COLUMBIA LIBRARIES OFFSITE HEALTH SCIENCES STANDARD HX00033561 FisKE Fund Prize Essay. No. LVIII. RECENT CLASSIFICATION AND TREATMENT OF PNEUMONIA, MOTTO: And they shall beat their swords into plowshares, and their spears into prun- inghooks: nation shall not lift up sword against nation, neither shall they learn WAR ANY MORE. — Isaiah II:4. Author : HARRY S. BERNTON, A. B., M. D. Washington, D. C. Columbia ^niberfiitp intteCitpof^etol^orfe CoUese of l^f)psimns anb ^urgeoniee ^tltvtntt Hibrarp FisKE Fund Prize Essay. No. LVIII. RECENT CLASSIFICATION AND TREATMENT OF PNEUMONIA, MOTTO: And they shall beat their swords into plowshares, and their spears into prun- inghooks: nation shall not lift up sword against nation, neither shall they learn war any more. — isaiah ii : 4. Author: HARRY S. BERNTON, A. B., M. D. Washington, D. C. Snow & Farnham Co., Printers Providence, R. I. 1919 TTHE Trustees of the Fiske Fund, at the annual meeting of the •'■ Rhode Island Medical Society, held at Providence, June 5, 1919, announced that they had awarded a premium of two hundred dollars to an essay on " Recent Classification and Treatment of Pneumonia," bearing the motto: And they shall beat their swords into plowshares, and their spears into pruninghooks: nation shall not lift up sword against nation, neither shall ■ they learn war any more. — Isaiah ll:4. The author was found to be Harry S. Bernton, A. B., M.D.,of Washington, D. C. DR. GARDNER T. SWARTS, Providence, R. I., DR. JOHN M. PETERS, Providence, R. I., DR. JESSE E. MOWRY, Providence, R. I., Trustees. HALSEY DeWOLF, M. D., Secretary of the Trustees, 305 Brook Street, Providence, R. I. Digitized by the Internet Arciiive in 2010 witii funding from Open Knowledge Commons http://www.archive.org/details/recentclassificaOObern Contents page Introduction ....... 7 Pneumococcus 9 Part I. Basis of Classification Methods of "Typing" Type I ..... . Type II and sub-groups . Type III Type IV Precipitable Substances And Spinal Fluids Blake's Optimum Dilution Methods . Rapid Methods of "Typing" 12 15 i8 i8 20 21 21 24 28 29 Part^II. Incidence of Types in Pneumonia Mortality of Types in Pneumonia Incidence of Types in Children Incidence of Types in Normal Persons Incidence of Types in Contacts . Incidence of Types in Convalescents . Occurrence of Pneumococcus in Dust Viability of Pneumoccous in Sputum . 31 34 35 37 39 42 44 45 Part III. Sero-therapy in Pneumonia Production of Antipneumococcus Serum i . Method of Administration of Serum . Action of Serum .... Results of Sero-therapy Serum Reactions .... Endless Chain Treatment . Vaccine Treatment .... Drug Treatment .... Prophylaxis ..... 47 47 49 49 52 53 55 57 59 6o Recent Classification and Treatment of Pneumonia INTRODUCTION The pneumococcus is the causative agent in more cases of lobar pneumonia than any other micro- organism. By virtue of this fact the intensive study of the biology of the pneumococcus, which the or- ganism merits, has yielded encouraging results. Both prophylactic and therapeutic measures have in consequence been instituted which promise to con- trol so ravaging a disease. Pneumonia exacts each year a heavy toll in human lives so that the economic loss exceeds that of other diseases. The young and the old, the weak and the strong are numbered among its victims. In this connection it may not be amiss to refer to the mortality statistics compiled by the United States Census Bureau. In 1916, 98,334 deaths oc- curred from all forms of pneumonia in the registra- tion area of the United States. Of this number 54,699 were due to lobar pneumonia. The remain- der were due to either broncho-pneumonia or to an unspecified form of the disease. The death rate for all forms of pneumonia in 1916 was 137.3 per 100,000 population. This rate exceeds that of pulmonary tuberculosis by 13.5, the death rate of the latter being 8 RECENT CLASSIFICATION AND 123.8 for the same year. The following table (Table I) shows that more deaths have occurred in Rhode Island from pneumonia than from pulmonary tuber- culosis in the six-year period from 1911 to 1916 in- clusive : TABLE I Showing Death Rate in Rhode Island per 100,000 Population. Year From pneumonia (all forms) From tuberculosis of lungs I916 1590 142.9 I915 164.9 132.7 1914 I43-I 134.6 I913 146.8 133-4 I912 157-7 132-4 19II 152-1 I45-I In the city of Pittsburgh alone, 1394 cases of the disease were reported in 1916, of which 1296 proved fatal. The analysis by Shattuck and Lawrence of 3291 cases of lobar pneumonia, treated at the Massa- chusetts General Hospital from 1889 to 1917, has furnished noteworthy deductions. The authors have noted a marked variation in the yearly mortality rate. This reached as high as 40% in 1896, and as low as 16% in 1906. The one thousand consecutive cases of the disease at the same hospital, compiled by Coolidge and Townsend, during the period 1882 to 1889, have been considered in addition to the number cited. A grand total of over 4200 cases is thereby obtained. If of this total the mortality rate is con- sidered by decades, it becomes evident that it has gradually increased from 10% in the first decade to 28% in 1917, and that since 1881 there has been no TREATMENT OF PNEUMONIA 9 significant change in the death rate. Acute lobar pneumonia in virtue of its prevalence must therefore be regarded as an endemic disease. The wide dis- tribution of the pneumococcus and its survival in dust form two very important factors in its epi- demiology. The pneumococcus was first observed by Pasteur and Sternberg in the saliva of normal individuals about 1881. Its association with pneumonia was es- tablished by Fraenkel in 1884; and later Weichsel- baum confirmed the association by demonstrating its presence in the blood and in the organs of victims of the disease. After this relationship had been es- tablished, many attempts were made to reproduce the disease in animals. Lamar and Meltzer in 1909 suc- cessfully produced experimental pneumonia in dogs. Under ether anaesthesia, a small stomach tube was introduced through the larynx into a bronchus, and from 5 to 10c. c. of a bouillon culture of the pneu- mococcus were injected into the lung. The animals were killed at intervals varying from one to six days after the time of injection. At autopsy, consolidation involving one-half lobe to that of the entire lung was found. The pathology of the pulmonic solidification was identical with that of lobar pneumonia in human beings. The pneumococcus stains readily with the aniline dyes and is Gram positive. It occurs in pairs of small oval or lancet shaped organisms, the broader ends of which are in apposition. During rapid growth ba- cillary forms occur, whereas in older cultures the or- 10 RECENT CLASSIFICATION AND ganism loses its ability to stain by Gram's method and involution forms appear. It is a non-motile, non-flagellated form which does not produce spores. Its optimum temperature is 37°C., and growth takes place equally well under anaerobic as under aerobic conditions. The pneumococcus, moreover, when freshly iso- lated from human or animal tissues possesses a well marked capsule. This is readily brought out by appropriate staining methods. In fact, the excessive capsule development of some forms of the pneumococ- cus appears evident in the ordinary Gram's stain as a halo around the bacterial cell. The organism soon loses the capsule on cultivation on artificial media, but regains it on being reinoculated into ani- mals. It seems as if the capsule serves as a protec- tion to the microorganism. Curiously enough the more severe types of pneumonia are associated with those organisms which show the pronounced capsule formation. Undoubtedly, they acquire thereby a greater resistance against the immune forces of the invaded host. The simple laboratory media yield ordinarily a growth of the pneumococcus. Its growth, however, is much enhanced by the addition of serum, defibri- nated blood, or carbohydrate to the medium. In bouillon, the pneumococci grow diffusely and exhibit chain formation. In gelatine, no liquefaction takes place. On blood-agar plates, the appearance of the pneumococcus colony is definite and characteristic. The colony is finely granular, showing a darker TREATMENT OF PNEUMONIA 11 centre and a well defined border. It is, moreover, flat, transparent, and moist and presents a ringed or wrinkled appearance. A green zone of discoloration which results from the changes wrought on the hemo- globin of the red blood cells usually surrounds the colony. This zone of methemoglobin is accentuated by incubation of the blood plates at ice-box temper- ature. In media containing the simple sugars, an initial growth of the organism usuajly takes place. Its growth is soon retarded by the formation of acid, in consequence of which the organism dies out. The acid producing power is to be seen also in the coagu- lation of litmus milk and of inulin serum water. The ability to ferment inulin has been regarded as a dis- tinctive cultural feature of the pneumococcus. There are some strains which do not possess this power, or possess it only temporarily. Thus certain morpho- logical and cultural characteristics have come to be ascribed to the pneumococcus. These do not ab- solutely distinguish the pneumococci from the allied group of streptococci. Some strains of the latter may show capsule formation, produce acid in car- bohydrate media, ferment inulin, and show a zone of green pigmentation around the colony. Neuf eld's observation in 1900 that pneumococci are soluble in bile has proven of fundamental importance. The cholic acid is a ready solvent for the pneumococcus, but not for the streptococcus. The bile solubility test furnishes, therefore, means of absolute diagnosis between the pneumococci and other organisms of similar morphology and cultural characteristics. PART I TECHNIQUE The work of the next decade (1900-1910) was crowned with results which radically affected the older conceptions of pneumonia. A new era dawned. Hitherto attention had been focussed on the grosser relationships of organisms. The study of the dif- ferences between families of microorganisms now be- came secondary to the study of differences between members of the same family. It will be recalled that as early as 1881 pneumococci have been observed by Sternberg in the saliva of normal individuals. In 1909, Neufeld and Handel have first shown that im- munologic differences exist between pathogenic and non-pathogenic pneumococci. Accordingly, the re- cent biologic classification of the pneumococcus has been based upon differences in immunity response. Many races of pneumococci were isolated from cases of lobar pneumonia. Animals were then im- munized by repeated injections with gradually in- creasing numbers of the organisms. Thus, a rabbit, inoculated repeatedly with the same strain of the pneumococcus, would show a specific immunity re- sponse against that particular strain. Herein was the keystone of the new classification. It has been de- monstrated that the blood serum of an immunized TREATMENT OF PNEUMONIA 13 animal, horse or rabbit, as the case maybe, possesses the power to agglutinate in vitro that particular strain of the pneumococcus against which it had been immunized. Not only is there a specific agglutinin present in the immune serum, but also a substance capable of conferring passive immunity. The viru- lence of the pneumococcus for the white mouse is an interesting phase of its biology. Neufeld and Handel have demonstrated that the serum of a vaccinated animal protects a mouse against what would other- wise be a fatal dose of living pneumococci of the same strain. The specific reaction consists, therefore, in the de- velopment of an agglutinin and of a protecting sub- stance in the serum of the immunized animal, both of which are specific for that strain of the pneumococcus with which the animal has been inoculated. This was the predominant fact established by Neufeld's work. The entire subject was next amplified by Cole and his associates at the Rockefeller Institute. Large numbers of pneumococci have been isolated from various sources and animals immunized against the strains isolated. The immunity reactions of each strain of organisms against its homologous serum and the other immune sera were studied. On the basis of protection tests, devised by Neufeld, and of the agglutination reactions, Cole in 1912 divided the pneumococci into four groups. These were desig- nated by him as types I, II, III, and IV. Further investigation has brought to light the interesting fact that 80% of pneumococci isolated from clinical cases 14 RECENT CLASSIFICATION AND of pneumonia belong to types I, II, and III; whereas the remaining 20% of pneumonias are caused by type IV. The latter is the organism commonly found in normal mouths. The organisms belonging to the first three types represent a group of highly parasitic forms to which the term "fixed types" has been applied. This in contradistinction to type IV, which consists of many heterogeneous strains. It is noteworthy that the strain studied by Neufeld corresponds to type I of the Rockefeller classification. One year later, in 1913, Lister of South Africa classified pneumococci into twelve groups, naming them alphabetically. Phagocytosis and specific ag- glutinations formed the basis of his classification. Group A, numerically most important in South Af- rica, is unknown in this country; and Lister's groups B, C, and E are identical with types II, I, and III of Cole. From the foregoing observations the following summary is drawn: the serum of a rabbit im- munized against pneumococcus type I agglutinates organisms of type I in vitro and has no effect on types II, III, or IV; an antipneumococcus serum of type I protects a mouse against a fatal dose of type I organisms; whereas no protection is afforded in case types II, III, or IV organisms be used for in- jection. In other words, there is no cross-agglutina- tion and no cross-protection. The value of the classification of pneumococci based upon immunologic differences is only too ob- vious. Epidemiology, prophylaxis, and treatment TREATMENT OF PNEUMONIA IS of acute infections are determined by etiology. It is therefore of paramount importance to study the pneumonias from the newer standpoint. Much at- tention has been given to this phase of the study by laboratory workers all over the country. The pro- gress already made augurs well for the future. An antipneumococcus serum against type I has been produced which, if administered early in the disease, affects the outcome favorably. More detailed con- sideration is given to the serum treatment of pneu- monia in another part of this paper. Our attention will now be devoted to the methods employed for the differentiation of the pneumococci into their four types. Pure cultures of the organism must first be procured from the infected host. These are obtainable from the blood or sputum of the pa- tient or by a direct lung puncture. The sputum is usually the material at hand. Accordingly it is re- ceived in a sterile Petri dish. The sputum from the deeper air passages, freed from the bacterial flora of the mouth, yields better results. A small portion is washed and ground and injected into the peritoneal cavity of the white mouse. The mouse acts as a se- lective incubator. The pathogenic pneumococci mul- tiply rapidly, whereas the growth of the saprophytic organisms is inhibited. Sufficient growth of the for- mer is usually obtained in six to eight hours. Under sterile precautions, the peritoneal cavity is opened; a smear is made from the exudate and a loopful of it is streaked on one half of a blood-agar plate. The smear, stained by Gram's, shows the character of the 16 RECENT CLASSIFICATION AND organism which has been recovered by the inoculation method. As a rule, the pneumococci are present in large numbers. By means of a glass pipette, the peritoneal cavity is washed out with about 8 to 10c. c. of bouillon or salt solution. At this time the presence or absence of a sticky exudate between the intestinal loops is noted. The peritoneal washings containing pus cells and bacteria are received in a sterile centri- fuge tube. To remove the pus cells, the tube is cen- trifugalized at low speed for a few minutes. The supernatent fluid is pipetted off into another centri- fuge tube. This is revolved at high speed until the organisms are thrown down. The resulting clear supernatent fluid was discarded in the earlier period of this work. It is at present retained and tested for a specific soluble substance excreted by the bacterial bodies. The bacterial sediment at the bottom of the second centrifuge tube is next emulsified by the addition of a few cubic centimetres of salt solution. This con- stitutes a pure culture of the pneumococcus, obtained during the course of several hours by animal inocula- tion from the sputum of a case of pneumonia. The autopsy of the mouse is completed by opening the thoracic cavity and exposing the heart. As a sep- ticaemia is usually present, two or three tubes of bouillon (-|-0.3 to -j-O.S acid) are inoculated with the heart's blood. A loopful of the blood is finally smeared over the second half of the blood-agar plate, used previously for the peritoneal exudate. These are now incubated for subsequent examination. TREATMENT OF PNEUMONIA 17 Agglutination tests are made by mixing the saline suspension of the pneumococci with antisera i and ii in small sterile test tubes. The tubes are shaken gently and mixed by tapping with the finger. They are then placed in the water bath at 37°C. for two hours, and afterwards kept in the ice box over night. Soon after the mixing of the suspension of the or- ganisms, the tubes are closely observed for reaction. This consists in the clumping of the organisms so that at first small flakes appear, as in a "miniature snowstorm." On standing, the clumps settle to the bottom of the test tube. If the organism belongs to type II, the clumping reaction in the No. 2 will be seen in one or two minutes. The reaction of type I organisms to their homologous serum is slower, the clumps appearing within half an hour. The presence of clumps in either tube No. 1 or tube No. 2 deter- mines to which of the two types the organism belongs. In the absence of clumping, the diagnosis of type III is justified in case a sticky exudate has been found at autopsy and in case smears from the peritoneal cavity show the organism to possess a large capsule. The diagnosis of type IV is reached by a process of exclusion. The tubes are again observed and the reactions read after the ice-box incubation over night. On the next day, also, the bouillon cultures, made from the heart's blood, are examined. If they are pure cultures of the pneumococcus, tubes of litmus milk and inulin serum water are inoculated with the growth, The acid production in these media fur- 18 RECENT CLASSIFICATION AND nishes confirmatory evidence of the cultural charac- teristics of the pneumococcus. Agglutination and bile solubility tests are repeated with the bouillon growth. The foregoing technique makes possible the prompt isolation of the strain of pneumococcus from the sputa of pneumonia patients and the determination of the type of infecting organisms by agglutination tests. A consideration of the four types is essential for an ap- preciation of the problems involved in the epidemi- ology and therapeusis of the disease. The biologic differences between the four types are most striking. Pneumococcus type I, the first member under dis- cussion, has the distinguishing capsule of the pneu- mococcus group. The capsule, however, is not as well developed as it is in types II and III. This may account for the success which has attended the production of an antipneumococcic serum against the homologous organism. This serum, as will be in- dicated later, is of distinct value in the treatment of pneumonia caused by type I. Such a serum care- fully prepared will agglutinate type I organisms in a dilution of 1:400 or over, and 0.2c.c. will protect a mouse against 0.1c. c. of a virulent culture. Thestand- ard of virulence is determined by the least possible amount of a bouillon culture which will kill a white- mouse in less than 36 hours. The amount consists of O.OOOOOlc.c. The agglutinating and protecting powers of an antipneumococcus serum against type II organisms are less effective. Despite identical methods em- TREATMENT OF PNEUMONIA 19 ployed in immunizing animals with type II organ- isms, the results markedly differ. Thus an anti- pneumococcic serum type II agglutinates the homol- ogous organisms in dilutions no greater than 1 :200 and 0.2c.c. affords protection against only O.Olc.c. of a virulent culture. The antiserum I is, therefore, ten times more powerful than antiserum II. Experi- ments, moreover, have proven the inadequacy of type II antiserum in treatment. In the description of methods of "typing" it has been noted that agglutination takes place promptly upon the addition of anitserum II to tube No. 2 containing the organisms. As the number of tests increased, the observation has been made that some of the type II pneumococci become agglutinated slowly. These have been termed "atypical 11 's." Not only is their agglutinating response delayed, but their virulence is slightly less than that of the typical members of the group. In 1915, Avery published his classification of the atypical II organisms, which was based on a study of ten strains. The ten strains have been divided into sub-groups: I la, lib, and llx, respectively. In the four years following Avery's work, Stillman has applied similar methods of study to a larger series of atypical II organisms, 204 in number. These strains have all shown the delayed agglutination with antiserum II. The strains have been obtained from the most varied sources, and include 77 from cases of lobar pneumonia, 5 from post-operative pneu- monias, 100 from normal mouths, 10 from convales- 20 RECENT CLASSIFICATION AND cents, and 6 from dust. By means of specific ag- glutinins, Stillman in 1919 has classified the atypical II pneumococci into twelve sub-groups. Sub-group lib is the largest, comprising 39% of all strains studied. Sub-groups I la and Ilh are essentially the pathogenic forms. Of the 458 strains of pneumococci, isolated from cases of pneumonia at the Rockefeller Hospital in three years, 52 or 11% belong to atypical II forms. In the 59 cases of atypical II pneumonias, there were 18 deaths or 32%. In summary it may be stated that the twelve sub-groups of atypical 11 's have an incidence of 11% in lobar pneumonia and 18% in normal mouths, and that the mortality of the former cases is 32%, indicating their high pathogenicity. It will be recalled that the diagnosis of pneumo- coccus type III is made on morphologic grounds. The organism is larger, rounder, and less lanceolate than the members of the other groups. It possesses, moreover, a very large capsule, and the inflammatory exudate which it produces is of a mucoid character. Until 1915, the diagnosis of tpye III by agglutina- tion test was impossible, because no antiserum was available. In August of that year, Wardsworth and Kirkbride had succeeded in actively immunizing horses against pneumococcus type III. The task is most difiicult, it being necessary to remove the sticky capsule from the organism by treatment with weak acids. The contribution in this regard has been most noteworthy, for the diagnosis of type III can now be made on a serologic basis with every assurance TREATMENT OF PNEUMONIA 21 of positivity. Subsequent experience has shown that mucoid exudates are formed also by some members of types II and IV. In my series, one atypical II and one type IV showed a definite stringy exudate in the peritoneal cavity of the mouse. Undoubtedly such organisms form connecting links between the various members of the pneumococcus family. The converse of this also obtains, in that some type III organisms produce no mucoid exudate. During the influenza epidemic, I isolated one strain from a case of pneumonia which agglutinated promptly with antiserum III, and yet the peritoneal cavity of the mouse was devoid of the mucoid exudate. Type IV represents all the remaining members of the pneumococcus family. It is the group commonly encountered in normal mouths and is responsible for about 20% of all pneumococcic pneumonias. The group is characterized by much heterogeneity and has been most extensively studied by Olmstead. The study has embraced the immunologic relationships of 94 strains — 46 from normal mouths and 48 from varied sources. Thirty-seven immune monovalent sera have been prepared. At least twelve sub- groups have been identified, the members of which show cross-agglutination with each other but not with the other groups. Undoubtedly some members of the group are transition forms between type IIx and type IV. In 1917, Dochez and Avery have added another noteworthy contribution to the biology of the pneu- mococcus. The authors have observed that a soluble 22 RECENT CLASSIFICATION AND substance is excreted by the actively growing or- ganism into the surrounding culture medium. This substance is of a protein character and is precipitated by the specific antiserum. In cases of pneumonia, it is found in the circulating blood and is eliminated from the patient through the kidneys. Type III produces the largest amount of soluble substance, and types II and I come next in the order named. The determination of type may be made, therefore, by the examination of the urine. An equal amount of each of the pneumococcic anitsera (0.5c.c.) is added to each of three tubes containing the patient's urine. The diagnosis of type is determined by which of the antisera produces a coagulum. The production of the soluble substance takes place early in the course of the disease. Its presence has been demonstrated in the urine as early as twelve hours after the initial chill. Its excretion may be continued for many days during convalescence. In some cases the prolonged excretion by the kidneys has been associated with an unresolved pneumonia. Obviously the production of the soluble and pre- cipitable substance is directly proportional to the amount of infection. Moreover, it is also of grave prognostic significance, for the mortality is high in those cases of pneumonia which show its presence in the urine. The most virulent strains of the pneu- mococci are more apt to produce the soluble pre- cipitinogen. Quigley disagrees with the deductions drawn by Dochez and Avery. The former investigator has TREATMENT OF PNEUMONIA 23 studied the precipitin reactions in the urine of 100 cases of lobar pneumonia. The incidence of type is as follows: Type I, 33 cases; type II, 36 cases; type III, 13 cases; and type IV, 18 cases. His findings are of timely interest. Of the 82 cases due to the fixed types, the urine of 67 (81%) have given at some time during the disease the specific precipitin reaction. Of the 67 cases, 23 or 34% have died. The strength of the reaction has gradually in- creased for a period of 3 to 4 days, persisted from 2 to 19 days, and gradually waned. Of the 27 deaths due to the fixed types, 23 have shown the reaction in the urine. Four have been negative. It thus appears that the precipitinogen is present in the urine of a large percentage of fatal cases of pneu- monia ; but the mortality of those patients who have not shown the precipitinogen in the urine has been only slightly less than in those who show its presence. The excretion of a soluble substance by the pneu- mococcus during its period of growth has proven of the greatest value in diagnostic technique. This sub- stance is present wherever the organisms find favor- able conditions for growth, — in the lung, in the pleural, synovial, or spinal fluids, or in the peritoneal cavity of the mouse. It will be recalled that the pneumococcus multiplies rapidly in the mouse after the injection of washed sputum; and that for diag- nosis of type, the cavity is washed out with a saline solution. After the removal of the pus cells by slow centrifugalization, the supernatent contains only the organisms. Upon high centrifugalization, the or- 24 RECENT CLASSIFICATION AND ganisms are thrown down, leaving a clear, supernatent fluid. This fluid has formerly been discarded. In the present technique, the fluid is retained and tested for the presence of the soluble specific pre- cipitinogen. Equal amounts of the three specific antisera are added to each of three tubes containing the clear, supernatent liquid. A precipitate forms promptly in the tube upon the addition of its homol- ogous antiserum. This test is done at the same time that the agglutination reactions are performed with the emulsion made from the sediment of pneu- mococci. The precipitin test acts, therefore, as a control for the agglutination test. It has, moreover, certain advantages in that the precipitin test demands no incubation and is not interfered with by the pres- ence of other organisms. It is specific and shows no cross-reactions. If a mouse, inoculated with sputum for diagnosis, dies during the early hours of the morn- ing, considerable autolysis may take place before it comes to autopsy. The pneumococci become con- taminated with saprophytic invaders, making clear- cut agglutination tests impossible. On such occa- sions, the precipitin reaction ofl^ers means for prompt diagnosis of type. As previously stated, the presence of contaminating bacteria does not interfere with the precipitinogen. Its presence can be detected after all the organisms have been removed by centrifugal- ization. Three opportunities have come to me for making rapid diagnosis of types in spinal fluids from cases of pneumococcic meningitis. The fluids upon with- TREATMENT OF PNEUMONIA 25 drawal were turbid. Stained smears showed the presence of pus cells and Gram positive diplococci. The fluids were centrifugalized long enough to yield a clear supernatent portion and a sediment of pus cells and organisms. With the former portion, the precipitin test was performed by adding the specific antisera. The reaction was obtained almost in- stantaneously. One spinal fluid showed a type III, and the two others gave a type I reaction. The diag- nosis of type was confirmed by inoculating mice with the sediment of organisms which were thrown down in the centrifugalizing tubes. The presence of the soluble precipitinogen in the spinal fluids of pneu- mococcal meningitis has not been entirely unexpected. The spinal fluid is an ideal medium to receive the soluble products which result from the metabolism of the invading pneumococci. The conditions are analogous to those which obtain in the peritoneal cavity of the mouse or in a test tube of bouillon. In fact, one case of meningitis comes to mind in which the spinal fluid has served as liquid culture medium for the pneumococcus. So overwhelming was the in- fection that no pus cells, the first line of defense, had been called forth. I have seen other cases of menin- gitis in which most diligent search failed to reveal the presence of bacteria in the purulent sediment. Pneumococci, however, were obtained in cultures from the same sediment. It is, doubtless, apparent that in lobar pneumonia the rapidly multiplying pneumococci secrete a soluble substance into the lung alveoli; that this substance 26 RECENT CLASSIFICATION AND is taken up by the capillaries and lymph vessels and is circulated through the blood stream; and that eventually the soluble precipitinogen is excreted by the kidneys and its presence is detected by the ex- amination of the urine. The question which now suggests itself is whether or not the soluble precipi- tinogen is found in the spinal fluids of those cases of pneumonia which are uncomplicated by any in- flammatory meningitis. Not infrequently in syphi- lis, the syphilitic antibody is present both in the blood and in the cerebro-spinal fluid, as determined by the Wassermann reaction. It may be present in the spinal fluids of those cases in which there are no symptoms referable to involvement of the cerebro- spinal axis. The mere absence of symptoms, how- ever, does not exclude the possibility of a pathological lesion of the nervous system. The conclusion may be warranted that in some cases of syphilis there is a passive transfer of the antibody from the blood stream into the cerebro-spinal fluid. In the light of the above facts, the question may be asked, "Does a similar transfer of soluble precipitinogen take place from the blood to the spinal fluid in the course of a pneumonia?" In my experience, I have encountered one case of pneumonia due to a type I infection which in part answers the question. The disease ran a very severe course. A purulent arthritis of the knee was an early complication, from which the type I pneumococcus was isolated. Several hours prior to death, symp- toms of meningitis presented themselves. Lumbar TREATMENT OF PNEUMONIA 27 puncture yielded a large amount of a clear fluid. The globulin test was negative, and there was no excess of cellular elements. It was obviously a case of meningismus. On testing the undiluted fluid with each of the three antipneumococcic sera, no pre- cipitin reaction was obtained. It follows, therefore, that in this case at least there was no passive trans- fer from the blood to the spinal fluid of the soluble precipitinogen. The severe course of the pneumonia with its septicaemic manifestations justifies the be- lief that the precipitinogen must have been present in the blood stream, and yet none was detected in the spinal fluid. The problem now assumes a physiologi- cal aspect. Despite the excessive production of spinal fluid, was the choroid plexus impermeable to the precipitinogen or was the capillary pressure too high to admit of the passage of the precipitinogen from the blood stream into the cerebro-spinal fluid? Whatever factors may be involved, the spinal fluids of pneumonia cases are deserving of study. The con- centration of the precipitinogen in the blood stream, the pressure of the cerebral capillaries, the presence or absence of pneumococcal precipitinogen in the spinal fluid, the permeability or impermeability of the choroid plexus may furnish evidence concerning the physiological mechanism of so formidable a disease. Any description of technique for the determination of type is incomplete without mention of Blake's modification. Blake, in 1917, has devised a means for overcoming a difliculty which has been encoun- tered in the agglutination tests. Some forms of pneu- 28 RECENT CLASSIFICATION AND mococci represent transitions between the various types, and accordingly show agglutination in all the antipneumococcic sera. The character of the ag- glutination is atypical. It takes place more slowly and is less complete than in the agglutination re- actions seen with the "fixed types." To render cross-agglutination impossible, Blake has established an optimum dilution of the antisera and an optimum incubation time. The use of the antisera undiluted, as has been the practice hitherto, has been dispensed with. Antiserum I is now used in the dilution of 1:20; antiserum II is used both un- diluted and diluted 1:20; and antiserum III in the dilution of 1:5. Equal parts (O.Sc.c.) of the suspen- sion of organisms are added to the dilutions of the antisera, and the tubes are incubated in the water bath at 37°C. for one hour. At the end of that time, the agglutination of the organisms is complete in the homologous antiserum. The following table (Table II) indicates the agglutination reactions of the different types of pneumococci, resulting from the use of diluted antisera, as suggested by Blake: TABLE II Showing Blake's Optimum Dilution Method Antiserum I Antiserum II Antiserum III i:20 undUuted 1 :20 1:5 Type I ++ — — — Type II — ++ +-I- — Type II (a, b, x) — + — — Type III — — — +-I- Type IV — — — — TREATMENT OF PNEUMONIA 29 It is evident that the typical II forms can be readily differentiated from the atypical. The former clump in the diluted and undiluted homologous an- tiserum. The latter show, at the end of one hour, partial to complete clumping only in undiluted serum. The use of the diluted antisera, as seen in the above table, makes cross-agglutination impossible with transition forms, and makes diagnosis of type more exact. The determination of type of pneumococcus in cases of lobar pneumonia is of practical importance. The efficacy of antipneumococcus serum I in therapy depends upon its early administration. In any acute self-limited infection, treatment must be begun as soon as possible to be of any avail. Each hour counts. The minimal time for diagnosis of type by mouse inoculation varies from 8 to 12 hours. Methods have, therefore, been introduced whereby the diag- nosis of type may be made more promptly. Mitchell and Muns, in 1917, and Krumurede and Valentine in 1918 have proposed shorter methods. These de- pend upon the presence in the sputum of the soluble precipitinogen, which is formed in whatever medium pneumococci find favorable conditions for growth. Avery, in 1918, has devised a rapid cultural method for the pneumococcus. It emphasizes some fundamental characteristics of the pneumococcus — that the addition of carbohydrate and blood protein to culture medium encourages an initial growth of the organism; that bile acts as a ready solvent for the pneumococcus; and that a soluble precipitinogen 30 RECENT CLASSIFICATION AND TREATMENT is formed during the period of vigorous growth. Avery has controlled the diagnosis of type by the above rapid cultural method with the mouse inocu- lation method, and in 60 cases the results have been identical. PART II STATISTICS — EPIDEMIOLOGY As a sequence to the recognition of types amongst the pneumococci, their relative incidence in pneu- monia becomes an interesting subject for investiga- tion. The experiences at the Rockefeller Hospital furnish the earlier estimates. Thus, in the year 1912- 1913, 74 cases of pneumonia have been studied. The incidence of type has been as follows: Type I 35 cases or 47% Type II 13 " or 18% Type III 10 " or 13% Type IV 16 " or 22% In the year 1913-1914, 71 cases of pneumonia have been treated at the institution. These have been classified as follows: Type I 21 cases or 30% Type II 28 "or 39%^ Type III 6 " or 8%, Type IV 16 " or 23% It appears, therefore, that in the first series 58 cases or 78% have been caused by the "fixed types" (I to III inclusive), whereas 16 cases or 22% have been due to type IV. This percentage closely corresponds with that of the second series, 77% due to the "fixed 32 RECENT CLASSIFICATION AND types" and 23% to type IV. The constancy of the percentage findings is most noteworthy. Three years later, Cole, in reviewing his greatly enlarged experience, concludes that 60 to 70% of cases of lobar pneu- monia are caused by types I and II; 10 to 15% by type III; and 20 to 25% by type IV. To control Cole's observations the following table has been compiled, which contains the results of different in- vestigators in different parts of the country: TABLE III Showing Type Incidence Investigators Location I II II a, b, III X. IV Total Alexander & Boston, Christian Mass. 25 22 12 45 104 Author [R. L] 24 22 15 89 150 Beals & Others Mich. 27 25 49 lOI Blanton & Irons Mich. II 16 16 148 191 Cole Texas 17 3 4 I 6 31 Cole & Others Rockefeller Hospital 151 133 19 59 92 454 Dick Arkansas 6 10 I 12 29 Hartman & Lacy Pittsburgh 41 29 6 23 99 Kirkbride N. Y. State 102 39 13 56 119 329 Litchfield Western Penn. 55 39 7 33 134 McLelland Alexandria, La. • 13 4 I 24 42 Medalia & Schiff Waco, Tex. 6 13 19 38 Mitchell & Syracuse, Muns N. Y. 19 4 3 4 29 59 Quigley Chicago 33 36 13 18 100 Thomas Camp Mead, Md. 57 26 II 27 121 Total 587 421 50 191 733 1982 TREATMENT OF PNEUMONIA 33 The incidence of types based on the above study of 1982 cases of pneumonia in widely scattered sec- tions is as follows : Type I 29.6+% Type II 21.2+% Type II (atypical) 2.5+% Type III 9.6+% Type IV 36.9+% The occurrence of the "fixed types" is less than that of Cole's earlier estimate, notably so in types II and III. The most striking discrepancy is seen in type IV, the percentage of which equals 36.9 as com- pared with the 20 to 25 percentage incidence noted by Cole. In two localities, the incidence of type IV reached as high as 45% and 50.8%, in Camp Pike, Arkansas, and Camp Upton, New York, respectively. Pneumococcus type IV, it will be remembered, is the organism found normally in the upper respiratory passages. This observation is not without an im- portant bearing on epidemiology, as will be dis- cussed later. The mortality rate of the four types is of equal interest with their incidence. The following table shows their percentage mortality, as determined by the several investigators : 34 RECENT CLASSIFICATION AND TABLE IV Showing Mortality-rate of the Four Types of Pneumococci No. of Investigators cases I II II a, b, X. III IV Alexander & Christian 104 32.0% 40-9% 50.0% 24.4% Hartman & Lacy 99 38.0% 34-0% 67.0% 26.0% Litchfield 134 47-0% 41.0% 57-0% 48.0% McLelland 42 7.7% 25-0% 4-2% Medalia & SchifT 38 7.0% 5.0% Quigley 100 39-0% 36.0% 8.0% 170% Thomas 1 2 1 5-3% 19-2% 18.: 2% 22.2% Total 638 Average 28.1% 29.0% 18.: 2% 45-5% 20.9% Of the 638 cases of pneumonia, therefore, the mor- tality rate has been 28.1% in type I cases; 29.2% in type II; 18.2% in type II atypical; 45.5% in type III; and 20.9% in type IV cases. The above percentages of deaths due to the fixed types are in accord with the earlier findings of Cole. He has estimated that the mortality of types I and II equals 25 to 30%; of type III, 50% or more; and of type IV, 10 to 15%. In the two compilations, the highest death rate is maintained by type III infections. Type IV infections, as a result of this study, do not warrant the favorable prognosis which has been ac- corded them. Their percentage mortality reaches 20.9 as compared with 15%, given by Cole. Many observers have called attention to the fact that the mortality of the different types of pneumonia varies in different years, more notably in type IV, and less so in the fixed types. Bovaird mentions a mortality of 5% in 1914 in pneumonias due to type IV, and a mortality of 25% in 1916 caused TREATMENT OF PNEUMONIA 35 by the same type. Litchfield cites a case of pneu- mococcic meningitis, compHcating a pneumonia due to type IV. My series includes fourteen cases of lobar pneumonia following measles. Pneumococcus type IV has been isolated from each case. The disease ran a very severe course and was accom- panied with a high mortality. The opinion originally maintained concerning the favorable prognostic sig- nificance of type IV in pneumonia needs be modified. The type of pneumococci occurring in the pneu- monias of infants and children merit investigation, for the disease is less fatal in the young than in adults. Pisek and Pease have, in 1916, published the following results of their study of 48 cases of pneumonia in children under six years of age: TABLE V Types of Pneumococci occurring in Children (From Pisek and Pease) Cases of Lobar Pneumonia Cases of Broncho- Pneumonia Percentage Pn, I 9 2 22.9 Pn. II II 3 293 Pn. Ill I 3 8.3 Pn. IV 7 12 39-8 Dunn has also found type IV to be the predomina- ting organism. Mitchell has classified the pneu- mococci in 90 cases of pneumonia which have oc- curred during the winters 1914-1915 and 1915-1916. The average age of the 90 children has been 2 years 36 RECENT CLASSIFICATION AND and 3 months. Six of the children have been over 6 years of age, and 62 under 2 years. The frequency of types has been as follows: Type I 11.1% Type II 11.1% Type III ?>.?>% Type IV 74.4% I have isolated pneumococcus type I from a child of 5 years, and type IV from 2 children aged 2, from one aged 6, and from one aged 8. Compared with the percentage incidence of 80 of the "fixed types" in adults, that of children reaches only 25.5% as shown by Mitchell. The mortality percentage of the types in Mitchell's series has been: Type I no deaths Type II 20% Type III ^Z.?>% Type IV 10.9% The above findings warrant the conclusion that pneu- mococcus type IV occurs more frequently in the pneumonias of infants and children than of adults, and that the mortality of the fixed types is lower in infants than it is in adults. From the point of view of epidemiology, there is no disease which is better understood than diphtheria. Four important factors are recognized as operative in its spread: first, the presence of the diphtheria ba- cillus in normal individuals; second, the presence of the organism in those exposed to patients or to con- TREATMENT OF PNEUMONIA 37 valescents from the disease; third, the "carrier" stage of convalescents; and fourth, the transmission of the bacillus by food substance, as milk, ice cream, etc. Are the same factors operative in pneumonia as in diphtheria? In answer to this important question, it becomes necessary first to consider the incidence of the four types of pneumococci in healthy persons. Stillman, in 1916, has presented the results of his study of the types of pneumococci present in the sputum of normal individuals. The mouse inocula- tion method, as previously described, was employed. Of 398 normal persons, pneumococci have been ab- sent in 226, and found present in the sputa of 172. Of the latter number, the type distribution has been as follows. See Table VI. The results of the study of another series of healthy persons, appearing one year later, in 1917, have proved confirmatory. Of 297 individuals pneumo- cocci have been isolated from 116, with the following grouping : TABLES VI AND VII Incidence of Pneumococci in Normal Persons (From Stillman) Incidence Percentage Incidence Percentage Type I 4 2.16 I 0.8 Type II 4 2.l6 0.0 Type I la o o.o I 0.8 Type lib II 5-84 7 5.8 Type IIx 15 7-79 14 11.6 Type III 44 23.40 34 28.1 Type IV no 58.51 64 52.9 38 RECENT CLASSIFICATION AND The high incidence of type III and type IV in normal persons is most striking as compared with the low incidence of types I and II. Type III is more com- mon in the mouth than has formerly been supposed. The incidence of types III and IV in disease is represented by a percentage of 10 and 36 respectively. The conclusion, therefore, becomes justifiable that pneumonias due to types III and IV are autogenic in character. The organisms live in the upper res- piratory passages as saprophytes until the balance of power between the host and potential invaders is disturbed. The cause of the disturbance may be attributed to a lowered resistance. This explanation, however, is unsatisfactory. Measures may be de- vised for determining the immunity or susceptibility of the individual to pneumococcic infection by some simple method, as the Schick cutaneous test in diph- theria. Certainly investigation along these lines may be productive of valuable information. The study of the bacteriology of post-operative pneumonias and of common colds lends support to the autogenic theory of type III and type IV infections. Olmstead in a study of 130 cases of post-operative pneumonia has isolated pneumococci in 28 instances. The types occurring amongst them are as follows: Incidence Percentage Type I 2 cases or 7.1% Type II 1 case or 3.5% Type II (atypical) 2 cases or 7.1% Type III 5 " or 17.8% Type IV 18 " or 64.2% TREATMENT OF PNEUMONIA 39 Similar observation has been made in forty-three cases of "common colds" with the following result; Incidence Percentage Type I 2 cases or 4.6 Type II 2 " or 4.6 Type III 4 " or 9.3 Type IV 35 " or 81.3 The percentage of types III and IV continues as consistently high in autogenic infections as under normal conditions of health. The striking fact remains that the incidence of both types I and II in normal individuals is low (0.8 to 11.6%), whereas in disease the two types comprise 50 to 55% of all cases of lobar pneumonia. Some ex- planation other than autogenic infection must be made. The alternative is infection from without. Therefore the transmissibility of the fixed types of pneumococci from patients infected to those in con- tact with them needs be ascertained. Information in this regard has first been obtained from the in- vestigations of Dochez and Avery in 1915. Thirteen associates of thirty patients suffering from pneu- monia, have proved to be carriers of the identical types of pneumococci which were responsible for the infection. Thus, eight cases of pneumonia due to type I gave rise to three positive contacts, each one of the latter having in the sputum a type I pneumo- coccus. The same was true of 24 cases of pneu- monia due to type II. They gave rise to ten car- riers of type II amongst their contacts. Stillman 40 RECENT CLASSIFICATION AND has studied the contacts of 48 pneumonia cases during the winter 1915-1916, with the following results: Type I — 24 cases ; contacts 44, of whom two proved positive, 8.3% Type II — 17 cases; contacts 40, of whom three proved positive, 17.6% Accordingly, 5 of the 84 contacts or 5.9% became carriers of the fixed types of pneumococci. In 1917, Stillman repeated the study. He found that of 107 normal persons exposed to cases of type I pneumonia, 15% showed pneumococcus type I; of 77 exposed to type II cases, 5 or 6% showed type II. Accordingly 11% of persons coming in contact with cases of lobar pneumonia harbored in the saliva the fixed types of pneumococci; whereas of 297 in- dividuals who were not exposed only 0.8% were car- riers. Furthermore, 15 cases of lobar pneumonia out of a total of 52 due to types I and II were responsible for one or more carriers amongst their associates. Examination of the 184 persons who lived in the 52 households of the pneumonia patients, showed a car- rier state of types I and II in 21 individuals or 11% of the total number. The foregoing experiences emphasize the import- ance of "contact" as a factor in the spread of pneu- monia. This is well illustrated by a series of six TREATMENT OF PNEUMONIA 41 cases of lobar pneumonia in one household, which has been recorded by Berry and Chickering. There were six persons, all related, occupying the same dwelling. On May 13, 1916, Mrs. D., age 34, de- veloped pneumonia and died 11 days later, on May 24. The other members in the following sequence developed the disease: May 14, Gertrude D., age 4; May 18, Mabel D., age 14; May 19, Arthur D., age 6; May 19, Fred D., age 20 months; and on May 22, Floyd D., age 11 years. Pneumococcus type I was isolated from five of the patients. Un- questionably, this represents infection with the same type of organism, spread by contact from one member of the family to another. I have studied the mouth flora of twelve contacts with four cases of pneumonia caused by type I. None of the contacts proved positive. My series, however, is too small to permit of any deductions. In support of the newer view of infection by contact, the incidence of pneumococcus types in rural dis- tricts furnishes interesting evidence. Richardson has analyzed the types of 46 cases of lobar pneumonia in country communities and has compared his find- ings with the type-incidence of the disease in the city. The following table shows the comparison: No. of cases Percentage Average percentage of Penn. Hospital cases Type I 7 15.2 28.6 Type II 5 10.8 21,2 Type III 2 4.3 8.3 Type IV 32 69.6 41.8 42 RECENT CLASSIFICATION AND The chance for infection by contact is obviously less in sparse!}^ settled communities than in the centres of population. The high percentage of type IV infection, which is autogenic in character, is in accord with expectations. Four of the type I cases and two of the type II cases give a history of a visit to the city prior to the onset of the disease. This suggests the possibility of having acquired the fixed types of pneumococci where opportunities for con- tact are greater. Of equal importance with the presence of the fixed types in "contacts" with pneumonia patients is their presence in convalescents from the disease. The persistence of diphtheria bacilli in the throat after all clinical symptoms have disappeared is a con- dition which may present itself with any patient. The same holds true of the typhoid bacillus, which may be discharged for months in the stools and urine of convalescents. The work of Dochez, Avery, and Stillman has shown that "carriers" may also be found in convalescents from pneumonia caused by the fixed types of pneumococci. In a study of 53 cases, it has been determined that the shortest carrying period is seven days, and the longest 85 days; and that the average carrying period for type I is 25 days and for type II it is 43 days. My observations in this regard have been as follows : TREATMENT OF PNEUMONIA 43 Case No. Type of Pneumococcus Number of days after onset of disease Type of Pneu- mococcus during convalescence I II 21 no pneumococci 2 I 3 22 3 I I8 IV 4 I 27 II 5 I 2 5 IV 6 III 4 8 23 III III III 7 I 3 7 8 I 3 14 25 9 I 35 54 72 102 251 IV no pneumococci lO I 26 I from nasal swabs IV from saliva Cases No. 5 and No. 9 represent the shortest and the longest carrying periods of 5 and 102 days re- spectively. The persistence of type III in case No. 6 for 23 days after the onset is consistent with the autogenic character of type III pneumonia. Case No. 10 is of unusual interest. On the 26th day after the onset of the disease, when convalescence was well advanced, swabs were rubbed over the nasal mucosa of each side and then soaked in bouillon. Some of the bouillon was injected intraperitoneally in the mouse. Type I pneumococcus was isolated from each swab. At the same time the sputum was injected and a type IV pneumococcus was recovered. 44 RECENT CLASSIFICATION AND On other occasions I have recovered from the nasal mucosa a fixed type of pneumococcus similar to the one found in the sputum. It is conceivable that more positive findings may be obtained by the use of a West tube in procuring organisms from the naso- pharyngeal mucosa. This method may furnish a more critical index of the bacterial flora than is pro- cured from the saliva of healthy persons or of con- valescents. Thus far, it has been indicated that "fixed types" pneumococci are transmitted from person to per- son by direct contact. "Is there any intermediary agent?" is a question which must now engage the attention. Stillman in 1917 has shown that dust may carry pathogenic pneumococci. Hitherto bac- teriological investigations of dust have been re- stricted to cultural methods. Stillman has made ob- servations of far-reaching importance by utilizing animal-inoculation methods in this study. Dust of rooms has been collected in sterile bouillon and in- jected into mice. In this manner the dust of 60 rooms has been examined and in 18 of them or 29% pneumococci have been recovered. The types of pneumococci correspond to those found normally in the mouth. In the dust of one room, however, a type I had been encountered. Curiously enough, a carrier of pneumococcus type I had visited in that room. The dust of 183 households, in which cases of pneumonia due to types I and II have occurred, has similarly been examined. Pneumococci have been found present in 74 or 40% of that number. TREATMENT OF PNEUMONIA 45 Type I has been isolated from the dust of 25 rooms, and type II from 23 rooms. Rooms in the house- holds other than those which the patients have oc- cupied have also had pneumococci of the fixed types in their dust. There are two epidemics on record which em- phasize the importance of dust as a factor. Of six cases of pneumonia in a Boys' Asylum, 3 or 50% have been caused by type I. The same type of or- ganism has been isolated not only from 10% of healthy contacts (56 in number), but also from the dust of the dormitory. The second outbreak oc- curred in a ward at the Rochester State Hospital for the Insane. Of six cases of lobar pneumonia, four were caused by type I. Two per cent of healthy con- tacts (148 in number) and the dust collected from the ward, have also yielded pneumococcus type I. Pneumococci may even remain viable in the sputum from patients for several days. The necessity of dis- infecting the sputum from active cases of pneumo- nia is emphasized by the following experience: On a handkerchief, I received a specimen of sputum from a pneumonia patient, who was infected with a type I organism. I placed the handkerchief behind a picture which was hung on the wall in the patient's room. Six days later, I excised a portion on which the sputum had dried and soaked it in bouillon. I then inoculated a mouse with some of the bouillon. Type I pneumococcus was recovered. On the eighth day I repeated the observation, and again isolated the type I organism. It is noteworthy that the 46 RECENT CLASSIFICATION AND TREATMENT second mouse died five days after the injection. Ap- parently the organism, though viable, had become less virulent. No further tests were possible. This single experiment, however, points out the danger which may result from the improper disposal of the sputum from active cases, as well as from careless spitting. Pneumococci, despite the effects of drying, retain their viability not only in sputum, but also in the dust of households. The deductions based on the foregoing considera- tions of the factors concerned in the epidemiology of lobar pneumonia are: Pneumonias due to types III and IV are auto- genic, for both types are found in the mouth flora of normal persons. Infections with type I and II may result from direct contact with an active case, or with a convalescent, or with a healthy carrier of those types; and from indirect contact, as an air-borne infection from dust or dried sputum. PART III TREATMENT Coincident with the increase in knowledge of the etiology of pneumonia has come an advance in the therapeusis of the disease. Serum-therapy has ac- cordingly been suggested by the earlier clinical and laboratory experiences as a means for combatting so acute an infection. Passive immunity, conferred upon the host as a protection against the parasite, has proven effective in diphtheria, tetanus, menin- gococcus meningitis, and gas-bacillus infection. Sim- ilarly, an antiserum has been devised for therapeutic use against pneumococcal infections caused by type I. The antisera produced against types II and III have thus far been of no avail in treatment. Further investigations will undoubtedly overcome the subtle causes for failure with these two antisera. The therapeutic serum is obtained from the horse after a course of active immunization. The first in- jections consist of washed culture sediment of fresh 18-hour cultures of type I organisms which are killed by heat. Three such injections are adminis- tered on three successive days; these are followed by increasing doses of living organisms. After an interval of one week, the series of injections is re- peated. Six days after the second injection, the animal is bled. Anderson has modified this technique 48 RECENT CLASSIFICATION AND by administering six daily injections in three courses with seven-day intervals. The potency of the serum thus obtained is de- termined by protection tests on mice. A serum, to warrant its use in therapy, must be of such potency that 0.1c. c. will protect a mouse against 0.4c.c. of cul- ture, which would be fatal in 0. 000001c. c. A serum of low potency is disappointing in its results. This is illustrated by Spooner and Sellards. Before the in- fluenza epidemic, they treated with an antiserum of low titre 24 cases of pneumonia caused by type I. Of the 24 patients, 19 recovered and 5 died, yielding a mortality of 20%. During the epidemic, 54 cases were treated with the same serum. In this number there were 31 recoveries and 23 deaths, a mortality of 43%. During the height of the epidemic, 15 cases were treated with a serum of standard titre. The results proved most satisfactory. Of the 15 treated cases, only one case died, giving a mortality of 7%. The fatal case was moribund at time of treatment. Curiously enough, agglutination tests applied to antipneumococcic serum furnish inconclusive evi- dence of its therapeutic value. Ordinarily a serum of high agglutinin titre has little protective power; and conversely, as emphasized by McCoy, a serum of high protective power may have no agglutinins. For effective therapeutic use in pneumonia caused by type I, not only is it essential that the serum be of standard potency, but it must also be administered as early in the disease as is possible. Herein lies the TREATMENT OF PNEUMONIA 49 importance of prompt diagnosis of type. The meth- ods of diagnosis by examination of the sputum and urine of the patient have already been discussed in detail. The antiserum, 75 to 100c. c. in amount, is administered intravenously. This is repeated every six to eight hours until a favorable result is obtained. The average dosage consists of 250c. c. Within an hour or two after the administration of the serum, a slight elevation of temperature usually precedes a marked fall. A second rise in temperature is an in- dication for more serum. Temperature records are to be made every two hours. Patients who are treated on the second or third day of their disease rarely require more than three doses of serum. In treated cases, crises are more frequent on the second, third, or fourth day. The febrile period is shortened. The antiserum, moreover, acts as an agent in sterilizing the patient's blood and in pre- venting the development of septicaemic complica- tions. In the severer infections, pneumococci varying in number from 900 to 2 per cubic centimetre may be present in the blood stream. The antiserum also produces agglutinins in the blood of the patient. Soon after the administration of the antiserum, the patient's blood may agglutinate the pneumococci in the dilution of 1:10. With the rise in the agglutinin index, there is a fall in temperature. The presence, however, of agglutinins in the patient is no criterion as to the effectiveness of serum-therapy or as to prog- nosis. As has been shown by Bloomfield, fatal cases of pneumonia have had agglutinins of high titre 50 RECENT CLASSIFICATION AND in the blood ; whereas cases with a favorable outcome have had a transient agglutinin power or none at all. One of my cases confirms the experience of Bloom- field. No agglutinins were demonstrable in the blood after the serum treatment, yet the patient made a prompt recovery from the disease. It would be of theoretical interest, at least, to determine to what extent the protective bodies of the antiserum are absorbed by the patient's blood. A better guide of efficacy in treatment may thereby be obtained. To what factors may, then, the curative value of the antipneumococcic serum be attributed? The work of Stryker, to which reference has been made, shows the effects of specific immune serum upon pneu- mococci of homologous type in vitro. It will be re- called that pneumococci grown in media to which immune serum has been added, lose their capsule and virulence. Their growth is unaffected, for the im- mune serum has no bacteriocidal action in vitro. Immune serum administered intravenously, how- ever, is possessed of bacteriocidal properties. Sterili- zation of the blood in pneumonia patients usually re- sults from this procedure. Explanation of the mech- anism of sterilization is gained by experimental evidence. A rabbit receives in one ear vein an in- jection of pneumococci, and one minute later re- ceives a dose of antipneumococcic serum in the vein of the opposite ear. The effect first noted is the im- mediate clumping of all the organisms in the blood and the removal of the clumps by the spleen, liver, and lungs. Active phagocytosis now ensues. The TREATMENT OF PNEUMONIA 51 polymorphonuclear leucocytes are chiefly engaged, and their activity is directed towards the clumps. A single leucocyte may contain 50 to 100 diplococci; and as the result of the digestion of the bacteria by the blood cells within the organs of the body, steri- lization of the blood is accomplished. It becomes necessary, therefore, to administer in the course of treatment sufficient antiserum to ag- glutinate and sensitize the bacteria and to render them harmless. Another requirement for successful treatment is the use of sufficiently large amounts of serum to neutralize in addition the soluble anti- genic substances formed by the pneumococci. These substances circulating in the blood stream have the power to fix the antibodies of the immune serum. In consequence, there is less antiserum available for the sensitization and agglutination of the bacteria. The earlier in the disease the antipneumococcic serum is administered, the more effective is the con- trol of the infection made possible. In the very severe infections or late in the infection, the soluble antigenic substances may be present to such an extent as to require impracticably large amounts of serum for their neutralization. Moreover, these sub- stances, resulting from bacterial metabolism, need to be saturated before there can be any effective concentration of antibodies in the blood stream. The presence of the soluble antigenic substances and their effects on immune sera can readily be demon- strated. Some pleural fluid from a case of empyema due to fixed type of pneumococcus is centrifuged. 52 RECENT CLASSIFICATION AND The organisms are removed and the sterility of the supernatent fluid is determined by cultural methods. This bacterial free fluid, added to anti- pneumococcic serum, not only lowers the agglutinin index of the latter against its homologous organism, but also diminishes its protective power, as de- termined by mouse inoculation. There must conse- quently be present some substances in the exudate which are capable of abstracting or neutralizing some of the immune bodies of the antiserum. These specific substances are also present in the blood in cases of severe septicaemia. Their presence may ac- count for the failure of antipneumococcic serum in the treatment of empyema by direct injection into the pleural cavity; and even in the treatment of pneumococcic meningitis. Litchfield reports twelve cases of meningitis complicating pneumonia due to type I. Despite the intraspinous administration of the serum, there were no recoveries. Consideration must now be given to the value of antipneumococcic serum type I in the treatment of pneumonia. Final judgment will of necessity be formed after the lapse of time and will be based upon the experiences of many trustworthy observers in dif- ferent parts of the country. Nevertheless, the re- sults thus far achieved in the saving of life give promise of a most favorable verdict. In the follow- ing table are summarized the results of sero-therapy in pneumonia caused by type I, as collected from the literature: TREATMENT OF PNEUMONIA 53 TABLE VIII Showing Results of Sero-therapy in Pneumonia Investigator < Number of eases treated Number of recoveries Number of deaths Number of moribund cases when treatment was begun Bloomfield II 8 3 2 Bovaird 27 22 5 3 Christian 14 12 2 Cole 105 97 8 3 Cole 8 8 Litchfield 12 7 5 5 McLelland 13 12 I Spooner 15 14 I I Thomas 50 47 3 3 Total 255 227 28 17 Of the 255 patients treated with serum, 227 re- covered and 28 died. Seventeen of the fatal cases were moribund at the time of treatment, so that fail- ure cannot properly be ascribed to sero-therapy. If the 17 unfavorable cases be disregarded, there remains a total of 238 treated cases with 1 1 deaths or a mortality of 4.6%. The mortality of untreated cases, as indicated in a previous table, reaches 28.1%. The mortality is almost seven times higher in the un- treated than in the treated cases of pneumonia of type I. Now the administration of horse serum, intra- vaneously, may be attended with symptoms of anaphylactic shock. This results from the parenteral introduction of foreign proteins in those subjects who are peculiarly sensitive. This phenomenon is not limited to antipneumococcus serum, for it is occasion- ally seen after the use of diphtheria antitoxin. The 54 RECENT CLASSIFICATION AND symptoms are varied, and include suffusion of the face, restlessness, tachycardia, urticaria, and collapse. A history of asthma or of hay-fever in the patient in- vites caution. Irrespective of history, the cutaneous test for the detection of protein sensitization may well be applied as a preliminary measure. The test is similar to the Von Pirquet test, the horse serum being substituted for the tuberculin. In case of sen- sitiveness to foreign protein, an erythema and a localized area of oedema appears at the site of in- jection within an hour. In order to desensitize pa- tients, even those who do not react to the cutaneous test, O.Sc.c. to Ic.c. of horse serum is administered sub- cutaneously one-half to one hour before the intra- venous injection. Another good rule to follow is to allow 15 minutes for the injection of the first 15c. c. of serum. If the patient reacts to the skin test, de- sensitization may be accomplished by giving small doses, l-40c.c., which are gradually increased. Of greater concern are the symptoms of non-speci- fic serum intoxication, which are similar to those seen in the administration of salvarsan. These may appear during the treatment or may come on an hour or two afterwards. The symptoms resemble those of ana- phylactic shock and may include in addition vomit- ing, sweating, and high rise of temperature. To minimize these alarming effects, the slow injection of the serum, maintained at body temperature, be- comes imperative. The administration of the serum may have to be interrupted, and resumed an hour or two later. Adrenalin and atropin happily afford prompt relief. TREATMENT OF PNEUMONIA 55 Unlike anaphylaxis and non-specific serum in- toxication, the symptoms of serum sickness appear delayed, — from seven to fourteen days after the ad- ministration. The latter are characterized by ele- vation of temperature, skin rashes, glandular en- largement, and joint pains. They are never serious. Urticaria is the most common manifestation, and it is readily controlled by adrenalin and carbolic washes. The encouraging results of antipneumococcus serum I in therapy have been disclosed by the fore- going analysis of available data. Statistics have furthermore shown a percentage incidence of type I infections amounting to 29.6. This equals the in- cidence of pneumonia caused by the other two "fixed types" of pneumococcus, types II and III. Serum from animals, actively immunized against these two types, has proven of no value in treatment of pneu- monia. To offset this failure, Humbert has made a most excellent suggestion, which merits serious con- sideration. He has proposed the "endless chain" treatment of pneumonia. He avails thereby of the immune bodies which are present in the serum of those persons who have recovered from the disease. Convalescents are bled. The serum is pipetted off after clotting has taken place and centrifuged. It is next inactivated by heating to 36°C. for thirty minutes. Tricresol (3-10 of 1%) is added as a pre- servative. The serum is kept in stock, properly la- belled so as to indicate the type of pneumococcus in- fection from which the donor has recovered. In 56 RECENT CLASSIFICATION AND treatment, the type of pneumococcus causing the pneumonia is determined, and the appropriate serum is administered intravenously, type for type. Humbert reports that of 11 cases treated in this manner, one death occurred in a type III infection. The incidence of types in his series was as follows : Type I, 2 cases; type II, 4 cases; type III, 4 cases; and type IV, 1 case. The number is too small to admit of definite conclusions. Nevertheless, anal- ogous experiences during the pandemic of influenza indicate that the use of human convalescent serum in acute pulmonary infections is attended with bene- ficial results. McGuire and Redden have treated 151 patients with influenzal pneumonia by intravenous injections of human convalescent serum. Of this number only six died, giving a mortality of 4%. O'Malley and Hartman have had similar gratifying results in the treatment of influenzal pneumonia by the same method. Only three deaths occurred in 46 treated cases, making a mortality of 6.5%; whereas of 111 untreated patients, 28 died, yielding a mortality per- centage of 25.2. The cause of influenza has not yet been definitely established. Nevertheless, patients recovering from the infection have in their blood stream antibodies which are capable of conferring passive immunity. It is not unreasonable to suppose, therefore, that serum obtained from patients who have recovered from type II and type III pneumonia may likewise have protective bodies. At least Humbert's sug- TREATMENT OF PNEUMONIA 57 gestlon may well be tried until potent antisera are experimentally produced. Sufficient time has now elapsed to justify the use of vaccines as a prophylactic against some bacterial infections. It has seemed inevitable that their use be extended to the treatment of disease. Some ob- servers maintain that active immunization with a pneumococcus vaccine is of value in the treatment of pneumonia. Accordingly, Rosenow has suggested the vaccine treatment of pneumonia for the type-cases for which there is no antiserum available. Pneumo- coccus vaccine, extensively employed in former years, is at present in disrepute as a therapeutic agent. The cause of failure has been attributed to the toxic ma- terial which is present in heat-killed pneumococcic vaccine. Rosenow has consequently devised a vac- cine of partially autolyzed pneumococci. He elimi- nates the toxic fraction of the bacterial cells and retains the antigenic fraction. Virulent strains of the organisms are grown in glucose broth from 18 to 24 hours. A suspension of the sediment, obtained by centrifugalization, is made with salt solution. The suspensions from various strains are mixed in bottles to which a small amount of ether is added. The bot- tles are thoroughly shaken and incubated for three to five days. During the period of incubation, the bottles are shaken twice daily. Each day a small quantity of the suspension is removed and a current of air is passed through the mixture to displace the ether. Slides are made from the suspension and stained. Incubation is discontinued when 95% of the 58 RECENT CLASSIFICATION AND organisms become Gram negative, and when 5c. c. of the suspension cause no toxic symptoms in guinea pigs. The Rosenow vaccine, therefore, differs from others in that it does not consist of an emulsion of or- ganisms killed by heat, but of an emulsion of organ- isms which have undergone partial digestion. One cubic centimetre of the vaccine, containing 200 billion organisms, is administered daily until the temperature becomes normal and remains normal for one to two days. Rosenow has employed his vaccine in the treat- ment of 200 cases of lobar pneumonia. Of this number, 186 have recovered and 14 have died. In the 186 cases of recovery, there have been strikingly good results in 63, good in 73, and indifferent in 50. Furthermore, the average mortality rate in 95 pa- tients receiving the vaccine early in the disease was 3%; and in the 105 cases receiving the first injection on the third day of the disease or later, was 11%. The average mortality of all cases treated was 7%. The mode of action of this vaccine is still obscure. Further experience will determine the usefulness of active immunization in the treatment of acute in- fectious diseases. In this regard, the opinion of Theobald Smith is as important as it is authoritative — "vaccines applied during disease will be rarely, if ever, life saving, but they may hurry a stationary or languid process which tends towards recovery, by bringing into play the unused reserve of various tissues." TREATMENT OF PNEUMONIA 59 Before the advent of serum and vaccine therapy, the treatment of pneumonia depended upon the no- tions of the individual practitioners. Search was made in vain for a specific remedy. Drug after drug was tried. During my interneship, a routine treat- ment for pneumonia was the administration of 10 grains of quinine every two hours for six doses. The temperature on the day following the treatment in- variably came down to normal, but soon became elevated. The course of the disease was otherwise unaffected. The discovery of optochin, a derivative of quinine, by Morgenroth and Levy, in 1911, be- came, therefore, a matter of unusual interest. These investigators found that optochin possessed a specific bacteriocidal action on pneumococci both in vitro and vivo. Hopes were aroused that the drug might prove of value in therapy. The results obtained in the treatment of 75 cases of pneumonia at the Rockefeller Hospital with optochin were disappoint- ing. No beneficial results were noted. However, the combined use of optochin and im- mune sera in experimental pneumococcic infections has proved more effective than if either one were em- ployed singly. It seems as if optochin enhances the action of immune serum. Moore has demonstrated this property of the drug by the use of mice. One- half cubic centimetre of a 2% oily solution of op- tochin and 0.2c.c. of antiserum II each fail to protect a mouse against 0.06c.c. of a highly virulent culture of pneumococcus type II. Now if the optochin and immune serum are injected simultaneously, pro- 60 RECENT CLASSIFICATION AND tection is afforded the animal against O.Sc.c. of culture — or eight times the amount of culture against which each alone would be ineffective. Chemosero- therapy still remains an inviting field for investiga- tion. The success which has attended the use of salvarsan lends encouragement to the search for chemo-bacteriocidal agents in vivo. The pneumococcidal property of optochin has, furthermore, suggested its use in prophylaxis. Kolmer and Steinfeld recommend a 1:10,000 solution of op- tochin in a 1 :10 liquor thymolis, as a gargle or mouth wash. They are of the opinion that such a solution "may ser\'e to destroy virulent pneumococci as they gain access to the mucous membrane of the mouth and upper part of the throat and prevent their pro- liferation in large numbers." The liability to in- fection with pneumococci by contact has been pre- viously indicated. This liability would, therefore, be appreciably lessened if systematic gargling with the optochin solution were adopted by physicians, nurses, and members of the family of a patient stricken with pneumonia. The application might also be extended to convalescents of the disease. The studies of the last decade have resulted in a better understanding of pneumonia, caused by the pneumococcus. The results achieved in the classifi- cation and treatment of this disease may be pre- cursive of still more effective measures in the con- servation of human life. TREATMENT OF PNEUMONIA 61 NOTE. The writer has drawn freely upon the recent Hterature bearing upon the subject matter of this thesis, and has been bold enough to include his experience and suggestions. Full acknowledgments are made to the investigators mentioned, by whose toil medical science has been enriched. COLUMBIA UNIVERSITY LIBRARIES 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 library rules or by special arrangement with the Librarian in charge. DATE BORROWED DATE DUE DATE BORROWED DATE DUE !I\N2 ^ 154:^ Jn'« ^ i i C2b (747^ Mice J^^ -"X> (