sometimes ; it occurs occasionally in the dog, frequently in the cat, and it is the rule in the rat and mouse (237, 522) ; the quivering movement ceases and after a brief quiescence, comparable to the pausej following a paroxysm of tachycardia (792), the heart will beat again in a normal fashion. Recovery is distinctly controlled by the size of the organ and possibly, therefore, by the degree of the original derangement. It will be convenient to defer a discussion of the ultimate nature of fibrillation to a succeeding chapter, in which this and several closely allied problems may be treated together, and meanwhile to describe the records of fibrillation and to note the events which lead up to that state. * Indirect stimulation such as happens when a fresh dose .of stronger vapour is inhaled may be cited ; as may also the simple act of lifting the animal. ■[■ Kronecker and Schmey came to this conclusion when they saw fibrillation follow the introduction (often repeated introduction) of a needle into the ventricular septum at a given level. The same result is often to be obtained when other regions of the ventricle are similarly treated and by many less violent procedures. t The pause is shorter than is the coihpensatory pause because impulses from the fibrillating ventricle are transmitted back to the auricle. 316 CHAPTER XXVI Records of fibrillation. When fibrillation of the ventricle is indaced by faradic stimulation, gradually increasing in strength or maintained ; when it is brought about by ligation of a coronary artery (432), by the administration of chloroform (426) or by the injection of digitalis ; or when it is precipitated by nerve stimulation or the injection of adrenalin, the essential events are the same. Fibrillation in its fully developed form does not come abruptly ; it is foreshadowed by a distinct train of events. The premonitory period may be long or short ; on occasion the steps are incomplete, but when complete they are definitely ordered. At first the heart's regularity is disturbed by premature contractions MdMM«M*y tJjUk: ^ il3- S . r^A;^4;;*f^'i^^---A -] '-V¥ v.E.*^i»- ^mMMm^^mSSimS^^M MMMMHHMMMSiBH IHIiiMi ■^■NH Fig. 297. Three electrocardiograms, illustrating the disorders produced in the cat's heart by the inhalation of low percentages of chloroform vapour. The first curve shows the normal rhythm, mterrupted by ventricular extrasystoles arising in at least three separate foci The second curve shows the fully developed tachycardia ; the beats are not uniform in shape their origin is variable, neither are they quite regular in incidence. The third curve exhibits the mechanism which immediately precedes fully developed fibrillation of the ventricles This disorder probably resembles closely what is termed flutter in the case of the auricle" Time in thirtieths of a second. coming from a ventricular focus,; these increase in number and eventually form groups ; other foci of activity are often added (Fig. 297) The heart accelerates in response to these new impulses and is eventually controlled by them exclusively ; a tachycardia originating in a single ventricular focus or in several ventricular foci, becomes estabhshed. It may be gradually, it may be more suddenly, that the rate of movement increases till the circulation can be maintained no longer ; the arterial pressure rapidly sinks to zero and the process develops further At this VENTRiaULAR FIBRILLATION. 317 -ZiCArofn^ -."S-- "r:r:tr1 L/-' t ". ". ■ .'^~£ Li: . 1 -T=Z kii-L 1^=^. r-ri Fig. 298. Fig. 299. Fig. 298 and 299. (xf .) Simultaneous carotid blood pressure curve, electrocardiogram and ventricular ( Vs) and auricular [As) myooardiograms from a dog. Fig. 298 was taken before and Fig. 299 shortly after faradising the ventricle ; the second curves show the early stage of ventricular fibrillation. In Fig. 299, the electrocardiogram displays large oscillations, following each other at a rate of 375 per minute. The excursions are not quite uniform in shape or sequence. The blood pressure (the carotid curve may be measured to the common base line of the photographs) has fallen almost to zero, though little undulations still show upon it. The auricle (As) continues to beat, its rate being enhanced ; its beats are not regular. The ventricular myocardiogram shows only a few coarse movements, which correspond to the changes in the muscle ; this lever lies in a position of not quite full systole. Time in fifths of a second. stage the movement witnessed in the ventricular wall is a rapid undulation, almost regular, but of small amplitude. The ventricle is not dilated but, on the contrary, is diminished in volume. If stimulation is withdrawn at this stage the action continues for a little while and ends in recovery, or progresses as it would if stimulation were continued. The electrocardiogram 318 CHAPTER XXVI. shows a movement of the fibre exceeding in ampUtude that of the natural ventricular beat ; the individual undulations (Figs. 297, last curve, and 299) are almost regular in incidence, but often varying in amplitude in a phasic way. The auricle beats with increased frequency and irregularity (208, 792) ; it is responding to the ventricle, for its movement becomes regular if the bundle is divided [76). The movements of the ventricle are scarcely strong enough to be recorded, but a few general or jerky movements may be seen (Fig. 299). The further change to the fully developed Kg. 300. (x f.) Similar curves from another animal and showing the final and full picture of veri- tricular fibrillation. The levers recording movements of auricular and ventricular mxiscle and carotid blood pressure are motionless. The electrocardiogram consists of very large and irregular oscillations. This type of curve corresponds to the phase of ventricular dilatation when small flickering movements are visible on the surface. Time in thirtieths of a second. condition, which in my view alone deserves the term fibrillation (see Chapter XXVIII), is so gradual that its onset cannot be signalled. The heart gradually distends and all movements of the auricular and ventricular levers cease (Fig. 300) ; the blood pressure remains at zero ; the appearance of the ventricular surface is that described in the first paragraph of this chapter ; the movements of the galvanometric string assume the characteristic irregularity [358) illustrated by Fig. 300. Clinical fibrillation. It is to be remarked that all the major forms of cardiac disturbance which have been produced experimentally are also known to occur clinically. Until recently a single condition, fibrillation of the ventricle, formed an exception to this statement. If fibrillation of the auricles is so frequent in the human subject, it is natural to ask why a similar disturbance of the ventricle is so rare ? Its rarity is presumably more apparent than real ; for fibrillation of the ventricles spells death.* We have now the strongest * This is the rule in the dog : recovery is rare ; it is almost certainly the rule in man, though an exception has been recorded by Robinson and Bredeck (652). VENTRICULAR FIBRILLATION 319 1 i ^M ' ■" 1 ifiiu i •■■ 1 H J 1 ■ 4- 1 ! tii 1 1 t §© I! 'I 1 I'f 'igT' i t'f iffl " i t '^'14 T't n 5jj' j| .f iuja\ Si- I'T" ■ |;'jt ' iS^ ' & '! li t II Rj » , il " ■ *^ Jp^i J-iL V' si" S'T ^J^'^^i!^ I']' Ilfl' ■-If ^wH ' i|; te[ h:| j H I' s S ,i3 O O GO 5 «J p fl r. a •-e ts i d & 2 Ti f. CI -fl <1) > S Lj P "^ bl) Si; :3 H '" 1 13 ns hn .-3 « T5 .2^ T3 O ^ S d j3 _ 1 i^; "" V ... . & g & ■■■ V. v ■§ ]■! ■ ^ 1#r ' Wi I* V cd 1 ^ .2 m ■ -i-^ ^ 1 ' ^t , « ^^ 1 ^ «E ^j c 1 C! u .pi :;::i ? t|L j I ^ •< H iri 1 '\', -fl^;-"' HX-i--J cc H ■ nil A* W - -_H '^ S ^ ^L a • ' K 1^' 1 -S ' B» n ^ K^ ■ m 'T3 oK" ' ' i^ : 1 M i - ~L^ -. _;=— - iirl: =-- ~ 1 1 ~- -X f "s m if -.il \ f^fK f^ :