r^^\ ow^S Columbia fHmbersittp in tfje €itp of J^eto |9orfe College of fi!tctan£( anb burgeons; Reference Hibrarp Digitized by the Internet Archive in 2010 with funding from Open Knowledge Commons http://www.archive.org/details/surgeryofpancreaOOswee THE SURGERY OF THE PANCREAS BY J. E. SWEET, A.M., M.D. Assistant Professor of Surgical Research, University of Pennsylvania, Philadel- phia. Awarded the Alvarenga Prize of the College of Physicians of Phila- delphia for 1915. " With each advance in technic we reach a higher level from which a wider field of vision is open to us, and from which we see events previously out of range." — Pawlow. 1. Introduction. 12. Pancreas vergiftung. 2. Anatomy of the Pancreas. 13. The Relation of the Pancreas to 3. The Blood-vessels. High Intestinal Obstruction. 4. The Lymphatics. 14. The Pathology. 5. The Nerves. 15. Chronic and Acute Pancreatitis. 6. The Microscopic Anatomy. 16. The Factor of Safety in Physiology 7. The Surgical Anatomy. and Surgery. 8. The Embryology. 17. The Diagnosis of Pancreatic Dis- 9. The Physiology. ease. 10. Internal Secretion. 18. Pancreatoenterostomy. 11. The Relation to the Ductless Glands. 19. The Surgery. 20. Conclusions. INTRODUCTION There comes a time in the affairs of men when it is well to pause for a moment and take an inventory of whatever is on hand, in order that one may know what he has as the result of his labor and what he needs to make his labor more productive. I propose in the following to take stock of our knowledge concern* ing the pancreas, particularly to assemble certain parts of our knowl- edge which lie on the experimental shelves, to dust off the old articles and add certain new ones ; to see, when we have our inventory com- plete, if we do not have greater surgical assets than we think. Of all the organs of the body, the pancreas seems still the one most feared by surgeons; there are certain good reasons for this respect, and equally good reasons why this proper respect should not be allowed to grow to the proportions of a noli me tangere. 1 "VuW** **^> ^«w\">fvV€vtv»V k €*u V OVWt/Vy # vK t %tArJ*>5v INTERNATIONAL CLINICS THE ANATOMY The pancreas (nav tcpta?, all flesh, though not necessarily mean- ing muscle tissue) retains the meaningless name given it by the ancient anatomists in most modern languages except the German, which often employs the term Bauchspeicheldruse; but even this im- provement does not do justice to the organ, which is far more than a salivary gland. The trend of modern investigation points strongly to the suspicion that the pancreas is the most important organ of the body, not because of its function in the digestion of food, which can be entirely dispensed with, but from the point of view of the control exercised by the pancreas over the functions of many of the other glands which possess an internal secretion, and from the point of view of the control by the pancreas of glucose metabolism. The importance of this latter function is seen not only in the history of our knowledge of diabetes, but even more, perhaps, in the recently- offered proof that a portion of every protein molecule is available for conversion into glucose. Once converted into glucose, this protein fraction becomes subject to the control of the internal functions of the pancreas ; therefore the pancreas must be considered as controlling, by its internal function, not only carbohydrate metabolism as it is ordinarily thought of, but also a part of j3rotein metabolism. The pancreas lies not only functionally, but anatomically, at the centre of the abdominal cavity, lying transversely across the spinal column at the level of the first or second lumbar vertebra, though occasionally higher, across the twelfth thoracic. In shape Meckel compared the pancreas to a sort of hammer; Verneuil to a cross placed on its side, the short vertical arm repre- senting the head; Winslow compared it to a dog's tongue; Piersol l compares it with a revolver, the head of the pancreas representing the handle of the revolver. These descriptions convey fully as accurate a concept of the form of the organ as does the division into many parts and surfaces of which the anatomists seem so fond. Since the surgeon has no occasion to investigate the normal organ, and since the normal picture is so complicated by adhesions and inflamma- tions when he does see it, the relation of the organ to surrounding- structures is of more concern to him than the normal form. The head of the pancreas lies in the curve of the duodenum, conforming to the variations in position and consequent form of the duodenum. On the front of this portion of the gland is a groove for SURGERY OF THE PANCREAS o the gastroduodenal branch of the hepatic artery, this groove marking anteriorly the division between head and neck. The ductus chole- dochus passes beneath the upper part of the head and the duodenum, and is not uncommonly entirely buried in the glandular tissue of the pancreas. Conforming to the variations in shape and position of the duodenum, the head of the pancreas may be in close proximity on the posterior surface to the vena cava, the right renal vein, or the right suprarenal body. The posterior surface of the portion called the neck is marked by a deep groove for the portal vein, which vein may be entirely surrounded by gland tissue. The posterior surface of the body and tail of the pancreas, from the deep portal groove to the end of the tail, or the splenic end, lies on the vena cava, then on the aorta between the cceliac axis and the superior mesenteric artery, which groove the pancreas above and below ; beyond this point it lies on the left pillar of the diaphragm, the left suprarenal body, and the left kidney. The end of the tail usually rests against the spleen, or may extend still farther toward the left. There are two horizontal grooves on the posterior pancreas surface, the smaller one above for the splenic artery from near the aorta to the tail, the larger groove below, from the deep portal groove to the tail, for the splenic vein. The actual point where the pancreas crosses the spinal column is represented by a tuberosity called the tuber omentale. The actual length of the pancreas is variously given by different anatomists, varying from Henle's 16 to 22 cm. to Schirmer's measurements of 22 to 26 cm. and occasionally 30 cm. The relation between the stomach and the pancreas is so intimate that the ancients supposed that the pancreas formed a pillow for the stomach to rest upon. The lower end of the vertical portion of the pyloric end of the stomach rests directly upon the tuber omentale. These relations with the stomach depend, however, upon the degree of gastroptosis present ; depending upon the position of the stomach, cysts and tumors of the pancreas may present above, behind, or below the stomach, and, if below, either between the stomach and transverse colon or below both stomach and transverse colon. These relations are shown in the diagram, Fig. 1.* In some patients with pronounced gastroptosis the normal pancreas may be palpated above the stomach. The relations of the pancreas to the peritoneum are such that the * My thanks are due to Mr. Erwin F. Faber for the accompanying drawings. 4 INTERNATIONAL CLINICS pancreas — at any rate, the body and tail — doubtless shares less in a general visceroptosis than any of the other abdominal organs. When first formed in the embryo, the pancreas runs upward toward the head, behind the stomach, and between the layers of the mesogastrium, so that it possesses a complete peritoneal covering. Subsequently, as a result of the changes which take place in the position of the stomach, the pancreas turns over on its right side, and becomes adherent to the posterior abdominal wall. The peritoneum of the posterior sur- face is soon lost by absorption, but it persists on the anterior surface ; the organ, arising entirely intraperitoneal, has therefore become extra- peritoneal. This retroperitoneal location, together with the relation to the great vessels, renders the organ less movable than any other viscus. The peritoneal covering of the anterior surface of the pancreas is derived from the prolongation of the two layers of the transverse mesocolon, which latter is attached to the anterior border of the gland, from the tail to the neck. At this border the two layers separate, the anterior, derived from the lesser peritoneal sac, passing back- wards and upwards over the superior pancreatic surface; the pos- terior, derived from the greater peritoneal sac, turning downwards and backwards along the inferior surface. The pancreas in cross- section is roughly triangular, with the base against the posterior body wall, the apex of the triangle, to which the mesocolon is attached, towards the front, one limb of the triangle forming the superior pancreatic surface, the other limb the inferior surface (Fig. 1). As the transverse mesocolon is followed to the right it ceases, as a rule ; that is, its two layers fail to meet about the neck of the pancreas. Beyond this the posterior surface of the colon is generally free from peritoneum, and is connected by areolar tissue to the front of the head of the pancreas. When the head extends below the level of the colon, the pancreas is covered by the continuation downwards of the peri- toneum from the under surface of the colon. Often the transverse mesocolon is continued to the right as far as the hepatic flexure, when the anterior surface of the head is completely covered by peritoneum. THE BLOOD-VESSELS The arteries of the pancreas are: (1) The superior pancreatico- duodenal, a branch of the gastroduodenal artery, which runs down on the front of the head, sending branches outwards to the duodenum, Fig. 1. G.H.L n c.e.L T.W. (7 PC The relation of the pancreas to the greater and lesser peritoneal cavities, and to the stomach, colon, duodenum, and the retroperitoneal space. (Modified from the diagram of Piersol.) G. H. L., gastrohepatic ligament; M., mesocolon; G. E. L., gastro-epiploic ligament; F. W., foramen of Win- slow, with arrow passing through the foramen from the greater to the lesser peritoneal sac; P., pancreas; £>., duodenum; G. P. C, greater peritoneal cavity. — V. fe M 'S3 B ro ess O C.S .2 c> ™ s S ° 10 SURGICAL ANATOMY The surgical anatomy of the pancreas becomes, therefore, ex- tremely complicated because of the intimate relation of the organ with so many important organs and so many vital structures. The head is in definite relation to the common duct and the duodenum, and to the common blood supply of head and duodenum ; this blood supply is, 8 INTERNATIONAL CLINICS in the dog, of such importance that the destruction of the common artery may result in necrosis of the duodenum. The head may be in relation to the right kidney, the right suprarenal body, or the vena cava. The neck of the pancreas is in immediate contact with the gastroduodenal branch of the hepatic artery anteriorly, and pos- teriorly with the portal vein, which it may surround. The body is intimately associated with the cceliac axis, the aorta, and the superior mesenteric artery, and with the vena cava; the body and tail with the artery and vein of the spleen; the tail with the spleen, the left kidney, and the left suprarenal body. The pancreas is more or less intimately associated with the stomach and with the lesser peritoneal cavity; it is also in relation, below the attachment of the mesocolon along the apex of the gland, with the greater peritoneal cavity (Figs. 5 and 1). The retroperitoneal location may result in involvement of the tissue surrounding the great vessels. These anatomical rela- tions show why disease of the pancreas may involve the common duct, the duodenum, or the pylorus, the cardiac end of the stomach, or even compress the whole stomach against the anterior abdominal wall ; the colon or the ureter may be involved ; or the portal vein, the vena cava, aorta, splenic vessels, or the superior mesenteric vein; further, these relations show how disease of these adjacent structures may, conversely, involve the pancreas. A tumor of the pancreas extending toward the right may produce jaundice and intestinal obstruction; extending upwards, the same symptoms, plus pyloric obstruction and gastric dilatation ; extending backwards, by pressure on the retroperitoneal lymphatics, ascites and oedema of the lower limbs. It is apparent, from these relations, and from a glance at Fig. 1, that the pancreas may be approached surgically either from the front or the back ; the approach through an anterior incision would proceed either above the stomach, through the gastrohepatic omentum, below the stomach, through the gastro-epiploic omentum, or below both stomach and colon, through the mesocolon. The choice of route would depend upon the position of the stomach in the individual patient. From the back the body is approachable at the costovertebral junction, the tail through an incision extending from the tip of the twelfth rib anteriorly. A result of the anatomical position of the pancreas is the very Fig. 4. The relation of the lymphatics of the gall-bladder to the head of the pancreas. (From Frankc.) SURGERY OF THE PANCREAS 9 complete protection afforded by this location from the effects of external violence. This protection is, however, not perfect, as shown by the cases of complete subcutaneous transverse rupture of the pancreas which have been reported. 11 ' 12, 13, 14 THE EMBRYOLOGY The pancreas arises from one dorsal and one or two ventral Anlagen, budding out of the primitive gut, the dorsal Anlage nearer the stomach, the ventral Anlage or Anlagen in close relationship with the ductus choledochus. These primary diverticula give off hollow buds, which in their turn give off others, and this process is continued until the mass of the gland is formed (Fig. 6). The terminal buds thus produced develop into the gland acini, while the others form the gland ducts. The latest authorities speak of three Anlagen, one dorsal and two ventral, of which the left ventral soon retrogresses, or never develops beyond a rudimentary stage. As the digestive tube now grows in length, together with the lengthening of the ductus choledochus, the persisting ventral portion becomes further removed from the duodenum, its ducts fuse with the duct system of the larger dorsal Anlage, 15 and then, curiously, the duct system of the smaller portion develops into the most important duct, the duct of Wirsung, while the glandular tissue developing from this Anlage forms only a part of the head and the uncinate process; the greater part of the pancreas develops from the dorsal Anlage, while the duct system of the original Anlage of this portion becomes the slightly-developed duct of Santorini. While this is the general rule, variations may occur, as seems natural from such an involved mode of duct growth. In a study of one hundred subjects Opie 16 found two ducts in every instance, but occasionally one or the other was so small that it was found with difficulty. In ten out of the hundred cases the two ducts failed to anastomose within the gland, and in four additional cases the two duct systems were united by such a minute twig that they might be regarded as independent of one another. In twenty instances the duodenal end of the duct of Santorini was not patent, and in a considerable number of specimens the orifice of the duct of Santorini, though patent, was so minute that its func- tional significance was slight. In eleven out of one hundred speci- mens the duct of Santorini, on the contrary, was equal in size to, or 10 INTERNATIONAL CLINICS larger than, the duct of Wirsung, so that during life it was doubtless the outlet for a considerable, if not the larger, part of the pancreatic juice. This matter of the ducts of the pancreas becomes of importance in the study of the relation of pancreatitis to gall-bladder disease, and particularly to the technical question of the possibility of drain- ing the pancreas through the gall-bladder. Not only does the develop- ment of the ducts result in variations in the relative size and func- tional importance of the two ducts, but variations in the form of the actual union between the duct of Wirsung and the ductus chole- dochus occur, 17 as is shown in Fig. 1. In Fig. 1 the type of communication is such that a small gall-stone could block the papilla of Vater, and drainage of the gall-bladder would drain the pancreas; in the type of Fig. 3 this same condition might possibly occur; in types 2 and 4, however, the pancreas could not be drained through the gall-bladder under any conditions. This matter of drainage will be discussed later, and is introduced here to suggest the surgical im- portance of a knowledge of the variations in the ducts, variations which seem clear in their origin and in their diversity when we think of the manner in which the ducts change in the course of development. Two important surgical possibilities also depend upon the em- bryological development of the pancreas, annular pancreas, and ac- cessory glands. If we assume that the ventral bud, before it becomes fused with the dorsal Anlage (Fig. 6), becomes turned once around the duodenum in the course of the turning movements followed by the stomach before the stomach reaches its final position, and then fuses with the portion developing from the dorsal Anlage, we have a clear picture of the manner in which such a pancreas as is shown in Fig. 8 could arise. 18 While very rare, this annular pancreas may, by the production of intestinal obstruction, be the cause for surgical interference. The occurrence of an accessory or aberrant pancreas is explained by Broman 19 on the basis of the phylogenetic development of the pancreas. He assumes as the earliest type of pancreas one in which many small glands are disseminated in the wall of the entire small intestine; as the pancreatic function, going upward in the scale, be- comes more or less dependent upon the liver function, we find those glands nearest the liver taking on the entire function, and then unit- Fig. 5. Diagrammatic presentation of the relations of the pancreas to the neighboring organs. Fig. 6. Reconstruction of the two pancreas anlagen in a human embryo of five weeks. (From Hamburger.) Ch, ductus choledochus; v, ventral anlage; dr, dorsal anlage; s, stomach. Fig. 7. The four types of anastomosis between ductus choledochus and duct of Wirsung, seen in cross- section of the duodenal wall. (From Letulle and Nattan-Larrier.) Fig. 8. Annular pancreas. (From Cords.) Ch, ductus choledochus; D, duct of the dorsal V, duct of the ventral anlage. SURGERY OF THE PANCREAS 11 ing in a definite single organ, while those scattered below disappear. An accessory pancreas would therefore represent an atavistic reversion to the primitive type. Whether this explanation covers the not -un- common occurrence of such aberrant pancreas tissue in the wall of the stomach, or whether this tissue in the stomach may be related to the left ventral bud which is supposed to disappear, matters little to the surgeon. He must remember that such glands do occur, in the stomach wall or at any point in the wall of the entire small in- testine, or even in relation with the umbilicus, and should not mistake them for neoplasms ; such aberrant organs may also become diseased, independently of the main organ. THE PHYSIOLOGY A study of the physiology of the pancreas leads to the conclusion that the pancreas is a complete, self-contained laboratory of physio- logical chemistry. Perhaps this is another reason why the surgeon hesitates to approach the pancreas. The development of modern surgery daily makes clearer the fact that physiology is of basic im- portance to surgery, and it is just this organ, the pancreas, which demonstrates that a knowledge of physiology is often of more funda- mental importance than a knowledge of anatomy: anatomy teaches only the " how," never the " when," of pathology or the " why " of physiology. Further, the manner of attacking any form of pan- creatic disease is, I believe, to be determined by a knowledge of certain of its physiological peculiarities. Like all the abdominal organs, the pancreas receives fibres from the central nervous system by two paths, the vagus and the splanchnic. The meaning of this dual nerve supply, and even the relation of the nervous system to the activity of the gland, is not yet clear. The role of the nerve relations in pancreatic activity is further complicated by the presence in the gland of the autonomous nervous system repre- sented by the numerous ganglia scattered throughout the gland; a still further complication is offered by the fact that the most potent stimulant of pancreatic activity is not nerve stimulation, but the action of a chemical stimulant, brought to the cells of the gland by the blood stream from its place of origin in the intestinal mucosa. This substance, called " secretin " by its discoverers, Bayliss and Starling, 20 and further classed as a hormone, or chemical messenger, 12 INTERNATIONAL CLINICS is formed by the action of hydrochloric acid upon a substance called " prosecretin," contained in the cells of the mucosa of the duodenum and the entire jejunum. When secretin is brought to the pancreas by the blood stream, the pancreas immediately begins to secrete. The normal process of pancreatic activity is, then, that the hydro- chloric acid, coming from the stomach in small amounts as the pylorus allows the food to enter the duodenum, changes the prosecretin into secretin, which is carried to the pancreas and causes that organ to furnish the digestive fluid necessary to complete the work of digestion begun by the ferments of the stomach. The creation of abnormal conditions by performing a gastroenterostomy, with closure of the pylorus, does not seem to seriously interfere with this relation of acid, prosecretin, secretin, and pancreatic juice, probably because the pro- secretin is found throughout the greater part of the mucosa of the small intestine, and the abnormal condition created by the gastro- enterostomy is quickly compensated. Alkalies hinder the production of pancreatic juice, perhaps by neutralizing the acid of the gastric juice ; the use of alkalies is suggested to lessen the flow of juice from a fistula of the pancreatic ducts. Physiologists reckon the amount of pancreatic juice secreted daily as about equal to the amount of gastric secretion, or in man about 1500 ccm. The most important constituents of the pancreatic juice are the ferments, or, better, the proferments or zymogens. These ferments are the proteolytic trypsin, which breaks down the proteins to the amino-acids; a lipase or steapsin, which transforms neutral fats into fatty acids and glycerin ; a rennin-like ferment ; a diastase or ptyalin, the same as found in the saliva; a maltase, which can transform the maltose, derived from the starch by the diastase, into glucose; a lactase is found in the pancreas of sucklings and in the pancreas of animals fed on milk-sugar. The pancreatic juice contains, further, a nuclease, which dissolves the nucleins which have been precipitated in the stomach ; the action of this ferment consists in transforming the gelatinizing nucleic acid a into the more soluble, non-gelatinizing nucleic acid b. This fact of digestion of the nucleins by the pancreatic juice is the basis for Schmidt's test 21 for pancreatic function ; he concludes, from the appearance of cell nuclei in the faeces, that a disturbance of pancreatic function exists. The theoretical value of this test is impaired by the SURGERY OF THE PANCREAS 13 work of Umber, 22 who found that considerable amounts of nucleins are digested in the stomach, and by the work of Gumlich 23 and Araki, 24 who showed that the succus enterieus can dissolve the nucleins. Aside from the doubtful value to the surgeon of the nuclease of the pancreas, or its absence, as an aid in diagnosis, another of the pancreatic ferments furnishes the only positive diagnostic factor in certain types of pancreatic disease, in regard to which factor there is universal agreement, that it gives a positive, unfailing diagnostic sign. This is the lipase or steapsin. To be sure, it is necessary to per- form a laparotomy in order to obtain the help of this agent in diag- nosis ; but in the present stage of our knowledge it would seem entirely justifiable to proceed to an exploratory laparotomy and obtain the advantage of an early operation rather than wait for other and by no means satisfactory methods of diagnosis. The steapsin of the pancreatic juice breaks down the neutral fats into fatty acids and glycerin ; this process occurs, of course, normally within the intestine, but if the pancreatic juice escapes into the peri- toneum or into the tissues it can act upon the fat of the body itself, forming characteristic, round, dead white patches, which, once seen, are never mistaken for any other condition (Fig. 14). The rela- tion of these areas of fat necrosis to steapsin was shown by Flexner, who demonstrated the ferment and its products in these areas. This fat-splitting ferment is not contained in the pancreas as such, but as a zymogen, steapsinogen, which is transformed into the ferment steapsin by some constituent of the bile. There seems to be a certain amount of active steapsin in the pancreas, from which we must conclude that the steapsinogen can be activated by some other substance than the activator contained in the bile. The occurrence of fat necrosis would not, therefore, necessarily justify the conclusion that the bile had been forced into the pancreatic ducts, and had there converted the zymogen into the active ferment before its escape into the tissues; the theory would suggest such a possibility, but there are no reports from which any deductions bearing upon this point can be drawn. While the presence of fat necrosis is the only positive sign of pancreatic disease, outside the pancreas itself, it is not necessarily always present. The proteolytic ferment of the pancreas, trypsin, is the product 14 INTERNATIONAL CLINICS all-important to the surgeon, determining not only the condition for which he must most often approach the pancreas, but, further, this trypsin determines the manner in which, as we shall see later, he must modify his treatment, as compared with his treatment of the inflammations of other organs. Out of the wealth of fact concerning trypsin and its relation to digestion certain things stand forth which the surgeon must know; to these we will limit our attention. The ferment is not contained as a ferment, neither in the gland-cells nor in the ducts, but is found in the form of the inactive proferment or zymogen, trypsinogen. This is the answer to the old question of why the pancreas does not digest itself. The normal activating factor for trypsinogen, the agent which makes of the harmless trypsinogen the powerful proteolytic ferment trypsin, is produced exclusively, under normal conditions, by the cells of the mucosa of the small intestine, and is called enterokinase. This body, concerning the relation of which or its mode of action little is known, is not continually secreted by the cells of the intestinal wall, but only when trypsin (respectively trypsinogen) is brought into the lumen of the intestine. The change from trypsinogen to trypsin is effected instantaneously at body temperature. This enterokinase seems to be the only substance normally present in the body which can change trypsinogen into trypsin. Delezenne 25 believed that an active enterokinase could be extracted from the leuco- cytes of the blood or the Peyer's plaques of the intestine, but this does not seem to be the case. Bayliss and Starling 26 state that no other substance can effect this conversion of trypsinogen into trypsin, a statement which, I think, we shall find later should be modified — that no other normal constituent of the body can effect this change. An activating substance similar in action to enterokinase, if not identical with it, appears to occur in many bacteria (Delezenne 25 and Hekma 27 ), and there is further, and surgically important, evidence that trypsinogen is also activated by some constituent of the pancreas itself when the gland is injured in such a manner that necrosis develops. The observation was made long ago by Heidenhain, 28 and confirmed by others, that in glands removed from the body and in watery extracts of glands trypsin appears, at first slowly, then rather rapidly (Vernon 29 ). The fact of some such change of trypsinogen into trypsin is seen SURGERY OF THE PANCREAS 15 in the single experiment from which the original of Fig. 14, show- ing fat necrosis, was obtained. A part of the pancreas of one dog, removed under strict aseptic precautions, was placed, again under aseptic technic, into the peritoneal cavity of another dog. The ensuing fat necrosis proves that the steapsin must have either been present in the gland or else have been activated from the steapsinogen by some substance not normally present in the gland. The fact that such a procedure inevitably kills the dog into whose belly the pancreas is placed shows that some powerful poison has been developed, and, as we shall see in a later chapter, this is either trypsin or the toxic products of tryptic digestion. On the other hand, the injection of large amounts of fresh, normal pancreatic juice into the belly of a dog, or the arranging of the ducts of his own pancreas so that they will discharge the pancreatic juice into his own peritoneal cavity, is followed by no symptoms at all. The pancreatic juice is also said (Bayliss and Starling 30 ) to contain a weak proteolytic ferment resembling erepsin; this will digest fresh fibrin, but not coagulated protein. Whether or not this proteolytic ferment enters into the conditions which interest the surgeon is not known. THE INTERNAL SECRETION OF THE PANCREAS The problem of the internal secretion of the pancreas, as it is called — that function which seems to be at the bottom of the meta- bolism of the carbohydrates — represents, so far as we know at present, the one vital function of the pancreas. A whole library 3 * has grown out of the study of this function, and, judging from current physio- logic and physiologic chemical literature, interest in the problem is as keen as ever, or as ever since the fact of the relation of the pancreas to diabetes mellitus was first established by Minkowski. A complete discussion of this problem would obviously be out of place in this surgical essay, for, while it is true that the surgeon must know all that the internist knows, and then more — this more being the whole subject of surgical diagnosis and technic — yet it is hardly necessary that the surgeon should be expected to follow the painstaking and painful mental gyrations by which the true internist finally success- fully arrives at the conclusion that our knowledge is limited. So in regard to this problem of the relation of the pancreas to 16 INTERNATIONAL CLINICS diabetes mellitus, let us confine ourselves to a statement of what seems to be the consensus of opinion. The physiologist and the physiological chemist agree that the complete removal of the pancreas is certainly followed by the most severe form of fatal diabetes : some blame this condition upon the loss of the islands of Langerhans ; others are not so sure, since cases of diabetes mellitus have been recorded in which no demonstrable lesions of the pancreas could be found. Most chemists are agreed that the pancreas furnishes in its in- ternal secretion a substance which facilitates the combustion of the carbohydrates ; possibly this substance activates a glycolytic or glucose- burning ferment produced by the muscles. The characteristic feature of pancreatic diabetes, according to some physiological chemists, is not the glucosuria, but the accompanying evidence of a disturbance of fat metabolism, as manifested by the acidosis ; therefore the finding in the urine characteristic of pancreatic disease would not be glucose, but glucose plus acidosis. In other words, according to this idea, glucose may appear in the urine as a result, or accompaniment, of many other conditions besides pancreatic disease ; acidosis may accom- pany other conditions, such as starvation, but glucosuria plus acidosis means pancreatic disturbance. I trespass upon the peculiar province of the internist to the above extent, in order to point out that the discovery of sugar in the urine is not enough to justify the diagnosis of pancreatic disease. A matter of general theoretical interest, and of practical surgical interest, in connection with the problem of the internal function of the pancreas, is the question of the increased susceptibility of the diabetic to infection. It is sometimes taught that this increased susceptibility is due to an increased sugar content of the tissues ; this, however, cannot be the explanation, since the characteristic fact con- cerning diabetes is that the blood sugar is only slightly increased : as soon as an increase occurs the sugar is carried off by the kidneys, and hence the glucosuria. My first work in connection with the pancreas concerned just this point; it was found 32 that there is a considerable loss of hsemo- lytic complement from the serum of the completely depancreatized dog; this loss of complement — the finding having been later con- firmed for the bacteriolytic complement of the blood — offers a reason- able explanation of this diminished resistance to infection, even though the mechanism of this loss of complement remains unexplained. SURGERY OF THE PANCREAS 17 THE RELATIONS OF THE PANCREAS TO THE OTHER ORGANS WITH INTERNAL SECRETION What shall be said of the relations between the pancreas and the other organs with an internal secretion ? Here again we are con- fronted by an extensive array of publications, from which more than the technical ability of the most skilful surgeon is required to dissect out the vital portion. I am therefore inclined to present simply the facts which I have personally observed; the bearing of these facts and their interrelations are not clear, nor is there, in my opinion, sufficient evidence to-day to permit of anything but the most vague theorizing. The first interrelation which has come to my personal observation is that the suprarenal exercises some sort of control over the pancreas. Such an interrelation has been found by others, partly on the basis Fig. 9. Respirations Artificial respirations, I ■ .Pressure here much lower, than alter first secretin when juice flowed although now juice is inhibited 8lood-pressure Flow after secretin | 5cc adrenalin ' iTime jn* The inhibitory effect upon the activity of the pancreas, induced by secretin, of an injection of adrenalin. of experiment, partly on the basis of theoretical labor ; it has even been claimed that the removal of the adrenals prevents the appear- ance of the glucosuria following upon the complete removal of the pancreas. The observations which lead us to the conclusion that the suprarenals and the pancreas are interrelated are seen in the accom- panying tracings. 33 These tracings were obtained as follows: A dog under constant ether anaesthesia was attached to a kymograph, by which was recorded, for the entire time of the experiment, a con- tinuous tracing of respiration, blood-pressure, with the time in seconds and a base line on which the flow of juice from the pancreas could be electrically recorded, each vertical stroke on the base line recording the passing by the column of juice of the graduations on a cannula inserted in the main pancreatic duct. As is shown by the tracing (Fig. 9), the injection of 5 ccm. 18 INTERNATIONAL CLINICS of secretin caused a flow of pancreatic juice, and the injection of 5 ccm. of adrenalin stopped this flow within three minutes. The converse of this proposition is also true: i.e., that the re- moval of the adrenals is followed, after a time, by a flow of pancreatic juice. This is seen in the tracings (Figs. 10 and 11), which were prepared in the same manner as just described, by the use of a con- tinuous kymograph. In Fig. 10 A, a marked flow from the pancreas is recorded; in Fig. 10, B, taken from a later period of the trac- ing, as shown by the lowered blood-pressure, the flow has become Fig. 10. A Cannula full 31 35 . 90 i . .. Cannula full .Emptied to 20 Cannula emptied to division 22 "1 I !<■ . . . ■ . | JO 20 B The induction of pancreatic activity by the removal of the adrenals. A. The flow at a time when blood-pressure is still fairly high. B. The flow at a later period, when the blood-pressure has become much lower. The upper line in each tracing is the respiratory tracing; the second line is the blood-pressure; on the base line is the record of flow from the pancreas, through a graduated cannula; below is the time in seconds. much more marked; in Fig. 11 is shown the composite picture of the effect upon blood-pressure and pancreatic activity, a trac- ing formed of sections of the entire tracing taken at consecutive hourly intervals ; in this tracing is seen the effect upon the general blood-pressure of the removal of the adrenals, the gradual progressive fall, and the appearance of the flow of pancreatic juice. This find- ing of some sort of interrelation between the suprarenals and the pancreas is, in my opinion, of surgical significance. It has been shown beyond question by the physiologists that the suprarenals are con- SURGERY OF THE PANCREAS 19 cerned in the formation of a substance, adrenalin, which is essential to the maintenance of normal blood-pressure. 34 The proof of an interrelation between pancreas and adrenals raises the question of whether the collapse and fall of blood-pressure so characteristic of acute pancreatic disease may not be due to a destruction of suprarenal function. I have as yet insufficient evidence to constitute proof, only sufficient to warrant the suspicion that the adrenals are concerned in acute pancreatitis. The second interrelation which I have observed is between the pituitary and the pancreas. 35 After the removal of the pituitary from dogs, the pancreas has always, in a long series of experiments, been found at autopsy to have a peculiar color ; namely, that of the normal Fig. 11. ti A.M Tic S ■"»-*> ll( *1 The effect upon blood-pressure and the secretion of pancreatic juice of the removal of the adrenals. Ine pancreatic secretion is shown by a series of vertical marks on the base line, each mark recording the instant at which the column of juice passed the graduations of a glass cannula connected with the duct. gland during digestion. This pigmentation of the normal gland during digestion is something apart from the blood supply, since it persists after washing the organ free from blood. The clinical fact that in certain cases of pituitary disease an in- creased sugar tolerance can be demonstrated is interpreted by Cush- ing 36 as pointing to the relation between pituitary and pancreas. The difficulties encountered in attempting to perform satisfactory experi- ments on sugar tolerance, the complexity of the problem of glucose metabolism, together with the natural dependence of glucose meta- bolism upon carbohydrate digestion and resorption, have prevented reaching a conclusion perfectly satisfactory. Work upon this phase of the problem has been in progress in my laboratory for some time, 20 INTERNATIONAL CLINICS in collaboration with Dr. A. I. Ringer, but we have as yet not been able to disentangle the numerous complexities. Some relation between pituitary and pancreas certainly exists. The third interrelation which I have observed is one between the pancreas and the spleen. Schiff, in 1862, observed that the spleen swells during digestion. Herzen 37 and Gachet and Pachon 38 claim that the spleen forms a product of internal secretion which can change trypsinogen into trypsin. My contribution to the question of the interrelation between spleen and pancreas is illustrated in Fig. 12, which shows the remarkable atrophy of the spleen following the complete removal of the external function of the pancreas. In other words, if that part of the pancreas of the dog be removed which lies along the duodenum, tying off the body and tail and the uncinate process and leaving them in the peritoneal cavity so that the internal function of the pancreas is amply provided for, this marked atrophy of the spleen, among other things, occurs. The change apparently consists in a simple atrophy, as is shown in the photo- micrograph, Fig. 13. This splenic atrophy has been a constant finding in a large series, and has been marked after three days' time. The possible surgical bearing of this relation between pancreas and spleen is suggested by the report of Musser, 39 of an acute anaemia in four of eight cases of acute pancreatitis. The well-known relation of the spleen to the blood, together with this finding of acute splenic atrophy, suggests an explanation of Musser's report of acute anaemia, and further suggests the importance of clinically following the blood- picture in cases of suspected or proved acute pancreatitis. The fourth interrelation which has come to my observation is that between the pancreas and the thyroids. Fig. 12 shows the results upon the thyroid gland of the operation just described, of the removal of the external function of the pancreas. The thyroids have in every case of this series shown a change which consists in a striking translucency or transparency of the gland; this apparently depends upon a great increase in colloid, though the alveolar cells do not seem much changed microscopically. In the plate this peculiar transparency is illustrated by drawing the lobe of the thyroid held against the light, with the blade of a scalpel interposed. The para- thyroids stand out against this translucent background as delicate pink bodies, and even the lower pair of parathyroids, which in the Fig. 12. The effect upon the spleen of the complete removal of the external function of the pancreas. The large spleen is drawn from the measurements taken at the operation; the small spleen is the actual size of the spleen found at autopsy. Above, a normal thyroid lobe of the dog; beside it, a thyroid lobe from the dog whose spleen is pictured. SURGERY OF THE PANCREAS 21 dog are buried within the thyroid tissue, can be seen on holding the transparent thyroid to the light, while normally the parathyroid thus buried can be found only by cutting and staining serial sections. Microscopically the parathyroids show no change. 'Now what do these findings mean, and what is their bearing, physiological or surgical ? I do not know, and I do not think that any- body else knows. The only conclusion which I feel justified in draw- ing is that the activity of the pancreas, as represented by its relation to the digestion and the absorption of food by the external secretion, together with its further influence on metabolism by its internal secretion or secretions, makes of the pancreas the mainspring which determines the movements of every one of the wheels in the complex mechanism of the living body. PANCREASVERGIFTUNG On December 4, 1914, the pancreas was removed from a normal dog and dropped into the peritoneal cavity of another normal dog, the entire procedure being conducted under strict aseptic precautions. Twenty-four hours later the animal was dead. The autopsy, per- formed immediately after death, showed a picture of beginning peri- tonitis, localized to the peritoneum which was in immediate contact with the softened, dirty yellow, disintegrating transplanted pancreas ; the animal's own pancreas was normal. Only one small area of fat necrosis could be found, this being in the immediate proximity of the transplanted pancreas. On the other hand, the peritoneal cavity contained a considerable amount of dark fluid, best described as a dirty fluid, which, in my experience, is always suggestive of pancreatic disease, even if not quite as definitely pathognomonic as the finding of fat necrosis. This acutely fatal poisoning inevitably follows the plac- ing of sterile pancreatic tissue, in sufficient amounts, into the peri- toneal cavity of a normal animal, whether it be the animal's own tissue, isolated from its blood supply so that autolysis or necrosis must follow, or whether the pancreatic tissue come from another animal of the same species or of a different species. This is the condition called Pancreasvergiftung by the Germans, and it seems to me that a study of this condition will place the prob- lem of acute pancreatitis before the surgeon in a different light from that which is shed upon the subject by the ordinary text-book treat- 22 INTERNATIONAL CLINICS ment, or even, indeed, by some of the best monographs upon the subject of the surgery of the pancreas. Thus Moynihan, in the supplementary volume of " Keen's System" (1913), confines his discussion of this subject to the fol- lowing words : " This infection, or entry of intestinal juice into the ducts of the pancreas, activates the pancreatic juice and leads to autolysis of the cells and gangrene and hemorrhage (Coenen)." And Korte 40 devotes but five lines to the discussion of the work of Doberauer 41 and of von Bergmann and Guleke, 42 a manner of treat- ment which hardly does justice to the subject. There is, however, abundant evidence 43 that this activation of pancreatic juice is not only what " leads to autolysis of the cells," etc., but it is just this activated pancreatic juice which is responsible for the entire picture, canvas and frame, glass and signature, of this most acute and terrible intoxication, the whole secret of acute pan- creatitis. A full understanding of this fact settles many of the apparent problems of the pathology and surgery of the pancreas ; chronic pancreatitis is an infectious condition, in which infection alone is the factor; acute pancreatitis is the condition, whether pri- marily traumatic or infectious, in which the necessary conditions for the activation of pancreatic juice have been met. No further sub- division is necessary, nor can be made, since the various classifications of acute pancreatitis become but different degrees of the same process. The problem of the time of operation becomes clear — the sooner the better; the question of drainage ceases to be a question, a drainage must be provided, not only of the infection present but of this activated pancreatic secretion ; and we must consider this same type of drainage when dealing with a trauma of the pancreas. The problem of the toxic action of the pancreas has not proved easy of solution, nor is this strange, in view of the complex nature and properties of the pancreatic activities; not only are many powerful ferments elaborated by the gland, but their activity depends upon the collaboration of other factors — the activation of the proferments into ferments. Therefore it has taken the combined efforts of many workers to follow the intricacies of the problem, and only by the com- parative study of their results can we reach a definite conclusion. The discoverer of the pathological condition of fat necrosis, Balser, as shown by the title of his paper, 44 " Ueber Fettnekrose, eine zuweilen Fig. 13. 'lfZ^? , ***0»r -; :t * Photomicrograph of a section of the spleen after complete removal of the external function of the pancreas. The simple atrophy of the cellular elements results in a shrinking and consequent thickening of the capsule and trabecular. SUEGERY OF THE PANCREAS 23 todliche Krankheit des Menschen," inclined to the belief that the fat necrosis represented the fatal element in acute pancreatic disturbance. But that the fat necrosis is simply an accompanying symptom, which takes more time for its development than the fatal element in the process (Fig. 14), and may therefore fail to appear, is shown in the experiment described at the beginning of this chapter, or in the following case reports : Case of Kirste, 4 * 1902. — A man of fifty years is seized, without prodromal symptoms, with very severe abdominal pain, vomiting, belching, and weakness. Temperature normal, pulse small and rapid. Death occurs after twenty hours. The abdomen contains a half litre of blood-stained fluid; peritonitis; necrotic in- fection with hemorrhage in over one-third of the pancreas. Case of Heinecke, 4 " 1907. — A man of thirty-six years fell beneath a loaded wagon. Immediate symptoms of peritonitis. Operation two days later, prognosis serious; small amount of blood found in the peritoneal cavity, the intestines dis- tended and inflamed; some tears in the mesentery are found and repaired, but no other injuries are seen. Death from shock occurs a few hours after operation. Autopsy shows frac- ture of several ribs, with blood in the left pleural cavity; hematoma of the bursa omentalis; pancreas torn clear across, splenic vessels not injured; dirty gray color of the edges of the pancreas, as though digested; extensive diffuse fat necroses. Two cases reported before the New York Surgical Society by Erdman 48 il- lustrate the relation of fat necrosis to pancreatitis. The first case was a man ot fifty years, who, after a history of gall-bladder disease dating back three years, had a sudden attack of sharp pain, with profound jaundice and some cyanosis and dyspnoea. He was operated on the third day of the attack; hemorrhagic pancreatitis with fat necrosis was found. Recovery followed cholecystectomy and choledochotomy. The second case was a man of thirty-five years, who gave a history of very sharp abdominal pain, temperature 102°, pulse 120. On opening the abdomen for a supposed appendicitis, free fluid was found. A median incision was then made higher up, and an acute pancreatitis, without any evidence of fat necrosis, was found. Liberal drainage of the pancreas by puncture and a cholecystostomy was performed. The direct experimental proof that the factor causing the fat necrosis is not the fatal factor has been offered by Lattes, 49 who shows that an activated pancreatic juice which exhibits no fat-splitting power may produce the typical fatal picture without the fat necrosis ; and, on the contrary, that an animal which has developed no symp- toms following the injection of an inactive pancreatic juice may show extensive fat necrosis. The factor which causes fat necrosis is, then, not the factor respon- 24 INTERNATIONAL CLINICS sible for the fatal course of Pancreasvergiftung; fat necrosis, if pres- ent, is a sure sign of pancreatic disturbance, but pancreatic disturb- ance, even to the point of producing a fatal result, may occur with no sign of fat necrosis. Even after having ruled out the fat-splitting ferment as a vital factor, the problem has become only partly simplified. This fatal pancreas poisoning can be experimentally produced by a great variety of procedures, besides that of placing sterile pancreas tissue into the peritoneal cavity: artificial ischsemia (Blume, 50 Oser, 51 Milisch, 52 Lewit, 53 Wolf 54 ); the injection into the ducts of the pancreas of various substances [oil (Oser, 51 Hess, 55 Guleke, 56 Eppinger 57 ), bile (Guleke, 56 Flexner, 58 Opie, 59 Oser, 51 Polya 60 ), hydrochloric acid and intestinal secretion (Hlava, 61 Flexner and Pearce, 62 Hildebrand, 63 Rosenbach 64 ), intestinal secretion and pancreatic juice, commercial trypsin, calcium and sodium chloride (Polya 60 ), papain (Carnot 65 ), adrenalin (Rosenbach 64 ), zinc chloride (Thirolaix, 66 Oser 51 ), nitric acid, chromic acid, formalin, soda (Flexner 58 )] ; or by injection into the blood-vessels of substances such as oil, paraffin, wax, lycopodium (Panum, 67 Lepine, 68 Bunge, 69 Guleke 56 ) ; by tying off all the ducts during digestion (Hess 55 ), or by mass ligatures and gross injury of the pancreas (Katz and Winkler 70 ). This entire line of experiment has one common factor, the injury to the pancreas which naturally follows any disturbance of its blood supply, and also follows the injection of any material into the ducts. The pancreas is a very delicate and sensitive organ (Pawlow 71 ), and the injection of any substance into the ducts must be followed by injury to the finer branchings of the duct system, especially if, as in the majority of the experiments cited, the substance injected is in itself caustic or irritating. The entire list may therefore be boiled down to the statement that injury to the pancreas, no matter how caused, may result in the production of the typical picture of Pan- creasvergiftung. The injection into the peritoneal cavity or into the veins of an animal of inactive pancreatic juice (i.e., pancreatic juice which does not show in the test-tube the power of digesting proteins) does not produce any symptoms, except possibly a transitory lowering of blood-pressure, a phenomenon common to the extracts of practically all glandular structures. A similar injection of pancreatic juice plus ¥ ! .?mi Experimental fat necrosis in the omentum of the dog. Found four days after placing about one- third of a sterile normal pancreas into the belly of a normal dog. SURGERY OF THE PANCREAS 25 succus entericus (i.e., a pancreatic juice in which the trypsinogen has been transformed into trypsin by the normal action of the enterokinase of the succus entericus) will cause the typical symptoms of pancreas poisoning and death, the same result in every way as that produced by placing sterile pancreatic tissue into the peritoneal cavity. This fact is established by the work of Lombroso, 72 Roger and Gamier, 73 Falloise, 74 Cybulski and Tartschanow, 75 Roger and Gamier, 76 Fleig, 77 Seidel, 78 and Kirschheim. 79 On the other hand, Schittenhelm and Weichardt, 80 and Fragoin and Stradiotti 81 claim that the injection of active juice produces no symptoms. From my own experiments, I agree with the majority opinion ; active pancreatic juice is toxic, inactive juice is non-toxic, and it has been my further experience that juice obtained under different conditions and from different animals may not act in the same manner. In the estimation of such experiments it is well to bear in mind that Pawlow 71 has shown that the ferment content of the pancreas varies with the diet ; the most striking example of this is the fact that a lactase is found only in the pancreatic juice of sucklings and of animals which have been fed on milk-sugar. Further, it has been observed clinically that injuries occurring to the pancreas during digestion are more serious than injuries during fasting. No other hypothesis than this of the activation of pancreatic juice seems to explain such facts as the following: If the ducts of the pancreas be so arranged that the pancreatic juice flows out into the peritoneal cavity, no symptoms follow; but if the enterokinase containing intestinal secretion be injected into the peritoneal cavity of such an animal whose pancreas is discharging into his peritoneal cavity, death with the typical picture results ; while a control experi- ment with the intestinal secretion alone produces no symptoms (Lattes 49 ). The rate of formation of the toxic products also seems to enter into the problem. Thus Maragiiano 82 shows that a pancreas ground to a fine pulp can be placed in the peritoneal cavity without producing toxic symptoms; whereas the introduction of this same amount of pulp, together with oil, into the peritoneal cavity does produce fatal results. I would explain this result by the experiments reported from my laboratory, which show that the leucocytes of the peritoneal cavity are inhibited in their normal work of phagocytosis and severely injured, if not killed, by the presence of oil. Maragli- 26 INTERNATIONAL CLINICS ano's results are therefore to be explained by the rapid phagocytosis of the finely-divided pancreas, which process is prevented by injuring the phagocytes with oil. In the first instance, the pancreas tissue is removed by the leucocytes before it has autolyzed and activated the proferment ; in the second instance, the inhibition of phagocytosis permits this activation to take place. This leads us, without entering further into the discussion of many of the factors involved, such as the role played by the weak, active, proteolytic ferment of fresh pancreatic juice which was de- scribed by Bayliss and Starling, 30 to the question of main surgical importance: Can the pancreatic secretion be rendered toxic by the substances arising from autolysis \ This particular question seems to find its answer — an affirmative one — most clearly in the work of Lattes. 49 Lattes finds that there is a substance in the autolyzed pancreas (that is, in a pancreas which has been left outside the body for a time until the disintegration of the cells known as autolysis, or self solution, has begun) which has the power of activating fresh pancreatic juice, or, as he prefers to consider the process, which has the power of greatly increasing the proteolytic activity of the juice. This activating substance is more stable toward heat than the proteo- lytic ferment, and can be obtained free from ferment by heating autolyzed pancreas to 60° C. for fifteen minutes. Such an extract, added to pancreatic juice, produces fatal results; such an extract, heated to 75° C. for fifteen minutes — a procedure which destroys this activating substance — added to pancreatic juice, produces no symptoms. It hardly interests the surgeon to enter into further details of the discussion, such as the question of whether the necrosis activates the juice, or the activated juice causes the necrosis ; of whether the toxic action is a property of the pancreatic secretion itself, or if the toxicity is the expression of the toxic properties of the products of the first cleavage of the proteins of the body by the pancreatic secretion. The fact stands clear, that an activated pancreatic secretion, set free in the body, either activated by enterokinase in the normal manner, as the surgeon might find in a case of coincident injury of pancreas and intestine, or activated by a substance produced by the necrosis (re- spectively autolysis) of the gland, causes the fatal symptoms of acute pancreatitis. The further question of whether bacteria can activate SURGERY OF THE PANCREAS 27 the juice or not seems almost secondary, when the fact that the gland- tissue can, on necrosis, produce an activating factor is established; the surgeon knows too well that bacteria can cause a necrosis. THE RELATION OF THE PANCREAS TO HIGH INTESTINAL OBSTRUCTION The clinical similarity between acute pancreatitis and high in- testinal obstruction has led me to the conclusion that this similarity of symptoms may depend upon the fact that the symptoms in both instances are due to the same underlying cause; i.e., the absorption into the body of activated pancreatic juice. 83 I am aware that this suggestion seems contradicted by practically all the experimental work which has been done; all other workers seem to feel that they have excluded the pancreas as a source of the fatal poison in high intestinal obstruction, even though they fail utterly to agree as to the source and nature of the poison. There are numerous reasons why I am convinced that none of these experiments has necessarily ruled out the pancreas as the source of the toxin. A discussion of the long line of experimentation which has led me to this conclusion would, however, lead us too far from our subject, and so I shall mention but one series of experiments, which, I think, confirms my belief. If a closed loop of the lower ileum be made in a dog, by cutting the gut at two places about ten inches apart, closing in both ends of this segment, and restoring the continuity of the intestinal tract by an end-to-end anastomosis of the upper and lower ends of the ileum, the animal will live for months with no disturbance of health. This was first proved by Halstead, 84 many years ago. If now this same operation be repeated, with the single exception that 25 ccm. of nor- mal pancreatic juice be placed in this isolated loop of ileum, the dog will die with the typical symptoms of high intestinal obstruction. This experiment, which has often succeeded, although it may fail if the pancreatic juice is not really active juice, is sufficient proof, to my mind, of the relation between the pancreas and high intestinal obstruction. In short, the symptoms of high intestinal obstruction are the symptoms of Pancreasvergiftung. I introduce this relation of the pancreas at this time for two reasons : First, to point out the extent of the relation of the pancreas to surgery ; and, second, because a further consideration of this ques- 28 INTERNATIONAL CLINICS tion points the way in which these symptoms of pancreas poisoning — shock and fatal intoxication — may be treated until an operation shall have removed the cause. Hartwell and Hoguet 85 believe as a result of their experiments on high obstruction that the cause of death is the dehydration of the tissues consequent upon the excessive loss of water from the body by vomiting. This point of view is, to my mind, untenable, because of the clinical fact that not all experimental animals nor all patients lose fluid in excessive amounts; vomiting is not always present to such excess. However, acting on their theory, they found the highly important fact that a dog which would ordinarily die about eighty-four hours after the production of a high obstruction could be kept alive for so long as twelve days by the introduction into the body of large amounts of normal saline solution. Therefore, in order to combat the shock, which is often so acute and extreme that operation cannot be undertaken, and to aid in the excretion of the poison, I suggest the intravenous introduction of normal saline; be- cause, whether we agree that the poison in the two conditions is one and the same, or whether we disagree, the saline infusion would mechanically raise blood-pressure and aid elimination. Because of the relation pointed out above between the adrenals and the pancreas, I would further suggest the addition of small amounts of adrenalin, continuously administered in dilution in the saline solution, before, during, and, if needful, after the operation for the relief of the obstruction or for the drainage of the acutely inflamed pancreas. THE PATHOLOGY OF THE PANCREAS There are two methods which may be followed in presenting the subject of the pathology of a given organ: To review the subject as though the whole of pathology were founded upon the findings in that organ, entering into a discussion of the numbers of cases of this or that which have been reported, with particular importance granted not to the common conditions but to the rare anomalies. This seems to be a favorite pastime of the so-called " clinical pathologists," yet it is only in so far as the rare case explains the common case that such pastime becomes anything more than harmless. The other method is to treat the pathology of an organ as a chapter of general pathology ; to consider, from the point of view of the fundamental laws of gen- SURGERY OF THE PANCREAS 29 eral pathology, what processes might be expected to occur in a given organ, and to add only those special lesions which depend upon the anatomical or physiological peculiarities of that organ. This latter method is the one that appeals to me; just as the developmental anomalies of an organ depend upon the embryology of that organ, so its pathology depends upon the peculiarities of struct- ure, function, and location, and a knowledge of these peculiarities leads the surgeon not only to a proper concept of the pathology but also of the surgery. Surgery in the broad sense must accomplish not only the removal of disease but also the restoration of normal function. The pancreas is liable to developmental anomalies, both of loca- tion — accessory pancreas; of external form, — annular pancreas, the one of chief surgical interest — and of the arrangement of its ducts, all of which were discussed under the heading of " Embryology." It may be abnormally small — hypoplasia ; since it seems to vary so much in size, an hypertrophic development has not been recorded. The pancreas is a glandular organ of enterodermic origin ; it is there- fore subject to all the ills to which the glandular structures derived from the entoderm are liable ; it has an external function in the form of a secreted fluid, and is therefore subject to the formation of cysts. Since the dangerous factor peculiar to the pancreatic secretion, the powerful protein digesting agent, trypsin, is harmless until trans- formed from its harmless zymogen stage by the action of a substance formed in but two ways — normally, outside the pancreas, in the wall of the intestine; abnormally, inside the pancreas, by the autolysis of the substance of the gland itself — these cysts do not digest their walls, although they may do so if inflammation, with consequent autolysis and activation of the trypsinogen, occurs. Since this secretion is highly alkaline because of the presence of alkaline salts, we may expect, under conditions of retention of the secretion, that a precipita- tion of these salts may result in the formation of calculi, just as in the salivary glands, the kidneys, or the bile-passages. The conditions favoring this precipitation are doubtless the same in all locations; they seem to depend upon the presence in the stagnant secretion of bacteria which by their action start the precipitation of the salts about them. The so-called pseudocyst of the pancreas is a misleading term; by it is meant a collection of fluid in the lesser omental sac, which fluid contains pancreatic products. 30 INTERNATIONAL CLINICS Since the pancreas has blood-vessels, hemorrhage may occur into it; since it has lymphatics which are in direct communication with the lymphatics of neighboring organs, infection and metastasizing tumors may pass from it to neighboring organs, and from neighboring organs to it. Since all of its blood drains into the portal vein, the pancreas must share in the passive congestion consequent upon portal interference; it is normally hypersemic during digestion. Tubercu- losis is rare, congenital syphilis not so rare ; a general fatty infiltra- tion occurs as in other parenchymatous organs, and, with them, the pancreas shares in amyloid degeneration. Tumors may be expected ; adenomas, cystadenomas, carcinomas ; sarcoma is rare. The arrange- ment of the lymphatics, previously discussed, explains fully why primary tumors of neighboring organs metastasize so readily into the pancreas, and likewise why neighboring organs are conversely so readily involved in primary pancreatic malignancy. The surgical interest in the inflammatory processes in the pancreas justifies a separate and more extended treatment. PANCREATITIS Before attempting a classification of pancreatitis, we must be sure that all the factors involved are clearly before us, and that their interaction is definitely understood. Thus Deaver and Pf eiff er 86 finally arrive at a correct conclusion, but by a process including some decidedly incorrect reasoning, when they write : " It is not beyond the range of probability that certain cases of hemorrhagic pancreatitis may be initiated by infection reaching the pancreas through the lymph- channels. It is certain that not all hemorrhagic pancreatitis is pro- duced by the lodgement of a gall-stone in the papilla of Vater, causing retro jection of the bile into the pancreatic duct. A fair percentage of such cases are not accompanied by gall-stones. Neither is it necessary that duodenal contents be regurgitated into the pancreatic duct in order that activation of the retained pancreatic ferments may occur. The hormone secretion arriving by the blood stimulates the secretion of pancreatic juice. Activation of the juice within the gland is not necessary to its digestive action, as the clinical observations of fat necrosis and the postmortem autodigestion of the gland will attest. With infection of the gland, obstruction of the ducts by a resulting engorgement, and stimulation of the active pancreatic secretion by SURGERY OF THE PANCREAS 31 hormone action we have all the conditions necessary for focal necrosis of the pancreas, erosion of the blood-vessels, diffuse hemorrhage, fol- lowed by extensive gangrene and suppuration." Now activation of juice within the gland is not necessary to its digestive action upon fats, as the clinical observations of fat necrosis show, and as is shown by the experimental work upon fat necrosis; the fat-splitting ferment steapsin has nothing to do with the protein- digesting ferment trypsin, and so fat necrosis proves nothing in regard to protein digestion. The fact that the gland does not digest itself during life, but that digestion does set in very soon after death, attests the fact that the death of the cells of the pancreas sets free a factor which can activate the tripsinogen. With infection of the gland, obstruction of the ducts by a resulting engorgement is not necessary, since infection may cause the death of the cells which sets free the activating factor. Stimulation of the active pancreatic secretion by hormone ation is not necessary, since trypsinogen is preformed in the acinar cells and is not formed at the moment of excretion from the cell ; the post-mortem digestion of the gland will attest this fact. With infection of the gland alone we have all the conditions necessary for necrosis and consequent activation of the trypsinogen into trypsin, with the result of erosion of the blood-vessels and local hemorrhage, perhaps ; or diffuse hemorrhage, perhaps ; or extensive gangrene, perhaps ; or we may have infection without activation of proferment, and therefore only suppuration. The autopsy in vivo doubtless supersedes the old-style pathology in the study of the pancreas, since the gland is subject to such rapid and extensive postmortem or even antemortem change; experimental pathology supersedes the autopsy in vivo, since all the conditions can here be controlled, pathogenesis as well as time and extent of an autopsy in vivo or of a postmortem. So as Deaver's autopsy in vivo leads him to the conclusion that pancreatitis, chronic and acute, may be the result of infection, the study and analysis of the experimental results in physiology and pathology which I have endeavored to present in these pages lead to the conclusion that Deaver's surmise is undoubtedly correct. Chronic pancreatitis is the result of an infection of the interlobular connective tissue of the pancreas, therefore it is a lymphangitis. Acute pancreatitis is the result of the setting free into the sur- 32 INTERNATIONAL CLINICS rounding tissues of trypsinogen from the gland-cells and the trans- formation of this trypsinogen into trypsin by one of two factors — either enterokinase, as may happen rarely in coincident injury of pancreas and intestine, or by an activating substance produced when the gland itself undergoes autolysis. This proteolytic enzyme can digest the living proteid of all the structures with which it comes in contact. Since the pancreas is richly supplied with blood-vessels, they are digested along with the rest of the gland, and hemorrhage, either localized or diffuse, is commonly a marked accompaniment of acute pancreatitis. Hemorrhage is not an essential factor, as seen in cases of experimental pancreas poisoning; since the pancreas lies in such close relation with large veins and arteries, an extraglandular, even postoperative, and very severe hemorrhage may occur. This setting free of activated ferment may be brought about by any factor or combination of factors which supplies the two necessary ingredients, setting free of proferment and of activator ; infection, trauma of the gland, either direct mechanical trauma, or indirect trauma, as by embolism, for example, or indirect trauma caused, experimentally at least, by occlusion of the papilla of Vater and forcible injection of bile into the pancreatic duct. Chronic pancreatitis = infection. Acute pancreatitis = liberation of trypsin, by infection or trauma. The only valid classification of infection is one based on the identification of the specific microorganism involved, as in all other pathological conditions of infection. Now such a classification is almost ruled out with the pancreas, especially in acute pancreatitis, since the early peritonitis of adjoining loops of intestine favors the addition of microorganisms from the intestinal tract to the causative microorganism. The rapid postmortem and even antemortem changes in the pancreas so obscure the picture that the pathologist seldom sees the pancreas in the condition to which the surgeon's attention is most frequently drawn. We therefore have the unusual condition in which the pathologist is decidedly limited; and likewise the surgeon is limited, as Deaver admits, when he says he has no definite informa- tion concerning the pancreas in appendicitis, not feeling justified in extending his autopsy in vivo to the extent of examining the upper abdomen as a routine procedure in appendicitis. The difficulties of the problem are increased because of the troubles SURGERY OF THE PANCREAS 33 encountered when we attempt to compare the results of one man's findings with those of another, it being a matter of opinion rather than demonstrable fact. Because of this we can do no better than to follow the reasoning advanced by Deaver and Pfeiffer in support of their conclusion that chronic pancreatitis is a lymphangitis, and to weigh their evidence. The avenues by which infection may reach the pancreas are, according to Deaver and Pfeiffer, 86 (1) the blood stream, (2) the ducts, and (3) the lymphatics. The circulation is ruled out because the diseases with which pancreatitis is associated are not character- ized by bacteremia, nor is the pancreas often involved in conditions of general bacteremia. That this mode of infection is possible is evidenced by the infection of the pancreas in miliary tuberculosis, in syphilis, and occasionally by abscess formation in pyemia. In addi- tion to these conditions, I would call the attention of the reader to the fact that the blood supply of the pancreas is doubtless most developed in connection with the functional gland tissue rather than with the interlobular connective tissue, so that the chance of localiza- tion of a blood-borne infection would be decidedly in favor of the acini rather than the connective tissue. The characteristic of chronic pancreatitis is an interlobular inflammation. The lymphatics form an anastomosing network in this interlobular connective tissue. The ducts of the pancreas, according to Deaver and Pfeiffer, can not be exonerated from all blame. The relation of gall-stones to pancreatitis, especially to acute pancreatitis, was clearly presented by Opie; but all cases of pancreatitis, acute or chronic, are not accom- panied by cholelithiasis, and trauma may present the same picture. While this manner of infection of the pancreas must be admitted, Deaver and Pfeiffer think it has been overworked. To their reason- ing, the logic of which is undoubtedly correct, I would add two further considerations. The operation to be described later under the name of pancreatoenterostomy, first devised by Coffey 87 and later modified in this laboratory, demonstrates that the pancreatic duct can be suc- cessfully anastomosed with the intestine by simply cutting off the end of the pancreas until a fair-sized duct is reached, and dropping this cut end of the pancreas directly into the lumen of the gut, then closing the intestine around the pancreas, the normal duct openings having been ligated. Now such an operation certainly does away with the 34 INTERNATIONAL CLINICS assumed valve action of the normal papilla, which is supposed to protect the duct from ascending infection, yet the pancreas does not become infected. I would further call attention in this connection to the experi- mental results reported by Sweet and Stewart 88 on the subject of the ascending infection of the kidneys. They report, as a result of a long series of experiments, that the lumen of the ureter must be ruled out as the pathway of the ascending infection, and that the lymphatics of the bladder, ureter, and kidney must be looked upon as the route followed by the infection. These conclusions were reached by various experiments; they show that if a section of one ureter be removed and replaced by rubber tubing tied into the ends of the ureter and the bladder be infected, with ligation of the urethra, the resultant infection travels upward to the kidney on the unoperated side, but only to the rubber tube on the operated side. This experiment seems to rule out the lumen of the ureter ; the mucosa of the ureter is ruled out by their final series of experiments, which show that the kidney pelvis can be directly anastomosed with the intestine without a result- ing infection of kidney pelvis or kidney tissue. The evidence that chronic pancreatitis is not necessarily associated with infections of the biliary passages is presented by Deaver and Pf eiffer in the following summary of the cases which they have studied since their attention has been directed to the lymphatics as a possible route of infection : Number of cases Pancreatitis present Percentage Cholelithiasis 99 40 40.4 Cholecystitis 14 9 64.2 Duodenal ulcer 16 2 12.5 Gastric ulcer 3 1 33.3 To this might be added the statistical report of Walther-Sallis, 89 of 250 cases of chronic pancreatitis. He finds 50 cases on record in which there were no associated lesions of the biliary system ; in 27 cases the entire gland was involved, in 23 only the head; in 110 cases cholelithiasis was present, in 75 other cases an angiocholecystitis ; in 15 cases the pancreatitis was associated with an intestinal lesion, ulcer, or cancer. The most convincing argument advanced by Deaver and Pfeiffer in favor of the lymphatics is the fact that in the majority of cases of chronic pancreatitis the entire gland is not affected, as it should be SURGERY OF THE PANCREAS 35 if the infection ascended the ducts, but only the " triangle of pan- creatic inflammation," which lies between the duodenum and the con- verging ducts of Santorini and Wirsung, the lymphatics of which, according to Franke (Fig. 4), can be injected from the gall-bladder. The relation of the inflammation of the pancreas to gall-bladder dis- ease is unquestionable. If the pancreas can be infected through the lymph-paths connecting the gall-bladder and pancreas which have been demonstrated by Franke 5 (Fig. 4), then the pancreas could be infected from almost any point in the abdomen through the lymphatic anastomoses demonstrated by Bartels 3 (Fig. 3). Walther-Sallis finds 90 males and 160 females afflicted; of the 50 cases of non-biliary pancreatitis, 16 were male and 34 females. This writer considers that pregnancy has an etiological relationship to chronic pancreatitis. Acute pancreatitis presents another element of trouble to the solution of the problem of pathogenesis; namely, the wiping out of the evidences of the primary causative factor by the hemorrhage and digestion of the gland. An embolus might cause the primary de- structive changes in the organ, and yet would itself be among the first structures to suffer digestion. We have already considered acute pancreatitis in fact under the heading of " Pankreasvergiftung " ; it was there sufficiently emphasized that the only essential factor is the activation of the proteolytic proferment, tripsinogen, which can be brought about by a substance liberated from the gland on autolysis. JSTordmann 90 has added to the study of his eight clinical cases an experimental study ; he reaches the conclusion above noted in regard to such experiments, that the injection of substances into the pancre- atic duct injures the finest branchings of the ducts and consequently forces pancreatic secretion into the tissues; this process, he thinks, is not what occurs in human pathology. His experiments on dogs led him to the conclusion that for the production of a pancreatitis there must be a simultaneous prevention of the escape of bile and of pan- creatic juice into the intestine, with also a coincident infection of the bile-passages. This point of view, championed by Opie in his classic work, seems to be sadly undermined by the facts mentioned in the discussion of pancreas poisoning — that an acutely fatal pancreas poisoning can be produced by dropping a sterile pancreas into the belly of a normal dog, and that the same result follows the separation 36 INTERNATIONAL CLINICS of a portion of the dog's own pancreas from its blood supply ; whereas the ducts of the pancreas can be ligated with no pancreatitis, provided the blood supply be unharmed; but if any amount of pancreatic tissue be included in the ligature of the duct, a fatal pancreatitis may follow (personal experiments). The clinical fact that numerous cases have been reported of acute pancreatitis following trauma to the pancreas also proves that a hindrance to the flow of pancreatic juice is not a necessary etiological factor. Arnsperger 91 reports three cases of acute pancreatitis associated with cholelithiasis and cholecystitis; there were no inflammatory changes in the bile-passages, the pancreatic ducts, or the duodenum. He believes that these cases, therefore, represented an infection of the pancreas through the lymph-channels. There seems to be no escape from the conclusion that the only essential etiological factor in acute pancreatitis is the activation of trypsinogen into trypsin ; because (1) IsTo other explanation covers the clinical cases associated with infection and the clinical cases associated with traumatism of the gland. (2) ~No other explanation offers an understanding of the fact that injury to the gland during digestion is more serious than injury at other times. (3) The picture of an acute pancreatitis can be fully reproduced by placing the sterile pancreas of one normal dog into the belly of another normal dog; or (4) By isolating a portion of the dog's own pancreas from its blood supply. (5) The ducts of the pancreas can be tied without causing any- thing except simple sclerosis ; but (6) A fatal pancreatitis may follow the ligation of the ducts, if any great amount of gland tissue is so included in the ligature of the ducts that autolysis occurs in the ligated bit of tissue (own experiment) . THE FACTOR OF SAFETY Before proceeding to the removal of all or any portion of an organ the surgeon must know whether that organ is essential to life, and, if so, whether the entire organ or only a part, and what part. This factor of safety is well understood in the case of the paired Fig. 15. The pancreas of the dog, seen from behind. Ch., ductus choledochus; TV., duct of Wirsung; S., duct of Santorini; V., island of pancreas developing from the ventral anlage; the normal line of separation of this island from the main pancreas is marked by the pancreaticoduodenal vessels, normally covered, but here the tissue is drawn apart to show the division between the two parts of the pancreas of the dog. SURGERY OF THE PANCREAS 37 organs of the body — at least one entire half can be done away with, and often more. It is not understood by the surgeon just how far he might go with the pancreas. As we have seen in our discussion of the physiology of the pancreas, we must recognize that the two func- tions of the pancreas are entirely separate : the external function, con- cerned in the elaboration of the specific digestive ferments, has noth- ing to do with the internal function, the one concerned with the control of glucose metabolism. The external function of the pancreas can be, theoretically, entirely dispensed with. The various digestive enzymes are duplicated in so many places throughout the gastro-intestinal tract that the loss of one is not vital; surgeons have long known that the digestive function of the stomach is not a necessity. So with the pancreas, the complete loss of external pancreatic function should result only in an inability to care for food in excessive amounts or unsuitable quality. An individual should therefore be able to exist in comfort on a proper diet. I confess that I was unwilling to accept this teaching of the physiological chemist, and therefore performed a series of experi- ments, of which the following may serve as illustrations : On December 9, 1913, the duodenal portion and the uncinate process of the pancreas of a dog was removed, the body and tail being tied off (as at B, Fig. 15) by a ligature including vessels and duct, and dropped back into the peritoneal cavity to care for the internal function of the pancreas. The dog weighed 11 K. 400 g. The dog recovered completely from the operation, and on December 17, 1913, weighed 11 K. 800 g. On January 3, 1914, the dog showed a marked loss of weight, now weighing 7 K. 200 g. ; the urine was examined on this date and contained no sugar. The stools showed the pronounced fatty, or fatty-acid, characteristic of the stools of an animal with loss of the steapsin of the pancreatic juice ; stools very massive. On March 11, 1914, the dog weighed 6 K. 80 g. ; April 22, 1914, 6 K. 560 g.; May 5, 1914, 6 K. 302 g. ; May 11, 1914, 6 K. 220 g. ; October 4, 1914, 6 K. 220 g. The urine shows no sugar on this date. The animal at this time seemed perfectly well, the only clinical symp- tom being the loss of weight and the typical stools. The animal was killed on this date for autopsy. The autopsy shows two striking changes, an extreme atrophy of the spleen, the organ being only about one-tenth size of the normal spleen of a dog of this size ; the thyroids 38 INTERNATIONAL CLINICS are very pale and translucent. The intestines show a marked brownish pigmentation. There is very little fat in the body. It is interesting to note that the weight remained practically stationary from May to October, 1914. This experiment is given as the type of a large series of similar experiments, in all of which the atrophy of the spleen, the colloidal change in the thyroid, and the loss of weight are constant, as is also the apparent good health, except this loss of weight, of the animals on the mixed diet fed to all the normal dogs. On November 4, 1914, the same operation was performed on a dog. The spleen measured 16 Cm. in length, with width of head of 5 Cm. and width of tail of 3 Cm. (Fig. 12). The animal was autopsied on December 3, 1914 ; the spleen measured 8 Cm. in length, with width of head .of 2.75 Cm. and width of tail of 1.75 Cm. This spleen is shown in natural size on Fig. 12, the larger spleen beside it being drawn from the measurements taken at the operation on November 4, 1914. The thyroids drawn in this plate (Fig. 12) show the peculiar effect of this operation upon the thyroid ; a normal thyroid of the dog is pictured, and beside it the pale, transparent thyroid lobe from the dog just described, the lobe being drawn as it appears when held toward the light, with a scalpel blade behind it to show the transparency. The parathyroids of the upper part of the thyroid lobe, which are normally on the surface in the dog, appear in both lobes ; the lower pair of parathyroids are normally buried in the thyroid tissue and can only be found by cutting serial sections; in this peculiarly transparent thyroid the lower parathyroid can be seen plainly, on holding the lobe against the light. Microscopically the thyroid shows an increase of colloid, with flattening of the alveolar cells. We see, then, that the physiological chemist is right as regards the theory of digestion ; he usually is. Two further facts, however, stand out from these experiments which make it doubtful if the external function of the pancreas can be completely removed, leaving a small portion of tissue to take care of the glucose metabolism ; these facts are the extreme atrophy of the spleen and the changes in the thyroids. This striking atrophy of the spleen and the speed with which it occurs — a marked atrophy occurred in one animal which died three days after operation from acute pancreatitis — and the SURGERY OF THE PANCREAS 39 relation of the spleen to the blood-forming organs, suggest a basis for the diagnostic finding reported by Musser 39 of an acute anaemia in four of eight cases of acute pancreatitis. The animal reported in the first case above was in good health on the ordinary kennel diet at the end of ten months, and was killed in order that the changes occurring at the end of that time might be studied. Whatever the changes in the thyroid and spleen may mean, they can hardly be considered vital. These experiments present for consideration the question of whether, in a case of carcinomatous involvement of the greater part of the pancreas, the surgeon would not be justified in performing the extirpation of the entire gland, excepting a small part of the tail which is tied off to care for the internal function ; such a patient should live in comfort with dietary care, so far as his digestive functions are concerned, and his chance of life would be prolonged, minus the external function of the pancreas, as compared with the certainty of death plus a cancer of the pancreas. The external secretion of the pancreas cannot be separated from the internal function and studied alone. The following experiment is therefore presented, since it demonstrates the extreme limit of removal of pancreas tissue, an extreme to which the surgeon would doubtless never approach, unless it might be in a case where he should choose the lesser of two evils. I present in Fig. 15 a drawing of the normal pancreas of the dog as seen from the posterior aspect of the stomach and duodenum ; the little island of pancreas tissue, the lobe which develops in the dog from the ventral Anlage and contains the duct of Wirsung, which in the dog remains the smaller duct, is shown slightly separated by blunt dissection from the main part of the gland. The ducts of the two parts anastomose in one or two places. This drawing is presented in order to visualize to the surgeon, who does not ordinarily deal in weights and to whom, therefore, the comparative weight of this island, as compared with the entire gland, would convey no definite idea, just how little of the pancreas is needed to completely cover the nitrogen metabolism of the body. On October 9, 1914, the pancreas was removed from a dog, No. 71, leaving this island with its normal blood supply and its own duct. The total weight of the pancreas, plus the estimated weight of this 40 INTERNATIONAL CLINICS island, was 16.9 g. ; the estimated weight of the portion left was 1.2 g. ; the dog died December 15, 1914, and the autopsy showed a portion of pancreas weighing 1.9 g. Practically eight-ninths of the pancreas was therefore removed. The nitrogen metabolism of this dog was studied in comparison Fig. 16. 12 il 10 9 6 7 6 5 t 00 4- o £ 1 * ...j) = *•■ r ; - ■ ► - w BZ /( > / \ > / A i A 1" s /2 /■f /6 ZO Z2 Z Reconstruction in the form of curves of the tables showing the rate of nitrogen excretion in a normal dog. No. 82, and the rate of nitrogen excretion in a dog from which eight-ninths of the pan- creas has been removed, dog No. 71. The curves show that the results are the same in both instances, and that, therefore, the removal of all but one-ninth of the pancreas produces no change in nitrogen metabolism, neither in rate of resorption nor in rate of excretion. with the study of a normal dog. The details of the study of the normal dog, ~No. 82, are given in the first of the following tables ; the results in the depancreatized dog, No. 71, in the second table. The results obtained from the study of both animals are presented in the form of curves, Fig. 16, from which it is seen that there are no differences in the rate of resorption or excretion of nitrogen. SURGERY OF THE PANCREAS 41 Table I Normal Dog No. 82 Date, 1914. Time Number of hours Total nitrogen Nitrogen per hour November 19 12 1.84 0.153 20 12 1.87 0.156 20 8.50-10.50 2 0.346 0.173 20 10.50-12.50 2 0.300 0.150 20 12.50- 2.50 2 0.313 0.156 20 2.50- 9.05 6.25 0.985 0.157 20 9.05- 9.05 12.0 1.89 0.157 21 9.05 a.m.* 21 9.05-11.05 2.0 0.875 0.437 21 11.05- 1.05 2.0 1.66 0.830 21 1.05- 2.35 1.50 1.21 0.807 21 2.35- 5.05 2.50 2.07 0.829 21 5.05- 7.05 2.00 1.72 0.860 21 7.05- 9.05 2.00 1.62 0.810 21 9.05- 9.05 12.00 4.66 0.388 22 9.05- 9.05 12.00 2.32 0.193 22 9.05- 9.05 12.00 2.49 0.207 * Fed 600 gms. of beef-heart=16.8 gms. of nitrogen. Table II Dog No. 11 Date, 1914. Time tt Total Hour3 Nitrogen Nitrogen per hour Total glucose Glucose per hour November 15 8.57- 8.57 12.0 1.27 0.106 0.00 0.00 15 8.57- 8.57 12.0 1.15 0.096 0.00 0.00 16 8.57 a.m.* 16 8.57-10.57 2.0 0.753 0.376 1.61 0.85 16 10.57-12.57 2.0 1.67 0.835 4.63 2.315 16 12.57- 2.57 2.0 1.61 0.805 5.02 2.510 16 2.57- 4.57 2.0 1.64 0.82 6.27 3.135 16 4.57- 8.57 4.0 3.07 0.768 11.50 2.875 16 8.57-10.57 2.0 1.37 0.685 4.93 2.465 16 10.57- 8.57 10.0 4.40 0.440 12.50 1.25 17 8.57-10.57 2.0 0.46 0.230 0.89 0.445 17 10.57-12.57 2.0 0.435 0.217 0.26 0.130 17 12.57- 2.57 2.0 0.290 0.145 0.00 0.00 17 2.57- 8.57 6.0 0.780 0.130 0.00 0.00 17 8.57- 8.57 12.0 1.150 0.096 0.00 0.00 * Fed 600 gms. of beef-heart=16.8 gms. of nitrogen. That portion of nitrogen metabolism represented by the rate of absorption and the rate of excretion is therefore normal in an animal which possesses only about one-ninth of the normal amount of pancreas 42 INTERNATIONAL CLINICS tissue (Fig. 16). A study of Table II brings out the fact that the portion of the protein molecule which is available for transforma- tion into glucose — a portion representing 57.6 per cent, of the entire protein molecule — is disturbed in its metabolism. This is seen in the fact that this animal could care for an amount of protein represented by a nitrogen output of about 0.150 gramme, but an increase over this amount was marked by the appearance of sugar; this was a constant finding in other experiments with this dog and with other similar dogs, and offers the experimental basis for the suggestion that the study of the glucose metabolism is the most promising of the functional tests for the diagnosis of pancreatic involvement. The feeding of glucose and carbohydrates to these animals shows that there is the same sharp line of glucose utilization, regardless of whether the sugar is fed as such, or derived from carbohydrates in general, or from protein. The disturbance is not of sugar, carbo- hydrate in general, or of protein resorption, but of glucose utilization. The results of these studies indicate that two-thirds of the pancreas at least could be removed, and these results, taken in conjunction with the results of the operation described as pancreato-enterostomy, show the practicability of pancreas resection. A study of the record of dog No. 71 indicates that we have gone beyond the limit of safety. Just how large a piece of pancreas is essential in order to preserve normal metabolism might be difficult to determine and would perhaps vary with different individuals. In the following experiment, dog No. 84, roughly four-fifths of the pancreas was removed. A study of the results shows that the pan- creatic control of glucose utilization increases out of all proportion to the relative increase of pancreas tissue. Dog No. 71 was on the very threshold of sugar tolerance. Dog No. 84 can tolerate between 30 and 40 grammes of glucose. At the autopsy of dog No. 71, 1.9 grammes of pancreatic tissue were found ; the piece of pancreas left in dog No. 84 was estimated at operation to weigh about 2.6 grammes. Approximately one-fifth of the pancreas would seem, therefore, able to care for ordinary metabolic requirements. Dog No. 84 December 4, 1914. — Operation performed. Pancreas removed, 13.1 Gms. Con- trol piece, 2.6 +. Total size of pancreas, 15.7 +. December 16, 1914. — 10.30 a.m., 50 Gms. Uneeda biscuit. SURGERY OF THE PANCREAS 43 December 17, 1914. — 8.55 a.m., catheterized and bladder washed. Urine sugar- free. December 17, 1914. — 9.30 a.m., 100 Gms. of Uneeda biscuit. December 18, 1914. — 8. 55 a.m., catheterized and bladder washed. Urine faint re- duction (pseudo). December 18, 1914.— 10.00 a.m., 600 Gms. of beef-heart fed. December 19, 1914. — 3.30 p.m., catheterized and bladder washed. Urine sugar- free. December 21, 1914. — 10.30 a.m., 20 Gms. of glucose in water given per os. December 22, 1914. — 9.00 a.m., catheterized and bladder washed. Urine sugar- free. December 22, 1914. — 11.15 a.m., 30 Gms. glucose as above. December 23, 1914. — 8.55 a.m., catheterized and bladder washed. Urine sugar- free. December 23, 1914. — 11 a.m., 600 Gms. of beef-heart. December 24, 1914. — 8.55 a.m., catheterized and bladder washed. Urine, no sugar. December 24, 1914. — 9.30 a.m., 40 Gms. of glucose given per os. Urine contains sugar. THE DIAGNOSIS There are three cardinal signs in surgical diagnosis: pain, dis- turbance of form or relation, and disturbance of function. Applying these to the pancreas, and bearing in mind the peculiarities of the pancreas which have been described, we find that we could not expect pain to be a definite sign of pancreatic involvement, or, even if it were, pain could not be localized to any one point. The parenchyma- tous organs of the belly are not afflicted with sensory nerves ; they are protected by a sensory device, which consists in the presence in the peritoneal covering of these organs of sensory fibres. Now the head of the pancreas is related to the peritoneum of the duodenum and of the transverse colon, while the tail may be in relation with the peri- toneum of the left kidney or the spleen. The pain of chronic pancre- atitis is therefore found to be of varying degree, perhaps more often referred to the umbilical region, and increased on deep palpation, but sometimes occurring elsewhere in the abdomen ; often a definite pain is not complained of, but a general discomfort. Tumors would not cause pain until they had reached dimensions which produced a strain of the overlying peritoneum. The pain of an acute pancreatitis is essentially the pain of an acute peritonitis of the upper abdomen, and presents no specific element, as compared with the pain of high obstruction, or ruptured gall-bladder, or ruptured stomach-wall, unless, perhaps, the extreme intensity of the pain might be suggestive of 44 INTERNATIONAL CLINICS pancreatitis. 92 It may be noted, in this connection, that the injection of active pancreatic extracts, or of trypsin, is reported to be imme- diately provocative of severe pain. 43 Under the heading of disturbance of form or relations we find sometimes a more or less definite mass in the region of the pancreas in chronic pancreatitis, and sometimes, because of a coincident or causative cholangitis, jaundice and a distended gall-bladder; theoreti- cally, at least, the same jaundice, without history of cholangitis or its accompaniment, cholelithiasis, might be expected, since it is mechanically possible for the swollen head of the pancreas to compress the common duct ; there seems to be considerable difference of opinion as to whether this actually occurs in fact or only in theory. In regard to cysts and tumors, the demonstration of a mass in the pancreatic region is the only definite sign ; perhaps this is also true of abscesses. It must be remembered, however, as shown in Fig. 1, that such cysts and tumors have no definite manner of presenting themselves as regards the stomach and colon ; according to the degree of gastroptosis present, tumors of the pancreas may appear above or below the stomach, or below both stomach and colon ; an abscess of the body of the pancreas may creep along in the retroperitoneal tissue and finally present far below the pancreas. A change of the normal consistence of an organ is a surgical sign which belongs under the general heading of disturbance of form or relation ; such a change of consistence is, for example, of importance in the diagnosis of breast conditions. A change in the consistence of the pancreas, as felt by the surgeon at operation, has often been the decisive factor in the diagnosis of chronic pancreatitis, yet it is difficult to judge the value of this sign. " It should be borne in mind that a hard pancreas is not neces- sarily a diseased pancreas. I know of cases in which the diagnosis of chronic pancreatitis was based on the hard feel of the pancreas at operation, and subsequently at autopsy a normal pancreas was found. It is a matter of common knowledge among pathologists that a normal pancreas may have a consistence of almost stony hardness." I do not find that pathologists are quite ready to agree with Pratt's 93 last sentence; " stony hardness " seems somewhat hyperbolic. Nevertheless, pathologists point out that the pancreas of the normal human being, because of its structure — extensively lobulated, with a SURGERY OF THE PANCREAS 45 large relative amount of connective tissue extending in between the lobules — is a firm, dense structure to the touch, and in respect to this tactile density or hardness is to be placed next the uterus ; beginners in autopsy technic commonly report the normal pancreas as sclerosed or the site of chronic inflammation, because of this natural hardness. 94 For these reasons the pathologists have often doubted the diagnosis made by the surgeon at the operating table, especially since many cases have been noted like the one mentioned above by Pratt, where a pancreas pronounced diseased by the surgeon was later demonstrated by the pathologist to be normal. But, while the surgeon may not appreciate the facts known to the competent pathologist concerning the normal density of the pancreas, the pathologist, on the other hand, does not seem to be cognizant of the fact known to the competent surgeon, that the gland is seldom uniformly affected in chronic pancre- atitis. This fact that the lesion is normally localized to the head of the organ — a fact presented by Deaver and Pfeiffer as proof that the infection is not a duct-borne infection — and that the body and tail are unaffected, so far as can be ascertained by palpation, makes it seem doubtful that the surgeon's sign of change in consistence is so fundamentally an error. Deaver and Pfeiffer report that in 42 cases of the series of chronic pancreatitis under discussion, or 42 out of a total of 52 cases, the head alone was affected ; in only 9 was the tail noted as showing any change in shape, size, or consistence; in 27 cases there was well-marked involvement of the lymph-nodes in the neighborhood of the head of the pancreas. This common localization of the disease process to the " triangle of pancreatic inflammation," lying between the duodenum and the converging ducts of Santorini and Wirsung, affords a means by which the surgeon can control his own findings, and we might conclude that, in order to escape the scorn of the pathologist, the surgeon should limit his definite diagnosis, as based on the consistence of the pancreas, to those cases in which a manifest difference exists between the differ- ent portions of the gland. Under the heading of disturbances of function falls an entire series of tests, all of which must be weighed with a constant considera- tion of the facts which have been mentioned in the foregoing pages concerning the normal physiology of the pancreas, and particularly the facts presented under the heading of the factor of safety ; the con- 46 INTERNATIONAL CLINICS sideration of the factor of safety, at first sight a surgical considera- tion, was presented before this discussion of diagnosis because the facts there related must be borne in mind in weighing the value of any functional test. The pancreatic juice is not secreted continuously, even during digestion, but only when the hydrochloric acid entering the duodenum causes the change of prosecretin into secretin ; the secretin entering the blood stream is carried to the pancreas and incites the gland to activity; but even now the activity of the pancreatic juice is not entirely unfolded ; the steapsin needs the aid of the bile for its com- plete effectiveness, although it develops a certain amount of activity without the bile. The proteolytic activity of the pancreatic juice is developed by the assistance of a further factor, the enterokinase of the intestinal mucosa, which again is not preformed in the intestinal lumen, but only when trypsinogen is present in the intestinal canal. The function of the pancreas is therefore not dependent upon the pancreas alone, but upon the successful collaboration of the pancreas with all the other factors involved: the motor function of the stomach; the formation of hydrochloric acid; the formation of prosecretin and the transformation of prosecretin into secretin ; the effectiveness of the action of this secretin upon the pancreas; the formation of enterokinase and its successful action upon trypsinogen ; the relation of the bile; and, finally, we must remember that the composition of the pancreatic juice depends upon the diet (Paw- low 71 ). Perhaps even fundamental to these considerations is the fact that the factor of safety, surgical and physiological, is so high that we could hardly expect functional tests to give positive results until the disease process had affected a large portion of the gland. The functional tests which have been suggested may be divided into four general groups : 1. Those showing a disturbance of the interrelations between the pancreas and other organs. 2. Those showing a disturbance of the internal function of the pancreas which is involved in glucosuria. 3. Those showing the presence in the stomach content, in the urine, or the faeces of the normal products of the pancreas. 4. Those which indicate the extent of pancreatic function by the SURGERY OF THE PANCREAS 47 extent of the digestion of those substances which are supposed to be digested by the pancreatic secretion only. Under the heading of the tests showing a disturbance of the inter- relations between the pancreas and other organs belongs the test sug- gested by Loewi, the dilation of the pupil after instilling adrenalin into the conjunctival sac. This test is explained by assuming some disturbance of the sympathetic system, and there is little evidence which would lead me to expect a constant result. Sailer 95 reports one positive among 30 cases. The fact that the tonus of the sympa- thetic system is maintained, in some degree at least, by the adrenals, and the facts of the interrelations between the pancreas and the adrenals which have been described in the preceding pages, together with the fact that some go so far as to look upon pancreatic diabetes as really an adrenal diabetes, make this test seem not irrational, but it certainly does not yield practical results. The interrelation between the pancreas and spleen, here pub- lished for the first time, suggests a reason for Musser's finding of four cases of acute ansemia among eight cases of acute pancreatitis ; 39 but Musser's findings have not been confirmed, so far as I find, perhaps because his findings are not generally known. The second group of tests, those showing a disturbance of the internal secretion which controls glucosuria, seems to offer the greatest promise of practical value, judging from the experiments detailed under the discussion of the factor of safety. It has already been pointed out that glucosuria alone is not necessarily indicative of pancreatic disease. Three factors are concerned in diabetes: the pancreas, the liver, and the muscles, so the fault need not necessarily lie with the pancreas. Glucosuria with acidosis — the presence of sugar plus acetone and /^-oxybutyria acid in the urine — is thought by some to represent a definite indication of pancreatic involvement, but acidosis is a late symptom. From the experiments outlined above it would seem that there is a fairly definite relation between the appear- ance of glucose in the urine and the amount of pancreas tissue still functionating. We have seen that there is no disturbance of nitrogen intake, neither in gross amount nor in the rate of resorption. There is a very definite and sharp line marking the point at which the glu- cose^ — either the glucose fed as such or the glucose derived from the protein — ceases to be completely utilized and begins to appear in the 48 INTERNATIONAL CLINICS urine. A lowered sugar tolerance, or, better, a lowered ability to utilize sugar, seems, therefore, on the basis of our experiments, to be the most hopeful method of diagnosing pancreatic insufficiency. Where the Cammidge test belongs in the above classification is doubtful, or even if it is a test relating at all to the pancreas. The physiological chemists have never taken much stock in it; the in- ternists find it positive in health and in a variety of diseases in which the pancreas is normal, and negative when the pancreas is extensively diseased. 93 From the more strictly surgical side, the reports based on the extensive material of the Mayo clinic are negative. 96 The test is included under the heading of the internal secretion, since Cam- midge very recently reports 97 that the substance concerned in his test is dextrin. The tests designed to demonstrate in the stomach content, the urine, or the fseces pancreatic ferments are very numerous. Miiller and Schlecht 98 introduced the serum-plate method, in which the presence of trypsin in the fseces is revealed by the digestion of the surface of a serum-sugar plate, producing a pitting. Gross " bases his test on the precipitation of casein from an alkaline solution on acidifying with acetic acid; casein digested by trypsin does not precipitate. Pratt 93 records his own experience with these tests, and then adds : " That an absence of trypsin in the fseces always indicates disease of the pancreas is doubtful. No trypsin could be demonstrated in the fseces of one of the workers in our laboratory. He presented no symptoms of disease." Trypsin has been sought in the duodenal contents obtained by causing a regurgitation of the duodenal content into the stomach by administering large amounts of oil. 100 This method, introduced by Volhard, 101 will doubtless be superseded by the more exact procedure of using the Einhorn bucket 102 or the Einhorn tube. 103 The final test of these methods which I would consider definite, to find in a given animal if the trypsin content of the duodenum after a given meal and a given time were constant, then to remove a part of the pancreas and repeat the test, seems to be still lacking. A study of the diastase of the fseces may be followed by the method of Wohlgemuth, 104 with numerous modifications. Reports concern- ing the value of this test are inconclusive. Recently Y. Noguchi 105 and Y. JSToguchi and Wohlgemuth 106 report the finding of diastase SURGERY OF THE PANCREAS 49 in the blood and urine of dogs with experimental pancreatic injury, and claim that the amount of diastase is in direct relation to the extent of the injury. Clinical confirmation is still wanting. Opie's one case 107 of the finding of lipase in the urine in acute pancreatitis wanders still alone and unsupported through the liter- ature, although Hewlett 108 showed experimentally that lipase can be found after obstruction of the pancreatic duct or after acute pancre- atitis. The demonstration of an increase in the ethereal sulphates of the urine is supposed to depend upon a lessened growth of intestinal bacteria when the pancreatic juice is absent. 109 The pioneer of the tests of pancreatic function by examining the fasces for substances which are supposed to be digested by pancreatic juice only is the Schmidt's nuclei test. 21 It was pointed out, in discussing the pancreatic nuclease under physiology, that the stomach can digest a certain amount of nucleins, 22 and so can the succus entericus. 23, 24 This test has seen various modifications, which may generally be taken to mean that the test is not entirely satisfactory. A similar test is that of Sahli, 110 who gave glutoid capsules — gelatin capsules treated with formalin, which are not supposed to be digested by the gastric juice — filled with a sub- stance, such as salol or iodoform, which can readily be detected in the urine. These tests are all grossly qualitative, and in all of them a proper valuation of the various factors concerned is difficult, if not actually impossible. It is interesting to note the conclusions of some of the writers who have been most enthusiastic. Thus Pratt concludes 93 : " Pancreatic insufficiency, even of mild degree, can be recognized by the functional methods of diagnosis now available. At the present time no single sign or symptom can be accepted as pathognomonic of pancreatic disease, but by the use of a number of different tests the diagnosis can be made. " The functional tests have already thrown much light on the pathological physiology of the pancreas. Observations made with these tests indicate that the diminished or altered secretion of the pancreas may occur without demonstrable anatomical change." Vautrain, 89 apparently disregarding these functional tests, writes : " Les symptomes de la pancreatite clironique, aujourdhui bien encore mal degages et insuffisamment connus, se grouperont bientot pour con- stituer un tableau clinique defini, grace auquel le diagnostic sera plus 50 INTERNATIONAL CLINICS facile. On decelera alors la pancreatite chronique simple, que Von trouvera frequemment independente de toute lithiase biliaire, et car- acterisee surtout par une tumefaction epigastrique, un peu d'ictere, des douleurs, des troubles digestifs, des nausees ou des vomissements, V amaigrissement et meme la cachexie." The conservative diagnosis of idiopathic chronic pancreatitis will therefore probably not often be made until some real functional test of the pancreas has been discovered. Since the majority of cases are secondary to infections of the bile-tracts, the diagnosis will, by the majority of surgeons, be ventured only in those cases in which the symptoms of cholangitis are associated with a palpable mass in the region of the head of the pancreas, or even to those cases in which, during the course of operation upon the bile-passages, the head of the pancreas is found to have a consistence differing from that of the body of the gland. The more enthusiastic surgeon will be inclined to diagnose the condition on the grounds of the presence of a mass in the pancreatic region plus the symptoms of serious digestive disturbance. And, finally, the enthusiast will diagnose the disease in all abdominal conditions in which the infection of the lymph-channels might be reasonably expected; but, while we shall know of the cases he finds positive, we will never learn the number of false diagnoses he has made in the course of his series. It is probably particularly true of the pancreas that our knowledge would be most effectively advanced if all the failures of diagnosis and of operative treatment were frankly reported. The matter of the diagnosis of acute pancreatitis stands on an entirely different basis. We are dealing with a condition in which early operation offers the only chance for the patient, and in which the differential diagnosis concerns only conditions which can like- wise be treated only by operative measures. While the differential diagnosis between acute pancreatitis and ruptured stomach or duo- denum or gall-bladder might often be easily made on the basis of the case history, it would serve no purpose to delay operation in order to accomplish it. The number of cases which have been opened for high obstruction with the finding of acute pancreatitis, and the reverse, together with the facts brought out in the preceding pages which tend to prove that the symptoms of high obstruction and of acute pan- creatitis are the same, because they are due to the same toxic agent, fully demonstrates the impossibility of a differential diagnosis. The SURGERY OF THE PANCREAS 51 only hope in either condition is immediate operation. So soon as the belly is opened the diagnosis may sometimes be definitely made by the discovery of fat necrosis, which is the only absolute symptom of pan- creatic disease; the finding of a peculiar, dirty fluid is suggestive; the examination of the pancreas establishes the diagnosis. PANCREATOENTEROSTOMY In 1909, Coffey U1 published the result of experiments designed to provide an exit into the intestine for the pancreatic juice in cases in which, for one reason or another, such as malignant disease of the head of the pancreas, or of the common duct, etc., it would be desir- able to completely remove the head of the pancreas. His operation, which he named pancreato-enterostomy, consists in uniting a U-loop of intestine after the manner of a Finney pyloroplasty, and anas- tomosing the pancreas into this loop. The same idea was worked out in this laboratory, but with a much more simple technic. The pro- cedure we adopted was simply to sew the cut end of the pancreas into a longitudinal slit in the intestine. As shown in Fig. 15, the pancreas of the dog is an extremely elongated organ, admirably suited to such an operation. The end of the uncinate process of the pancreas was cut off, as at A, Fig. 15; the end of the pancreas was then attached to the gut by a row of fine Lembert sutures, placed at a little distance from the cut end around the part which would later lie underneath the pancreas. This stitch compares to the first row of stitches in a gastro-enterostomy. A longitudinal slit is then made through the wall of the intestine, the end of the pancreas is tucked into the slit, and the Lembert stitches continued around the pancreas to the point of beginning — comparable to the fourth row of a gastro- enterostomy. The operation is precisely that of an end-to-side in- testinal anastomosis, excepting only the fact that the end of the pancreas, corresponding to the end of gut in an end-to-side anastomo- sis, is tucked into the lumen of the gut, instead of being simply approximated to the wall. The results of these experiments were as follows: In a series of three dogs the uncinate process was cut off at A, Fig. 15, and the duct of the proximal end ligated. The cut end of the distal portion of the pancreas was implanted into the gut by the method described. All three showed an uneventful and perfect clin- 52 INTERNATIONAL, CLINICS ical recovery; autopsies after six, five, and four weeks, respectively, showed that the duct was patulous in all three dogs and of normal size. There was no fat necrosis, no evidence of leakage or suppura- tion, and few adhesions. In two there was no atrophy of pancreatic Fig. 17. Pancreato-enterostomy. The pancreas, after amputation of the uncinate process, being fastened to the wall of the intestine by a running Lembert suture. tissue ; one showed a shrinkage to one-third of the original size of the transplanted piece. In a second series of three dogs both ducts were tied with double ligatures and cut between the ligatures. Instead of transplanting the duct of the uncinate process, the uncinate process was amputated and the proximal end of the pancreas was anastomosed with the gut in the manner described above (Figs. 17, 18, and 19). Two of these dogs were killed and autopsied live weeks after operation; both SURGERY OF THE PANCREAS 53 had made an uneventful recovery and had shown no clinical symp- toms; in both cases a new duct had formed about the double ligatures of the main pancreatic duct, and the implanted duct was not patulous. There was no evidence of fat necrosis, leakage, or sup- puration, or of pancreatitis or atrophy. The third dog of this series Fig. 18. Pancreato-enterostomy. The first row of Lembert sutures is completed and the gut incised. died a week after operation from distemper pneumonia. The autopsy showed a perfect operative result, except that even in this short space of time a new duct had formed about the ligatures. In a third series of two dogs the re-formation of the ducts was prevented by ligating in two places, cutting between the ligatures, and then drawing omentum between the cut ends of the duct. The 54 INTERNATIONAL CLINICS proximal end of the pancreas was then anastomosed with the intestine as in series 2. Four weeks after the operation, which was followed by uneventful recovery, the autopsy showed that the implanted duct was patulous, the ligated ducts permanently closed; there was no atrophy of the pancreas, the gland showed normal consistence, there were no evidences of fat necrosis or leakage or suppuration, and few adhesions. The technic in the third dog of the series was slightly Fig. 19. Pancreatoenterostomy. The ends of the pancreas have been tucked into the slit of the intes- tinal wall, and the Lembert suture has been continued to the point of beginning. The small sketch shows the appearance of the completed operation from the inside of the intestine. varied ; the cut end of the pancreas was implanted in a cut of the bowel wall which extended only to the mucosa, since we wished to find if the pancreas, by virtue of the digestive action of its secretion, would provide an escape for itself through the mucosa. The animal died at the end of a week, and the autopsy showed that the digestive action of the secretion was indeed present, but that it had turned upon the pancreas itself. Sufficient necrosis or autolysis of the cells of the cut end of the pancreas had developed to activate the tryp- SURGERY OF THE PANCREAS 55 sinogen — or possibly enterokinase had gained entrance from the cells of the mucosa — the result being acute pancreatitis marked by fat necrosis, and death. This same operation was performed on another animal on Decem- ber 13, 1913. On November 14, 1914, the animal, which was in Fig. 20. The result of the operation of pancreatoenterostomy, eleven months after operation; drawing of Kaiserling specimen, a. Probe in the closed papilla of the main duct. 6. Double loop of silk- worm-gut, passing through the new opening of the duct and out through the duct, which had been slit open to see if it had become dilated. perfect health, was autopsied in order to determine the result of the operation. The pancreas was entirely normal, the original duct was closed, and the new duct opening of ample size was found at the bot- tom of a little dimple of the intestinal wall (Fig. 20). The results of these experiments show that: 1. There are no valves in the pancreatic ducts, and the flow can be reversed. 2. Nature is surprisingly insistent that the normal exit be pre- 56 INTERNATIONAL CLINICS served, and will quickly overcome even the barrier of a double ligature and excision between the ligatures. 3. The entire pancreas, or a portion of the pancreas, can be joined to the intestine without danger of pancreatitis, either chronic or acute, and without danger of sclerosis for so long as eleven months. 4. Ascending infection of the pancreas is certainly not prevented by the assumed valve action of the sphincter of the papilla, or the assumed mechanical valve action of the papilla itself ; nor is it likely, in view of these experiments, that infection ascends up the lumen of the duct, or that it ascends upwards along the mucosa of the duct. This operation, therefore, taken in conjunction with the facts presented under the heading of the factor of safety, lead to the con- clusion that at least two-thirds of the pancreas could be safely removed and the duct of the remaining portion could be safely and simply anastomosed to the intestine. The human pancreas is, of course, larger in proportion to the human intestine than is the pancreas of the dog in proportion to the dog's intestine ; but the study of the ex- periments described above leads me to the conclusion that the pancreas could be safely anastomosed to an opening of the intestine, even though the end of the pancreas did not project into the intestinal lumen as in these experiments. THE SURGERY OF THE PANCREAS If the method of reasoning which has now brought us to the ques- tion of the actual surgical procedure is logical and correct in its basic ideas, the discussion of the surgery of the pancreas can be reduced to simple terms. If we agree that chronic pancreatitis is a lymph- angitis secondary to some primary focus outside the pancreas, then the primary focus must be found and treated. If we agree that the fatal element in acute pancreatitis is the active proteolytic ferment of the pancreas, which ferment is made active by the addition of a product which develops on autolysis of the gland substance, then two things must.be always borne in mind — prevention of the autolysis of the pancreas tissue which would liberate the activating factor, and thorough drainage of the pancreatic secretion which may have escaped into the surrounding tissue ; or, in simpler terms, removal of infection or drainage, and prevention of the escape of the secretion of the gland ; or, if it has already escaped from its normal confines, drainage. More than ordinary care must be used to prevent tying off bits of SURGERY OF THE PANCREAS 57 gland tissue with ligatures. Drainage is more fundamental with the pancreas than with any other organ ; not only an infection must be drained, but also the normal product of the organ. A study of the reports of the operative procedures on the pancreas will establish the truth of this reasoning, even though these procedures have not always been based on a clear concept of the principles involved. Let us begin with the more simple surgical operations, rather than follow the usual method of beginning with the more simple patho- logical conditions. We find all surgeons agreed that a drainage of the biliary system represents the curative procedure in chronic pancreatitis. We have seen that a simple drainage of the gall-bladder cannot necessarily drain the pancreatic ducts, even if the infection had ascended the lumen of the ducts, since the mode of development of the ducts and their diverse manner of entering the duodenum may make it a physical impossibility to drain the pancreas in this manner (Fig. 7). But, furthermore, agreeing that chronic pancreatitis is a lymph- angitis, we must agree that the infection is a deeper lying one than the mucosa of the gall-bladder or the mucosa of the ducts; it must be in the wall of the bile-passages. Therefore we may disregard the various propositions which have been made, exploratory laparot- omy alone, or drainage of the gall-bladder, or cholecystenterostomy, or cholecystogastrostomy (Jaboulay 112 ), the reason being perfectly clear why Mayo reports: 113 " In at least one-half of the cases operated on in our clinic the following sequence has occurred : cholecystostomy had been done for chronic cholecystitis without stones, and with a complicating chronic pancreatitis. The patient was relieved for some weeks or months and then the symptoms returned. Recognizing the need of more pro- longed drainage, the gall-bladder was reopened and drained for a considerable period. There was complete relief so long as drainage of the gall-bladder continued, but sooner or later, after the fistula in the gall-bladder healed, the symptoms returned. " These cases are characteristic, and I have no doubt have been observed by many surgeons who have been puzzled to know just what course to pursue. It has been our experience that removal of the gall-bladder primarily relieves the symptoms and permanently cures the patient. Just what the future condition of the pancreas may be one has no means of knowing, but I have found that chronic pancre- 58 INTERNATIONAL CLINICS atitis, the result of gall-stone disease, is usually cured by the removal of the stones and drainage of the biliary tract, and that in the chronic infection of the gall-bladder with secondary involvement of the pancreas, in the absence of infections with biliary drainage, chole- cystotomy furnishes a satisfactory symptomatic cure." The proper procedure in so-called idiopathic chronic pancreatitis is not clear, nor will it be clear until we are agreed whether such a condition really exists. I personally suspect that all cases of chronic pancreatitis represent a chronic lymphangitis secondary to some other focus of infection; primary infection of the glandular tissue of the pancreas would lead to suppuration with certain microorganisms, while with others it would lead to necrosis, autolysis, activation of trypsinogen, and acute pancreatitis. There is certainly no reason why the numerous lymph-channels, anastomosing with surrounding structures, should not carry infection toward the pancreas as well as do the lymphatics of the bile-tracts. The role played by a small stone lodged in the ampulla of Vater in the production of both chronic and acute pancreatitis justifies the routine careful examination of the ampulla, even to the extent of direct examination through an incision of the duodenum. We have seen that acute pancreatitis can only be explained by the facts of the activation of the proteolytic proferment. We have seen that this activation can be brought about by the enterokinase of the intestine or by a product formed on the autolysis of the gland itself, and that the inactive juice is not toxic. A consideration of these facts makes clear the surgical procedures indicated when handling the pancreas — to avoid trauma which might result in autolysis, to prevent the escape of pancreatic secretion, which, while itself harmless, might become activated, and, in the presence of activated secretion, acute pancreatitis, to provide ample opportunity for its escape — thorough drainage. We find in the report of von Mikulicz 114 the records of 75 cases of operation for acute pancreatitis, in 36 of which the pancreas was directly attacked, with 25 recoveries, while in 41 where the pancreas was not attacked only 4 recovered, and in these free drainage was provided. In other words, the toxic product is not in the peritoneal cavity, but in the gland, and a means of escape for it must be provided. If we feel that such statistical records are of doubtful value, since it is impossible to compare the work of different surgeons, even in the SURGERY OF THE PANCREAS 59 same field, let us turn to Korte, 40 who has had the largest experience in acute pancreatitis of any single surgeon. His opinion, based on the study of 44 cases, is expressed as follows : " So empfehlen jetzt die meisten Chirurgen, das entzundete Pankreas freizulegen. Wahrend nun Einige sich damit begnugen wollen, nur die Oberflache desselben durch Gaze und Rohr zu drainiren, rathen Andere, in das erkrankte Organ einzudringen. Wie ich oben auseinandergesetzt habe, schliesse ich mich den Letzteren an und glaube, dass man dadurch die Entzundung mildern und ausge- dehnterem Absterben vorbeugen kann." As to the time of operation, there can be no doubt, on the basis of the reasoning thus far followed, that the earlier the operative inter- ference, the better. The destructive process once started may un- doubtedly be limited and walled off by nature, though it may just as well involve the whole gland, and there can certainly be no way of determining the extent, or of limiting the extent, except by operation. Definite information on this point is given by Korte 40 ; of sixteen cases operated within the first two weeks after the beginning of the disease, eleven recovered and five died. Of the fourteen operated within the third and fourth weeks, seven recovered and seven died. The four operated in the fifth to the sixth week all died. Deaver 1 15 would not undertake operation in the state of primary shock. Because of the experimental results in high obstruction which have been described in the preceding pages, and because of the relation between the pancreas and the suprarenals, and the role of the suprarenals in shock, I would suggest the use of intravenous saline infusion with adrenalin in dilution in the saline to combat this shock and make operation possible, beginning the infusion before the operation, con- tinuing it during the operation, and after if necessary. The recognition of the toxic factor in the pancreas leads to the question of the manner of treatment of pancreatic cysts and of wounds of the pancreas. The pseudocysts of the pancreas cannot be extirpated, since their walls are formed from the tissues of the parts adjacent to the pancreas; they can be attached to the parietes and drained — " marsupialized." True cysts — cysts in connection with the ducts of the pancreas — must be extirpated, since if drained they leave fistulse which will not close (Korte 116 ). The treatment of wounds of the pancreas must be conducted on 60 INTERNATIONAL CLINICS the same lines. Given a case of complete transverse rupture, which cases are fairly numerous in the literature, 12, 13, 14 the only rational treatment would consist in either complete removal of the distal portion of the gland, which, as we have seen in the discussion of the factor of safety, is a perfectly rational surgical suggestion ; or else, since technically this is no easy suggestion, I would propose the liga- tion of both ends of the torn duct. The distal portion of the gland itself, having been ligated, would suffer sclerosis of the excreting cells, while enough of the elements of the pancreas would persist to aid in the preservation of the internal functions of the pancreas, as we have seen in the preceding discussion of the experiments in which the external function of the pancreas was completely removed. The traumatized portions of the gland should be carefully removed and drainage should be carried to the pancreatic stumps, either through an anterior excision, or, perhaps, even better, drainage would be more thorough by draining through a stab wound through the back. The importance of removing traumatized tissue has several times been made evident in our experiments upon the complete removal of the external function of the pancreas. I have seen a fatal pancre- atitis which evidently started in the portion of the gland which had been exposed to autolysis by the ligature which tied the duct. The procedure to be followed in dealing with tumors of the pancreas is undoubtedly, as elsewhere stated, complete extirpation. The location of the pancreas itself and the involved anatomical rela- tions; the free anastomoses of its lymphatics with so many regional lymph-nodes, together with the late time of diagnosis, make the hope of early radical removal of carcinoma seem very remote. The opera- tion of pancreatoenterostomy described above offers a solution for the treatment of non-malignant tumors of the head of the pancreas, as well as the suggestion for the symptomatic treatment of malignant tumors in which secondary symptoms of pressure might justify a palliative procedure. CONCLUSIONS Our inventory is complete. On assembling our notes of stock on hand, we find that the pancreas is an admirably protected and there- fore surgically difficult organ. The factor of safety is high, so that removal of two-thirds of the pancreas, at least, and probably more, is perfectly possible, the only difficulty being one of technic. SURGERY OF THE PANCREAS 61 This very factor of safety makes it clear that the tests of the func- tional activity of the pancreas can hardly be expected to be of much value in the diagnosis of early pancreatic involvement. The duct of the pancreas can be transplanted without danger of an ascending infection. Chronic pancreatitis is a lymphangitis, probably always secondary to a focus of infection in some neighboring organ, the lymph-vessels of which anastomose with the lymph-vessels of the pancreas. The treatment of chronic pancreatitis is therefore the surgical removal of the primary focus of infection. Acute pancreatitis is a disease process determined by the terrible digestive power upon the tissues of the living body of the proteolytic ferment of the pancreas, which ferment may be set free and trans- formed from the harmless zymogen stage in which the ferment is normally found in the cells and ducts of the pancreas by infection or by sterile autolysis. Infection does not necessarily produce an acute pancreatitis; an infection which caused the toxic death of the cells of the pancreas would compare with autolysis, whereas an infection which determined the complete dissolution of the cells might destroy the zymogen also, and result in suppuration and localized abscess formation. We find in this consideration of the pancreas the importance to surgery of what we may call a new principle — the avoidance not only of the septic digestion of tissue, but also the avoidance of sterile self- digestion, or autolysis. Not only in the pancreas but in the brain can a fatal process be loosed by autolysis : in the pancreas, acute pancre- atitis; in the brain, cerebral softening, as after cerebral hemorrhage. The prevention of autolysis can be accomplished by avoiding trauma — the direct trauma of rough handling or the trauma of ligat- ing a mass of tissue instead of the actual bleeding vessel — and by the careful removal of all pancreatic tissue which has been crushed or cut off from its blood supply. Once started, this autolytic process and its consequent setting free of the protein-digesting ferment of the pancreas can be treated only by drainage, drainage not only of the infection which may have been the causative factor, but drainage of this digesting fluid, drainage of the gland itself and of the tissues into which the secretion of the gland has penetrated. 62 INTERNATIONAL CLINICS BIBLIOGRAPHY 1 Piersol: "Human Anatomy," Philadelphia, J. B. Lippincott Company. 2 Do Rio Branco : " Anatomie et medicine operatoire du tronc coeliaque," etc., Paris, 1912. 3 Bartels: Arch. f. Anat. u. Phys., Anat. AMI, 1904, 299; 1906, 250; 1907, 267. *Oser: Nothnagel's Spec. Path. u. Ther., Vienna, 1898. * Franke: Deutsch. Zeitsch. f. Chir., September, 1911. "Pensa: Boll. Soc. med.-chir., Pavia, 1904. Inter. Monatschr. f. Anat. u. Phys., 1905, 22, 90. t Sobotta: "Anat d. Bauchspeicheldriise," Jena, 1914. (Bardeleben's Handbuch d. Anat. d. Menschen, Band vi, Abt. 3, 1.) 8 Sauerbeck: Ergeb. d. Allg. Path. u. Path. Anat., 1902, Abt. ii. •Sauerbeck: Virchow's Archiv, 1904, 177, Suppl. Heft. 10 Koch: Virchow's Archiv, 1913, 211, 321. "Guleke: Hunch, m. Wchnsch., 1910, 75. "Heineke: Arch. f. klin. Chir., 1907, 84, 4. "Garre: Beitr. z. klin. Chir., 1905, 46, 1. 14 Walther et Guillemin : Congres francais de Chir., 1905. "Hamburger: Anat Anzeiger, 1892, 7, 707. 16 Opie: American Medicine, 1903, 5, 996. "Letulle et Nattan-Larrier : Bull. Soc. Anat., Paris, 1899, 987. "Cords: Anat. Anzeiger, 1911, 39, 33. "Broman: Ibid., 1913, 44 Erganzungsheft, 14. ^Bayliss and Starling: Journ. of Phys., 1902, 28, 325. "Schmidt: " Functionspriifung des Darmes," Wiesbaden, 1904. 22 Umber: Zeitschr. f. klin. Med., 1901, 43, Heft 3 and 4. ^Gumlich: Zeitschr. f. phys. Chem., 1893, 18, 508. m Araki: Zeitschr. f. phys. Chem., 1903, 38, 84. ^Delezenne: C. R. de la Soc. de Biol., 1902, 282, 283, 328, 590, 693, 890, 893, 896, 998. ^Bayliss and Starling: Journ. of Phys., 1902, 28, 325. "Hekma: Arch. f. Anat. u. Phys., 1904, 343. 28 Heidenhain : Pfliiger's Archiv, 1875, 10, 557. 28 Vernon: Journ. of Phys., 1901, 27, 269. ^Bayless and Starling: Journ. of Phys., 1903-04, 30, 61. 31 Allen: " Glucosuria and Diabetes," Boston, 1913, 68 pages of literature references. 32 Sweet: Journ. of Med. Research, 1903, 10, 255. 33 Sweet and Pemberton: Archives of Int. Med., July, 1908; May, 1910. 84 Sweet: Annals of Surgery, October, 1910. 35 Sweet and Allen: Ibid., April, 1913. ^Cushing: "Pituitary Body and its Disorders," 341 pp., 112 pi., 8vo, 1912. 37 Herzen : Arch des sciences physiques et naturelle, 1897, 4. 38 Gachet et Pachon: These de Bordeaux, 1897. ^Musser: Univ. of Penna. Med. Bull., 1909. "Koete: Arch. f. klin. Chir., 1911, 96, 557. "Doberauer: Bruns. Beitr. zur klin. Chir., 1906, 48, 2. 42 von Bergmann und Guleke: Munch. Med. Wochen., 1910, 1673. 43 The following references, not specifically referred to in the text, all bear on the subject of Pancreasvergiftung : von Bergmann: Ztschr. f. exp. Path. u. Ther., 1906, 3, 401. SURGERY OF THE PANCREAS 63 Brugnatelle: Boll. Soc. med-chir., Pavia, 1910, 49. Bunge: Arch. f. klin. Chir., 1903, 71, 726. Dittrich: Viertlj. f. ger. Med., 1890, 43, 52. Doberauer: Arch. f. 'klin. Chir., 1906, 79, 4. Dreesmann: Ztschr. f. drztl. Fortbildung, 1912, 9, 5. Egdahl: Journ. of Exp. Med., 1903, 9, 4. Flexner: Journ. of Exp. Med., 1897, ii. Guleke: Arch. f. klin. Chir., 1906, 78, 44. Hess: Munch. Med. Wochen., 1903-1905. Hess: Grenz. d. Med. u. Chir., 1909, 19, 637. Hildebrand: Arch. f. klm. Chir., 1898, 57 Hlava : C. R. du XII Congres Inter, de Med. Moscow, 1897, III Section, 106. Kircheim: Arch. f. Exp. Path. u. Pharm., 1911, 66, 352. Kirste: Nihrnberger med. Ges., December, 1902. Lattes: Arch, di farmacol. speriment. e sc. aff., 1912, 13. Lattes: Pathologica, 1912. Lewit: I. Diss. Konigsberg, 1906. Maragliano: Policlinico sez. Chir., 1912, 19, 49. v. Mikulicz: Grenzgeb. d. Med. u. Chir., 1913, 26, 1. Polya: B. K. W., 1906, 49. Pfluger's Arch., 1908, 121, 483. Grenzgeb. d. Med. u. Chir., 1911, 24, 1. Roosen-Runge: Zeitschr. f. klin. Chir., 1901, 45, 56. Rosenbach: Arch. f. klin. Chir., 1911, 94, 2. Seidel: 38 Vers. d. d. Ges. f. Chir., Berlin, 1909. Zentralbl. f. Chir., 1910, 51. Simmonds: Munch. Med. Wochen., 1898, 6. "Balser: Virchow's Archiv, 1882, 90, 520. 45 Erdmann: Annals of Surgery, 58, ii, 688. 48 Kirste : Nurnberger med. Gesell., December, 1902. "Heinecke: Arch. f. klin. Chir., 1907, 84, 4. ^Erdmann: Trans. N. Y. Surg. Soc, May 14, 1913; Ann. of Surg., 58, ii, 688. ^Lattes: Virchow's Archiv, 1913, 211, 1. 60 Blume: Festschrift zur Naturforcherversammlung, Braunschweig, 1897. "Oser: "Nothnagel's Handbuch," etc., 1898, 18. B2 Milisch: I. Diss. Berlin, 1897. 53 Lewit: /. Diss. Konigsberg, 1906. "Wolff: Verhandl. d. Ges. deutsch. Naturforsch. u. Arzte, 71 vers., Miinchen, 1899, ii, 2, 550. 65 Hess : Munch, med. Wochen., 1903, 1905. bb Gtjleke: Arch. f. klin. Chir., 1906, 78, 44; 1908, 85, 43. "Eppinger: Ztschr. f. exp. Path. u. Ther., 1905, 2, 216. 08 Flexner: Journ. of Exp. Med., 1897, ii. 59 Opie : " Diseases of the Pancreas." 60 Polya: B. k. W., 1906, 49. 61 Hlava : C. R. du XII Congres Inter, de Med., Moscow, 1897. 62 Flexner and Pearce: V. of Penna. Med. Bull., 1901, 94. 83 Hildebrand : Arch. f. klin. Chir., 1898, 57. "Rosenbach: Arch. f. klin. Chir., 1909, 89, 2. 65 Carnot: These de Paris, 1898. m Thiroloix: Bull, de Soc. anat. de Paris, 1891, 573. "Panum: Virchow's Archiv, 1862, 25, 308. 68 Lepine: Lyon med., 1892, 302. 64 INTERNATIONAL CLINICS 69 Bunge: Arch. f. Mm. Chir., 1903, 71, 726. 70 Katz und Winkler: Arch. f. Verdangskrank., 1898, 4. "Pawlow: "The Work of the Digestive Glands," London, 1902. 72 Lombkoso: Giornale del ace. di med. di Torino, 1903, 66, 225. 73 Roger et Garnier: C. R. Soc. Biol, 1905, ii, 388, 674, 677. 74 Falloise: Arch, intern, de Phys., 1905, ii, 299. 70 Cybulsky et Taschanoff: Arch. int. de Phys., 1907, 5, 257. "Roger et Garnier : C. R. Soc. Biol., 1908, ii, 1910. "Fleig: C. R. Soc. Biol., 1908, ii, 718. 78 Seidel : 38 Vers. d. d. Ges. f. Chir., Berlin, 1909. 79 Kirschheim : Arch. f. exp. Path. u. Pharm., 1911, 66, 352. 80 Schittenhelm und Weichardt : Miinch. med. Wochen., 1911, 843. 81 Fragoin e Stradiotti: Arch, per le Sc. med., 1910, 34, 38. 82 Maragliano : Policlinico sez. chir., 1912, 19, 49. 83 Sweet: Penna. Medical Journal, April, 1913. 84 Halsted: Am. Journ. Med. Sci., 1887, 94, 436. 86 H'artwell and Hoguet: Am. Journ. Med. Sci., 1912, 143, 357. J. A. M. A., 1912, 59, 82. 86 Deaver and Pfeiffer: Annals of Surgery, 1913, 58, ii, 151. 87 Coffey: Annals of Surgery, 1909, 50, 1238. 88 Sweet and Stewart: Surg., Gyn. and Obst., April, 1914. 89 Walther-Sallis : Rev. di Chir., 1913, 48, 2, 907; 1914, 49, 1, 446. 00 Nordmann : Arch. f. Min. Chir., 1913, 102, 66. 91 Arnsperger: Deutsch. Ges. f. Chir., 1913. 82 Archibald and Mullally: Canadian Med. Journ., February, 1913. 98 Pratt: Am. Journ. Med. Sci., March, 1912. 94 Pearce : Personal communication. 96 Sailer: Am. Journ. Med. Sci., 1910, 140, 330. 96 Pilcher: Annals of Surgery, 1910, 51, 89. Wilson: Surg. Gyn. and Obst., 1910, 11, 156. 97 Cammidge: J. A. M. A., 1914, 63, 2063. 98 MtJLLER und Schlecht: Munch, med. Wochen., 1908, 55, 225. 99 Gross : Arch. f. exp. Path. u. Pharm., 1907, 58, 157. 100 Bold yreff : Centralbl. f. Phys., 1904, 18, 457. 101 Volhard: Miinch. med. Wochen., 1907, 54, 403. 102 Einhorn: N. Y. Med. Journ., 1908, 87, 1179. 103 Einhorn: Med. Rec, 1910 77, 98. 104 Wohlgemuth : Biochem. Zeitschr., 1909, 30, 432. 105 Noguchi, Y.: Langenbeck's Archiv, xcviii, Heft 2. 106 Noguchi, Y., und Wohlgemuth: B. 'k. W., 1912, No. 23. 10T Opie: Johns Hopkins Hosp. Bull., 1902, 18, 117. 108 Hewlett: Journ. Med. Res., 1904, 6, 377. 109 Tileston: Trans. Ass. Am. Phys., 1911, 26, 522. U0 Sahli: Deutsch. med. Wochen., 1897, 23, 6. 111 Coffey : Annals of Surgery, 1909, 50, 1238. U2 Jaboulay: Lyon Med., 1898. 113 Mayo, W. J.: Am. Journ. Med. Sci., 1914, 147, 469. 114 Von Mikulicz: Annals of Surgery, 1903, 38, 1. 1X5 Deaver: Am. Journ. Med. Sci., 1909, 138, 829. 11g Korte: Deutsch. med. Wochen., 1914, 40, 424. U7 Loewi: Arch. f. exp. Path. u. Pharm., 1909, 59, 83. COLUMBIA UNIVERSITY LIBRARY This book is due on the date indicated below, or at the expiration of a definite period after the date of borrowing, as provided by the rules of the Library or by special ar- rangement with the Librarian in charge. DATE BORROWED DATE DUE DATE BORROWED DATE DUE £C7 1 6 '39 C2S(239)MIO0 "C0«A UNIVERSITY LIBRARIES ,hsl,tx, RD591S\N3d ™ s ii|p|i»