it : i i ' - HH iH} ; Wil i at HT LALLA SEO OR ; t : : ; ith Hi iit o LLTALI ARE HLL i —$S— -— = = i Hl ci —— Digitized by the Internet Archive in 2010 with funding from University of Toronto http://www.archive.org/details/harveylectures10harv THE HARVEY SOCIETY THE HARVEY LECTURES Delivered under the auspices of THE HARVEY SOCIETY OF NEW YORK Previously Published FIRST SERIES . . . 1905-1906 SECOND SERIES . . 1906-1907 THIRD SERIES. . 1907-1908 FOURTH SERIES. . 1908-1909 FIFTH SERIES. . 19009-1910 SVD Mal SEI UBS) Ne. A doidocigoyer SEVENTH SERIES . r91%1-1912 FIGHRAG SERIES Vai) LOTE-TODS NINTH SERIES . . 10913-1914 TENTH SERIES . . 1914-1915 « hice Sueno nee Raps eee 73 Same experiment showing outlines of the curves obtained from five of OHO WORMS | 6. isco teins cee eM bidleowe Gle die oe ese ae 74 Serial leads from sulcus and superior cava in another auricle........... 75 Spread of the excitation wave over the surface of the right auricle..... 76 Direction of spread of the excitation wave on the front of the heart.... 77 Times at which the excitation wave appeared on the front of the same Front of a dog’s heart upon which five contact points were investigated. 80 External and internal contacts placed on the epicardium and endocardium FOSDOCEIVELY 0.4 2 520/00 wveje.o mp) evdyouk ole wiarerk. anes" ay oye oho hie 81 Diagram of the ventricles, showing path of excitation wave through Purkinje tissue... ... oc 0 errs atevere one olevelctereasutnabenene in tks gets iste (a) eee 82 The spread of the excitation wave in the auricle from a central node.... 83 Spread in the ventricle through branches of the Purkinje system...... 84 Growth curves of rats on diets containing a single protein...........-. 111 Growth curves showing effect of the addition of the amino-acids to zein. 111 Growth curves after addition of lactalbumin to zein...........00+005- 112 Growth curve showing effect of 18 per cent. of casein as the sole protein 114 Growth curve showing effect of 18 per cent. of edestin as the sole protein 115 EXPERIMENTAL STUDIES ON METHODS OF PROTECTIVE IMMUNIZATION AGAINST TYPHOID FEVER* PROFESSOR FREDERICK P. GAY University of California ROGRESS in combating bacterial diseases either by pre- ventive inoculation or by vaccine therapy might seem to lie largely in the application of recognized and fairly success- ful methods in new diseases. I venture to suggest, however, that great progress is possible in the present methods that have been regarded as more or less efficient in such diseases as anthrax, cholera, plague, and typhoid fever. The strikingly favorable results of protective immunization against typhoid fever have only recently been fully recognized and emphasized, and are therefore fresh in the minds of us all. Particularly brilliant have been the results attained in this country, and you were so fortunate last year in this Society as to hear the latest report of them presented by Major Russell, who is to a great extent responsible for their general acceptance. I hope that nothing I may say will in the least dampen your enthusiastic support of the current methods of antityphoid inoculation; at the same time I venture to suggest that they may be still further perfected. Let me outline to you the reasons which have led me and my associates to undertake, during the past two years, rather exten- sive experimental studies of the methods of typhoid vaccination. I may further suggest that similar studies are indicated in connection with other bacterial diseases; I have only to mention the first, and in many respects the most perfect, instance of antibacterial prophylaxis, that of vaccination against anthrax in cattle. You doubtless know that, removed from the careful laboratory conditions that marked Pasteur’s triumphant experi- ments, the method is often popularly claimed to be a failure; the living attenuated culture of B. anthracis dispensed by com- * Delivered October 10, 1914. 19 20 HARVEY SOCIETY mercial firms may either fail to protect, or, on the other hand, cause a high mortality among the inoculated animals. Typhoid fever is now the human disease of recognized bac- terial origin that is most efficiently prevented by previous administration of the etiologic agent. And yet the relatively good results that have been attained should, it would seem, only serve to stimulate us to strive for even better results. That further advance is needed is evident for the following reasons: First: We are not certain what preparation of the typhoid bacillus is best for immunizing purposes. Over twenty different typhoid vaccines have been suggested, and almost every one of them is still being advocated as the best. In addition to this number of preparations suggested, wide variations exist as to the size and number of the doses, the intervals of inoculation, and the like. The fact is that the development of the present methods, apart from the animal experiments on which they were based, has been largely empirical. A return to the experi- mental study of antityphoid immunization is therefore advo- cated as the only means of determining the best vaccine prep- aration and the best methods of administering it. Second: The methods now employed protect only relatively well even for short periods of time. The protection afforded has been estimated as from two to six times as great as that enjoyed by the unvaccinated individual under similar cir- cumstances. I am fully aware of the exceptionally good results that have occurred in the United States Army and in certain French garrison towns, but the identical methods employed in these instances have, in other groups of cases, been far less protective. Third: As compared with the enduring immunity that fol- lows recovery from typhoid fever, the protection afforded by artificial immunization is very short, at most of some two years’ duration. Fourth and correlatively: We have at present no sure method of estimating the actual duration of protection in the individual. When does the preventive treatment wear off? When should I be re-vaccinated ? IMMUNIZATION AGAINST TYPHOID FEVER 21 Fifth: The usual methods of antityphoid inoculation, although advisable under any circumstances, are frequently, perhaps in the majority of cases, attended with more or less unpleasant symptoms, which it might be possible to avoid with- out detriment to the protection afforded. HISTORICAL REVIEW Beumer and Peiper? were undoubtedly the first fully to appreciate the possibility of an active immunization against infection with the typhoid bacillus. In 1887 they were able to prove that mice that have recovered from a non-fatal infection with living typhoid bacilli are frequently protected against subsequent, larger, and usually fatal doses of the same organ- ism. In their most successful experiment they found that the best results were obtained by the gradual increase in dosage on successive inoculations, and they further suggest that it may be possible to immunize by means of sterilized cultures, which, as had already been shown, contain the toxic principle of the typhoid bacillus. They raise the question as to whether it might not be possible to immunize human beings by means of gradually increasing amounts of such killed cultures. In the following year Chantemesse and Widal,? following the work of Salmon and Smith * on hog cholera, and of Roux and Chamberlain‘ on malignant cedema, found that they could protect mice against infection with living typhoid bacilli by means of sterilized cultures of the organism. The practical application of these experimental results in animals to the prevention of typhoid fever in human beings did not come until eight years later, following the discovery of the lysins by Pfeiffer. It was A. E. Wright, who, in a preliminary publication in 1896 followed by a fuller account by Wright and Semple ® in the beginning of 1897, first outlined a method of immunizing human beings against typhoid fever. The method as outlined is notable on account not only of the essential facts involved, but also of the orderly and systematic method by which the problem was approached. Wright grew the culture of the typhoid bacilli in bouillon for two or three weeks and 22 HARVEY SOCIETY then killed them by heating to 63° C. (145.4° F.) for an hour, and preserved them with 0.5 per cent. phenol (carbolie acid). These vaccines were then tested for sterility and their toxicity carefully standardized by determining the minimal lethal dose for guinea pigs; the dose chosen for injection in human beings was measured by this toxicity for animals. Wright further in- troduced a method of counting the number of the bacteria in the preparation employed by comparing their number in a given dilution when mixed with a suspension of red blood-cells, the number of which could be accurately determined. The dose of bacteria usually employed was from 750 to 1,000 million. In the same year (1896) Pfeiffer and Kolle* described their method of immunizing and their method of estimating the pro- tection that is produced in human beings by the agglutinins and the bactericidal potency of the serum. They employed agar cultures of an avirulent strain of the typhoid bacillus sus- pended in salt solution and killed by heating to 56° C. (132.8° F.). An amount of this suspension, corresponding to one-tenth of an agar culture or something like 2 mg., was usually given on the initial injection, which was, as a rule, the only one. It is frankly admitted that the symptoms produced by this amount of culture were severe, and the method has since been modified in several ways to avoid these symptoms without essentially chang- ing the principle involved. So much for the first two communications on typhoid immu- nization in human beings. They form the groundwork on which subsequent methods of vaccination against typhoid fever have been built. The various methods that have since been advocated are numerous. Metchnikoff and Besredka * estimate that at least twenty different methods of vaccination have been described and advocated. Friedberger,® in his systematic review on typhoid immunization, enumerates twelve recognized methods. Paladino Blandini*® has actually attempted to test the comparative im- munizing value of seventeen preparations. It is not our pur- pose to describe all these methods in detail, and one who wishes further information on them may consult the systematic deserip- tion of Friedberger ® or of Fornet ™ in regard to them. It will IMMUNIZATION AGAINST TYPHOID FEVER 28 be well, however, to outline the most important of these methods as evidence of the scope that the investigation has taken in per- fecting this type of immunization, and as indicating the tendency which would seem to be leading to its eradual perfection. PREPARATIONS OF THE TYPHOID BACILLUS THAT HAVE BEEN USED AS VACCINES A. Killed Cultures of the Typhoid Bacillus—We have already mentioned that the first two preparations, those of Wright and of Pfeiffer and Kolle, consist essentially in killed cultures, the one, a bouillon culture and the other a suspension of an agar culture. These two original methods have been fol- lowed by many modifications. Thus Loeffler 12 took advantage of the fact that ferments, when dried, resist heating to a con- siderable degree without deterioration, and, regarding the an- tigenic property of the bacillus as ferment-like, dried suspended agar cultures of the micro-organism and then heated them to from 120° to 150° C. (248° to 302° F.). These dried cultures were then pulverized and used in weighed amounts for immuniz- ing animals. He states that such a culture has lost little of its property to produce antibodies. Friedberger and Moreschi ** use a similar dried and heated culture, and administer it intra- venously in very small doses; for example, an amount corre- sponding to 1/4,000 cese in immunizing human beings. It should be noted at this point that the method currently employed of determining the immunizing value of these preparations lies in estimation of the antibodies (agglutinins, lysins, etc.) pro- duced. As we shall later have cause to consider, these esti- mations offer an indication rather of the reaction of the animal body than a sure means of determining the degree of protection that has actually been afforded. As we shall see in a moment, the use of living instead of dead cultures has been warmly advocated by certain observers, and their assertions have apparently convinced several who are not quite willing to adopt such preparations owing to their possible danger, although they endeavor to approach them as far as possible without actually using living micro-organisms. 24 HARVEY SOCIETY It is apparently now the consensus of opinion that bacterial cultures are most antigenic when employed as nearly as possible in their living, unaltered condition, and that heat in particular tends to alter or destroy essential, characteristic, antigenic prop- erties. Several methods have been advocated for avoiding or obviating so far as possible the destructive influence of heat, and at the same time killing the bacteria. Leishmann ** advocates killing the typhoid bacillus at 53° C. (127.4° F.) instead of 56° or 60° C. (132.8° or 140° F.). Vincent * while fully recog- nizing the superior value of living cultures, regards their use as dangerous, and therefore kills the vaccines that he employs by means of ether. As will be later mentioned, we have used alcohol for the combined purposes of killing the typhoid bacilli and accelerating their flocculation and drying. Levy and Bruch ** killed their preparations by shaking the micro-organ- isms in a medium containing galactose, and find that organisms prepared in this way immunize guinea pigs as well as the living cultures, and that these two preparations are far superior to killed cultures in corresponding amounts. Fornet ™' regards the unpleasant effects that are produced by the heated vaccine as due not only to the heating itself but also to the presence of a large amount of albumin in the culture medium. He therefore grows his micro-organism in a medium containing only a small amount of peptone and kills them by heating to 55° C. (181° F.) for fifty-five minutes. Courmont and Rochaix?’ killed their preparations by heating to 538° C. (127.4° F.), and further modified the usual method by administering the antigen through the rectum. Nicolle, Connor and Conseil ** heat their bacteria to 55° C. (131° F.) for forty-five minutes, and then to 52° C. (125.6° F.) for thirty minutes more, and inject intravenously. Wassermann ”® insists that the antibodies produced by typhoid bacilli heated to 53° C. (127.4° F.) are not markedly better than when they are heated to 56° C. (132.8° F.). Renaud *° has advo- cated the use of ultra-violet rays to kill the bacteria. B. Extracts of Bacteria.—In addition to killed cultures of bacteria, numerous extracts and preparations derived from the bacteria have been advocated for the purpose of immunizing IMMUNIZATION AGAINST TYPHOID FEVER 25 against typhoid fever. Hahn has recommended the extract obtained from masses of bacteria by means of the Buchner press. McFadyen and Roland ”? have utilized liquid air as a means of killing bacteria and obtained from them an extractive substance. Neisser and Shiga ** have utilized free receptors obtained by autolysis of bacteria at body temperature in salt solution. Wassermann * has suggested a similar method with the autolysis produced by distilled water. He subsequently dries the extract obtained in this way and uses it as a vaccine powder (Impfpul- wer). Brieger and Mayer * have used a watery, filtered extract of shaken bacteria. Bergell and Meyer *° have used an extract of dried bacteria obtained by treating them with dilute hydro- chloric acid. Various so-called soluble toxins of the typhoid bacillus have also been suggested for immunizing purposes by Chantemesse,”’ by Werner,”® and by Rodet, LaGriffoul and Wahby.?® The ex- tract of bacteria obtained by the method of Jez *° has also been suggested. C. Living Cultures of the Typhoid Bacillus—tiving cul- tures of the typhoid bacillus, usually more or less modified in their pathogenicity, have been warmly advocated by certain observers as producing the best immunizing preparations in a manner similar to the methods that have been employed in deal- ing with other diseases, notably in cholera (Strong and Kolle). Castellani *t uses an avirulent strain of the typhoid bacillus in the form of recent bouillon cultures which are then partially killed by heating to 50° C. (122° F.) for one hour. Such a modified culture produces rather severe local and general symp- toms, but when given twice would, to judge from Castellani’s results, produce a most satisfactory degree of immunity, which apparently has lasted in a number of cases on which he reports for at least four years. He suggests, as an alternative, that the first injection may consist of a killed culture followed by a living culture on the second inoculation. In addition to the superior immunizing properties of the living culture, it is also pointed out by Fornet ** that the killed cultures in a given dose give more reaction because the split products of proteins, which are recog- 26 HARVEY SOCIETY nized to be toxic, are liberated by heat. Living cultures have also been employed by Pescarolo and Quadrone.*? The form of living cultures which has been advocated by Besredka will be considered under the next heading. As has already been mentioned, living cultures are generally admitted to be of superior immunizing value by many who are not willing to adopt them, owing to the real or fancied dangers coincident with their use, and this has led to an attempt to approach the condition of living bacteria without actually employing them. (Compare Vincent,?? Levy and Bruch,'* and particularly Metchnikoff and Besredka, to be considered presently.) D. Sensitized Cultures of the Typhoid Bacilus—The method of active immunization by means of sensitized vaccines, that is to say, by cultures that have been first treated with an immune serum and then killed, was introduced by Besredka *° in 1902. This method is not infrequently referred to as sero- vaccination, but it differs from the method properly called sero- vaccination suggested by Leclainche ** in swine erysipelas and by Calmette and Salimbeni* in plague, in that the excess of immune serum which these authors used is removed from the treated bacteria. It was found, as Besredka notes, that this excess serum tends to produce simply a passive immunity instead of the active immunity which is produced by the cultures treated with immune serum, and washed. Apart from his original experimental work, Besredka did not deal with the practical aspects of sensitized vaccines until the experimental work on typhoid fever in apes was taken up by him in collaboration with Metchnikoff in 1911. In the meantime, however, sensitized vaccines had been tried out apparently with considerable suc- cess in at least three instances. Marie *° had been able to utilize the principle in treating rabies virus, Dopter *’ in vaccination against dysentery, and Theobald Smith ** in a similar way found that he could produce active immunity by a balanced mixture of diphtheria toxin and antitoxin. This method of active im- munization against diphtheria by a balanced (sensitized) mix- ture of toxin and antitoxin has recently been applied to human beings by von Behring.*® The principal advantages of this IMMUNIZATION AGAINST TYPHOID FEVER 27 method, as originally pointed out by Besredka and as apparently proved, are, first, it produces little or no violent reaction on inoculation in instances in which the untreated bacteria them- selves are distinctly irritating, as, for example, in plague ; sec- ond, it gives rise to an immediate though transitory passive immunity; third, it produces, eventually, an active immunity which is as enduring and as rapidly formed as when untreated bacteria are used, Not a little experimental work was done with sensitized typhoid vaccine before the work of Metchnikoff and Besredka, which will be taken up later. Paladino Blandini* made a very careful study of seventeen different typhoid vaccines, com- paring their relative immunizing properties. His experiments were carried out on guinea pigs, which were treated by the various preparations and subsequently given an intraperitoneal dose of living typhoid bacilli. The vaccines tested include small doses of living cultures; killed cultures prepared after the method of Pfeiffer and Kolle, and of Wright; several ‘* soluble toxin preparations ’’ of the typhoid bacillus; nucleo-albumins ; and extracts of the typhoid bacillus prepared in different man- ners. Compared with these methods was the sensitized vaccine employed by Besredka, and Blandini was able to demonstrate that the latter method was by all means the most protective. Not only were guinea pigs protected for at least four months, but it was found that their serum, which contains sensitizers, also protects normal animals against infection. Ardin-Del-teil, Negre, and Reynaud *° find that the use of sensitized typhoid vaccine in rabbits and human beings gives rise to relatively small amounts of agglutinins, but the bactericidal properties of the serum of those treated in this manner are much higher than of those treated by the ordinary cultures. These authors have also obtained very favorable results in treating cases of typhoid fever with this vaccine. This failure of sensitized or agglutinated cultures to produce potent antibodies had already been noted by Neisser and Lubowski “! and by Pfeiffer and Bessau.*? Garbat and Meyer ** immunized rabbits either with sensitized or with whole cultures 28 HARVEY SOCIETY and compared the properties of the serums of the two sets. The serum obtained by immunizing the rabbits with sensitized cultures agglutinated and gave the fixation reaction much less strongly than the corresponding animals treated with unsensi- tized cultures. The serums of the sensitized animals, however, were more bacteriotropic and protected animals experimentally much better than the serum of the animals treated with the plain vaccine. This fact is incidentally evidence of the unreliability of the agglutinin test as a measure of the grade of protection against typhoid infection, a matter which we shall later consider. In 1911 Metchnikoff and Besredka § first reported their very important work on experimental typhoid fever in anthropoid apes. They found that the chimpanzee and the gibbon when given the dejecta from cases of typhoid fever, or lavishly con- taminated food, or, in their latter experiments, when the mouth of the animal is painted with cultures of the typhoid bacillus, after eight days’ incubation develop characteristic typhoid fever. Fifteen or sixteen monkeys that were tested gave positive blood cultures on the tenth day, their serum contained agglutinins, and, in three instances, death due to the typhoid infection fol- lowed. The temperature of these animals ran as high as 40.8° C. (105.4° F.). Peyer’s patches were found swollen but not ulcerated. The only detail in which this experimental disease would seem to differ from typhoid fever in human beings is the fact that the spleen in the chimpanzees is not distinctly enlarged. Having determined in this way that the typhoid bacillus is in reality the cause of typhoid fever and not some adherent filter- able virus, as in the case of hog cholera, they turned their atten- tion to methods of immunization against the disease. Their first series of experiments are far from convincing, although they draw from them rather sweeping conclusions. In the first place, the number of animals employed was necessarily small, owing to their expense and the great susceptibility of the animals to extraneous infection, so that in this first series, in which there were five experiments, there was only one vaccinated animal in each case tested with an untreated control. They come to the conclusion from their preliminary attempts that neither Vin- IMMUNIZATION AGAINST TYPHOID FEVER 29 cent’s typhoid vaccine, an autolysate of the typhoid bacillus, nor a killed culture sensitized according to the method of Besredka will immunize anthropoid apes against a subsequent infection by the mouth. Apart from the necessarily small number of animals in each experiment, certain other objections may be made to their experi- ments and conclusions, part of which Vincent ** has pointed out. In the first place, the animals were tested very shortly after finishing the immunizing treatment, usually in from four to six days, which may reasonably be regarded as too brief a period for the establishment of the highest grade of immunity; and, second, the dose given the animals by the mouth, which is not stated very definitely, seems in all events to be extremely large, much larger, indeed, than in the natural infection in man. They conclude from their experiments that heated vaccines do not protect anthropoid apes against typhoid infection, and, owing to the analogy of the disease in these animals with the human syndrome, they regard such vaccines as unfitted to protect human beings. Inasmuch as the protective vaccination of man against typhoid fever by means of heated vaccines is generally recognized as reducing the morbidity to one-half or one-sixth of the normal, the logic of their position does not seem sound. It would seem better to have concluded that the immunization or the disease in anthropoid apes, as they have produced it, are not similar to their analogs in human beings. In their second communication, Metchnikoff and Besredka *5 present experiments which, with the exception of one experiment in which Vincent’s vaccine is used, are more convincing. Their second series deals with the protective effect of vaccination by a living paratyphoid B culture, and by sensitized living cultures of the typhoid bacillus against subsequent infection of the latter organism in the manner described. These experiments comprise more animals than the previous series, and in addition the infec- tion is not given until from ten to fifteen days after the comple- tion of the treatment. Vincent’s culture again failed to protect, whereas the living sensitized typhoid culture and the living paratyphoid B cultures uniformly saved the animals. It would 30 HARVEY SOCIETY be far from safe, however, to conclude from these experiments that no preparation of heated typhoid bacilli has the power to protect even anthropoid apes against subsequent infection by the typhoid bacillus. We have examined the experimental details of the work of Metchnikoff and Besredka, not in a spirit of unfavorable criti- cism, because, as will later be shown, we personally believe that their conclusions were in part correct, though the data on which they have based them have not convinced us. We may here examine the results that have already been obtained in immunizing human beings by the living sensitized cultures recommended by Metchnikoff and Besredka. Broughton Alcock,** working under their direction, was the first to report on the harmlessness of the method, its freedom from untoward effects, and its apparent protective value. He advocates a dose corresponding to 1/100 of an agar culture, which is equivalent to 500 million typhoid bacilli. Their organisms are sensitized by a few drops of a strong antityphoid serum, the exact potency and amount of which is not stated either by Metchnikoff and Besredka or by Broughton Aleock. After standing for twenty- four hours the clumped bacteria are washed in salt solution and resuspended in an aliquot portion of normal saline. A living sensitized vaccine prepared in this manner keeps for at least four months on ice. Objections have been raised by Vincent and others as to the danger of injecting living micro-organisms into the human body, but Metchnikoff and Besredka ** have reported that careful tests fail to show that human beings treated in this manner become carriers of the typhoid bacillus or eliminate the micro-organism through the fees or urine. In view of the uniformly good results that have been obtained, there seems little reason further to suspect the harmfulness of the method. In a recent report Metchnikoff and Besredka ** have further endeay- ored to determine the optimal vaccinating dose and interval of this sensitized vaccine, and Besredka*® reports some of the favorable results that have been obtained. Cadeau °° vaccinated twenty-five patients, and only two showed slight general symp- toms. In an asylum at Briqueville where many eases of typhoid IMMUNIZATION AGAINST TYPHOID FEVER 31 fever existed in 1912, 516 persons were vaccinated, and in the twelve months subsequent no cases occurred among them, whereas four occurred in the 343 persons who were not immu- nized. Marie reports similar absence of general reaction. In all, Besredka has dispensed 10,000 doses of the vaccine and no unpleasant results have been reported except when the injections were given intramuscularly. He recommends a dosage of from 500 million to 1,000 million, although larger amounts could be given without harm. It is, of course, too early to draw any conclusive figures as to the actual protective value of this method, although its harmlessness and freedom from untoward symptoms seem to have been proved. PERSONAL INVESTIGATIONS In view of the general acceptance of the relative protective value of immunization with killed cultures of the typhoid bacil- lus, and the claim that Besredka’s method of vaccination with sensitized living typhoid bacilli insures all the advantages of the more unpleasant method of treating with dead cultures, and protects as well or better, it would seem desirable to devise some method of animal experimentation which would enable us to settle conclusively this and many other debatable questions that arise in connection with this extremely beneficial process. There are still many questions in connection with typhoid vaccination which remain unsettled, such as the dosage to be employed, the interval at which the injection should be given, and, most im- portant of all, the duration of the protection that is afforded by any particular method and to any particular individual. Animal experimentation has, to be sure, been used, for the most part on guinea pigs that after treatment with various typhoid vaccines are infected in the peritoneal cavity. It seems justi- fiable, however, to question the transference of results obtained in preventing a banal peritonitis in this animal to human beings. The method of testing typhoid vaccines on anthropoid apes that Metchnikoff and Besredka have described is undoubtedly the ideal one, but practically impossible to employ for the settling 32 HARVEY SOCIETY of details of method owing to the expense of the animals and the difficulty in obtaining them in any numbers. We have already described in some detail our method for testing the immunizing power of a given typhoid vaccine by means of the ‘‘rabbit-carrier’’ condition.** This carrier con- dition, as originally pointed out by Blackstein, may be produced in rabbits by the intravenous injection of certain strains of the typhoid bacillus, It presents a very close analogy to the disease known as typhoid fever in man. It is a bacteriemia, with a particular localization of the micro-organism in the gall-bladder, which viscus shows the characteristic lesions described in human beings, including the formation of gall-stones under favorable conditions. By employing one-half of a standard 10 per cent. blood agar culture of a certain strain of B. typhosus, we have been able to produce this condition in normal rabbits with rela- tive certainty. Thus of thirty control rabbits, twenty-eight, or 93 per cent., either became permanent carriers with positive cultures from blood or gall-bladder, or, in a few instances, died acutely within twenty-four hours with symptoms of intoxication. For further discussion of the experimental advantages of this method over others proposed for this purpose, excepting experi- ments on anthropoid apes, we refer readers to our previous communication. As we have just suggested, the presence of antibodies, par- ticularly of agglutinins, is not a measure of the degree of pro- tection against typhoid fever, either in animals or in human beings. Antibodies indicate reaction, which usually parallels protection but is by no means synonymous with it. In typhoid fever in particular the eventual absence of agglutinins in those recovered from and thereby protected from the disease indicates that serum tests are by no means to be confused with protection, which is more likely to be of a cellular than of a serum type. This indicates emphatically the necessity of testing the prophy- lactic immunizing value of any given vaccine by actual infection of the vaccinated animal rather than by any test of its anti- bodies. In our previous paper we have put forward our reasons for regarding the artificially produced bacteriemia in rabbits IMMUNIZATION AGAINST TYPHOID FEVER = 33 as analogous to typhoid fever in human beings, which with in- creasing agreement is now regarded primarily as a bacteriemia rather than a purely intestinal disease. The fact that we are able to prevent this carrier condition by previous immunization with one typhoid vaccine, whereas another preparation of the typhoid bacillus in the same amount does not prevent it, is further evidence that the method we have chosen is a delicate one and further suggests the variations in result that have been obtained in using different typhoid vaccines in human beings. Preparation of Typhoid Vaccines for Comparative Tests.— The first essential in comparing various vaccine preparations seemed to us to lie in a standardization of the amount of sub- stance employed. The methods usually employed in standard- izing vaccines for clinical use may be accurate enough in view of the fact that the optimal dose is largely empirical and a little less or more may make no difference in the results. The fact that the usual methods of enumeration of bacteria by counting by the Wright method are being constantly changed is evidence enough that it is a far from satisfactory condition. We had already adopted the far more accurate method of dried, weighed bacterial preparations when the work of Wilson and Dickson *? came to our notice. These authors have carefully determined the actual weights of the doses of various bacteria commonly employed, checking their weighing results by counting the bac- teria in a given volume of suspension. They find, for example, that there are on an average 8,000 million of typhoid bacilli in 1 mg. of a typhoid vaccine dried on platinum foil. We were surprised to find how closely our estimations, although obtained in different manners, agree not only among themselves, but also with the results of these authors. Our method of preparing dead typhoid vaccine has been to add normal saline (0.85 per cent.) in proportionate amounts to suspend forty-eight-hour cultures grown on Blake bottles containing 125 ¢.¢c. of 2 per cent. agar (titrated +1) or, in certain experiments, of the same medium containing 10 per cent. fresh defibrinated rabbit blood. To such a suspension is added an equal volume of absolute alcohol, which kills the culture in the proportions used in less 3 34 HARVEY SOCIETY than fifteen minutes, flocculates out the organisms so that they can be easily collected by centrifugalization, and accelerates their drying in an unglazed porcelain plate in partial vacuum over sulphuric acid. Our experiments on the relation of the number of bacteria to weight agree very closely with those of Wilson and Dickson (8,000 million of typhoid bacilli to 1 mg. of dried pulverized typhoid powder). In all our experiments with vaccine prepara- tions, then, we have used weighed amounts of dried and ground bacilli as a basis for comparison, except in those in which living cultures of the organisms have been employed, in which case comparisons were made in another way. Methods of Immunization—Having determined on the use of weighed amounts of dried, ground bacteria for comparative testing of typhoid vaccines, the next essential was to decide on the best method of immunization of the rabbits that are to be tested for protection against the carrier state. Fornet and Miiller®* and Bonhoff and Tsuzuki** have already shown that for the production of certain antibodies an intensive method of immunization may be employed by re- injection at short intervals. In collaboration with Fitzgerald °° I have still further perfected this method. We were able, for example, to produce high-grade hemolysins to sheep corpuscles by injecting rabbits intravenously on three successive days with 1 c.c. of washed sheep blood. These lysins we find to be present in high degree four or five days after the last injection. Similar, although not quite so perfect, results were obtained in the pro- duction of precipitins and also in the production of bacterial agglutinins by Miss Edna Locke.®* On the basis of these experi- ments, we had already determined that, provided a vaccine did not produce too severe symptoms, the period over which im- munization extended might be markedly decreased without any corresponding deduction in the intensity of the immunity effected. It is of interest to note that Fornet * has also adopted a rapid method of immunization in vaccinating against typhoid. We have ourselves recommended, and Force *? has employed, immunization at intervals of two or three days in human beings, IMMUNIZATION AGAINST TYPHOID FEVER 35 as will be mentioned later. The rapid method of immunization at short intervals not only may be shown in the case of rabbits to effect an equally high grade of immunity as determined either by antibody content or by actual resistance to infection, but also has the further advantage of actually producing less harm- ful effects when large amounts of the untreated vaccine are used. We have repeatedly noted that if a large total amount of vaccine is given to rabbits during a period extending over three weeks there is a larger percentage of mortality due to cachexia than when the same amount of vaccine is given within three days. The immunizing doses have invariably been three in number and have been administered either intravenously or subcutaneously ; in all our later experiments it has seemed better to adopt the intravenous method as giving greater uniformity of results. The advantage of immunization at short intervals as against longer intervals was first tested by comparing the resistance to the carrier state of two sets of rabbits, the one immunized by three intravenous injections on three successive days, the other by three equal doses administered intravenously at three-day intervals. There was found to be no marked difference in rela- tive resistance of these two sets of animals when given the test dose two months later. The agglutinins averaged the same titer in the two sets. In a more systematic experiment, a comparison of three daily with three five-day interval doses was made. The results of this experiment are expressed in the following table: TABLE 1.—Resvuits or LONG- AND SHort-INTERVAL METHODS OF TypHomw IMMUNIZATION IN RABBITS Died During Vaccination Carriers Recovered ——E nl Short intervals. .............ceeeeee Long intervals ..........+.+++++-e0+: 0 3 4 4 3 0 This experiment as well as other similar ones would seem to indicate the superior value of the short-interval intensive method of immunization. 36 HARVEY SOCIETY Dosage of Vaccine Preparations——We have no contribution to make on the optimal dose of typhoid vaccine for prophylactic use in human beings. As will be seen later, our choice for such purposes has been largely empirical and has followed accepted usage. In our comparative experimental work, however, it has been necessary not only to use preparations derived from weighed, dried bacteria, but also to determine with increasing accuracy the duration of the protection afforded by a given amount of a standard preparation. In our first experiments we immunized separate series of rabbits each with a given preparation and then tested representatives of each series at several successive periods. Inasmuch as we started with rather large doses of vaccine, the time before the protection wore off was rather long and the method time-consuming. By decreasing the total vaccinating dose we were finally able to estimate with certainty that a given amount of one type of vaccine would protect for five or six weeks; with this point determined, it was easy to compare the immunizing value of other preparations with this one. In nearly all experiments, two or more control normal animals have been injected, although their use is not absolutely imperative since the less well-protected animals be- come carriers and the percentage of carriers in each series gives the basis of comparison. Determination of the Typhoid Vaccine of Election for Im- munization.—The carrier condition was at first determined in the immunized and control animals by cultures taken from the blood ten and twenty days after inoculation; the results were checked by the post-mortem condition and cultures from the bile when the animals died, or in the case of survivors by chloro- forming them from fifty to seventy days later. A later method which gives quicker results is to chloroform all the survivors at the end of ten or fourteen days and make cultures from the blood and bile. A perfectly immunized animal is one that gives negative blood cultures and at death presents no lesions, par- ticularly of the gall-bladder, and gives negative cultures from the bile. IMMUNIZATION AGAINST TYPHOID FEVER 37 Throughout our experiments it has been our effort to kill our cultures of B. typhosus in a manner designed to obviate, so far as possible, the deteriorating effect of heat that has been evi- denced by the work of numerous observers mentioned before. The introduction of heat, at all events for the purpose of obtain- ing sterile cultures, seems to us quite unnecessary, and, as I pointed out a number of years ago,°** cultures of the dysen- tery bacillus may be sterilized and preserved by the simple addi- tion of 0.5 per cent. trikresol or phenol, without destroying their toxic and antigenic properties to the same degree as in corre- sponding cultures that are first killed by heat at 60° C. (140° F.) and then preserved by trikresol. Very little attention appears to have been paid to this observation, although we believe that several have utilized the method with advantage, killing the cultures either with liquor formaldehydi or with trikresol. In the present series of experiments we have desired not only to kill the bacteria with little harm but also to collect and dry them in order that the dosage employed might be reckoned from weighed amounts of the dried bacilli, which had been previously ground in an agate mortar. For this purpose, as already stated, we precipitate our various bacterial suspensions with equal parts of absolute aleohol, which accelerates the centrifugalization and the drying of the preparation. If the culture is to be sensitized, a definite amount of a potent, immune serum from the rabbit is previously added, the dosage being estimated in accordance with the agglutination titer of the serum in question and the number of bottles from which the bacterial suspension was obtained, the amount being 1 ¢.c. of a serum that agglutinates the organism in question in a dilution of 1: 20,000 to the sus- pended growth from each of the seven Blake bottles. After standing for two or three hours in an incubator at 37° C. 98.6° F.), the sensitized culture is left in the ice-chest over night, centrifugalized, the supernatant fluid removed, the sedi- ment washed in the original amount of normal saline, re- centrifugalized, and the volume restored to normal with sterile saline. This washed sensitized culture is then flocculated by 38 HARVEY SOCIETY adding an equal volume of absolute alcohol, dried, ground, and weighed as in the unsensitized culture. We first compared the protective value of sensitized and unsensitized cultures. The results of our experiments are given in full in our fifth article on antityphoid immunization, in the Archives of Internal Medicine.*® When parallel series of rabbits treated with sensitized and unsensitized vaccines are tested some weeks later, it is found that the serum of the unsensitized vaccine animal will agglutinate the typhoid bacillus rather better than the serum of a sensitized vaccine animal. The animals treated with sensitized vaccine are found, however, to be much better protected from becoming carriers on the intravenous administration of living typhoid bacilli. The method which we have described for preparing and grinding the vaccine is essentially similar to the one described by Besredka for the preparation of endotoxins, and it may, in- deed, be shown in the case of the untreated vaccine that the supernatant fluid contains the greater part of the toxic sub- stances of the whole vaccine (Besredka,® Stenitzer™). It seemed to us important to determine at this point whether or not the supernatant fluid of such a ground and suspended prepa- ration of typhoid bacilli, or the sediment of bacterial bodies left on centrifugalizing this mixture, contains the immunizing principle. In experiments in which the immunizing value of the two preparations is shown, it is very evident that it is the sedi- ment of bacterial bodies deprived to a great extent of toxicity of the whole vaccine which is the essential immunizing prin- ciple. This immediately suggests the advisability, from the point of view of symptoms produced and of protecting value assured, of utilizing simply the remnants of the bacterial bodies obtained in such a vaccine rather than the whole vaccine itself. Experiments designed to test out the comparative value of the whole vaccine, on the one hand, and the sediment, on the other, show very distinctly that the sediment possesses not merely as much but apparently actually more immunizing value than the whole vaccine. In comparing the whole suspended vaccine with the sediment derived from such a vaccine, it should be IMMUNIZATION AGAINST TYPHOID FEVER 39 noted that every technical error necessary to produce the sedi- ment from the whole vaccine would tend to lessen the amount of material in the sediment, and the superior immunizing value of the latter is therefore so much the more striking. In one experiment the percentage of failure to become carriers shown by rabbits immunized with the four different vaccines was as follows: Rabbits given whole unsensitized vaccine recovered in 33 per cent. Rabbits given unsensitized sediment recovered in 60 per cent. Rabbits given whole sensitized vaccine recovered in 66 per cent. Rabbits given sensitized sediment recovered in 75 per cent. We seem so far to have determined, then, that killed, sensi- tized cultures are more protective than plain, unsensitized, killed cultures of the typhoid bacillus and that the protection is due to the bacillary sediment and not to the supernatant endo- toxic fluid. These sediments, moreover, seemed distinctly more protective than the whole vaccines from which they were derwed. In view of the assertions of Metchnikoff and Besredka, it remained to determine by our method whether or not living, sensitized cultures have any superior immunizing value over sensitized cultures killed by alcohol. For this purpose we have compared the living sensitized culture first with whole dried sensitized culture and then with the sensitized sediment. In these experiments the original bacterial suspension is sensitized, washed, and divided into two parts, one of which is left intact and the other precipitated by alcohol, dried, ground, weighed, and suspended in the original amount of fluid. Here again technical errors would militate against the immunizing strength of the dried culture, particularly when the latter is carried further to the sediment preparation. Such experiments show, first, that living sensitized cultures have a slight superiority over whole dead (alcohol) cultures, but, second, that dead sensi- tized sediment protects better than the living sensitized vaccine. 40 HARVEY SOCIETY Protective Vaccination in Human Beings by Means of the Modified Sensitized Culture—Our experimental results had reached a stage nearly a year ago when we felt justified in recommending the whole sensitized culture that we have de- scribed for use in protecting human beings against typhoid fever. It seemed demonstrated from the work on rabbits that the sensitized vaccine, washed, precipitated, killed with alcohol, and ground, would protect at least as well and probably better than the ordinary vaccine that had not been sensitized, and we antici- pated much fewer untoward symptoms following vaccination by the use of this sensitized vaccine. This absence of untoward effects has been fully justified in experience. Up to the present time over 1,000 students in the University of California have been immunized by Professor Force of the Department of Hy- giene. The absence of reaction or the mildness of reaction has been very striking. Force” has already reported on the first 261 cases by this method, and less complete analyses of the subsequent cases agree with the results obtained in the smaller number. In these 261 cases—apart from the local symptoms of induration, redness, and slight tenderness for one or two days—general symptoms, even of a mild grade, were very seldom observed. The vaccine was administered on alternate days, the treatment being completed within a week. The very fact that it was possible to give the second and third vaccinations on the second and fourth days after the first indicates strongly that even the local condition was so mild as to allow the subse- quent inoculations. The dosage at first employed was */,, mg. of dried culture suspended in 0.5 per cent. carbolated salt solution, corresponding, as we have determined, to 500 million bacteria. Since this produced little or no reaction, we have since increased it by one-half (?/,, mg.). In only six cases in the total number of inoculations practised (671) did the tempera- ture rise above 38° C. (about 100° F.). The symptoms were somewhat more pronounced in women than in men. In two cases of arrested tuberculosis the reaction was more marked and in a few cases that gave a history of previous typhoid fever the reaction was also slightly more pronounced. More recently we IMMUNIZATION AGAINST TYPHOID FEVER 41 have employed the sensitized sediment vaccine in accordance with the experimental results that have just been obtained, which show conclusively that this substance is more immunizing than the whole sensitized preparation. This sensitized sediment gives even fewer reactions than the whole sensitized vaccine. Thus in a recent series that Force has vaccinated with sensitized sedi- ment, using a dosage corresponding to 750 million bacteria, in- cluding 672 separate injections, malaise was noted in 1.8 per cent.; fever (above 99° F.) in 4.5 per cent.; pain in back in 2.4 per cent.; pain in head in 9 per cent., and pain in the arm in 83 per cent. In cards returned to us from 231 outside patients to whom the vaccine was distributed, 215 report total absence of local and general reactions (93 per cent.). Ten cases were marked as having shown generalized symptoms and three cases of typhoid recoveries and three patients who had been previously vaccinated also showed slight symptoms. These reactions may be compared with the following: Hartsock * obtained a mild reaction, by which he specifies ‘‘temperature up to 100° F. with slight general reaction, malaise, and considerable local tender- ness,’’ in 83 per cent. of men treated with U. S. Army vaccine. Hatchel and Stoner * in 1,326 cases vaccinated with a polyvalent vaccine found malaise in over 58 per cent, and fever in 43 per eent. Albert and Mendenhall ** used the U. S. Army strain in regular doses and intervals and found that the local reaction lasted for from three to five days. Fever as high as 103° F. was usual, and malaise, headache, and insomnia were common. It must not be understood that we recommend this prepara- tion of vaccine because it produces less unpleasant symptoms on administration but primarily because, so far as can be judged from our animal experiments, it protects better. Both Wassermann ** and Vincent ** have suggested the use of polyvalent vaccines; although the typhoid bacillus has been regarded as an unusually fixed species, the occurrence of minor biologic peculiarities, particularly in the fermentation of sugars, has been noted in the collection of organisms which we have studied. In accordance with these facts we have recommended a polyvalent sensitized sediment for prophylactic immunization 42 HARVEY SOCIETY against typhoid fever in human beings. The organisms, five in number, which have been used to produce the polyvalent mixture were isolated from recent cases of typhoid fever in the vicinity. An additional advantage in the use of sensitized over unsensitized cultures is that any number of strains of the typhoid bacillus may be used as in our polyvalent vaccine without pro- ducing any more symptoms than with a monovalent vaccine. It is particularly unnecessary to seek a ‘‘mild strain’’ as in the case of the U. S. Army vaccine. Our various strains are grown separately on Blake flasks, and the suspended cultures mixed and treated with a polyvalent immune serum obtained by im- munizing rabbits with each of the strains in turn. This poly- valent sensitized sediment vaccine is now being manufactured and distributed free of cost by the California State Board of Health to any physicians in the State who may apply. THE TYPHOIDIN REACTION One of the greatest difficulties that has been present in deter- mining the protective value of typhoid immunization as a whole has been the impossibility of determining the protection of a given group of persons by other means than the careful study of morbidity statistics among vaccinated people over a long period of years (Firth®). This difficulty delayed the final acceptance of typhoid immunization for at least eight years (1896-1904). Still less have we any means of determining whether or not a given person who has been vaccinated is actually protected against typhoid fever. We have already described in full the skin test which Gay and Force ** have em- ployed as of presumptive value in testing resistance to typhoid infection. A new summary of our results with the typhoidin may be of interest: In 44 cases giving a definite history of typhoid fever from four and one-half months to forty-one years previously, 40 have given a clear-cut skin test to the concentrated growth of the bacillus. Of the four negative tests, two gave a strong skin reaction to a similar preparation made from B. paratyphosus A. which we have found not to oceur in anything like the same IMMUNIZATION AGAINST TYPHOID FEVER 43 intensity in those cases that react to typhoidin. Only one of the remaining negative cases was further tested with the paraty- phoidin solution and therefore only one of two cases was unex- plainably negative and 95 per cent. (40 of 42) positive. We have a growing set of observations on comparative tests with two paratyphoidin solutions (A and B) in recovered typhoid and paratyphoid cases, and in those vaccinated against typhoid, which we may wish to present at a later time; they apparently show certain interesting group reactions. Of the controls giving no history of typhoid fever, 38 out of 44, or 86 per cent., have been clearly negative. Only five of these control cases (11 per cent.) reacted distinctly positively, and we believe may be explained as cases of aborted or mild, undiagnosed typhoid fever. This explanation is rendered prob- able by the observation, brought to our attention by Dr. Edward von Adelung, of a perfectly controlled clinical experiment on this subject. The observer, Dr. von Adelung, was a member of a family party of four which visited Germany nineteen years ago. About two weeks after drinking at a suspected water source, two of the members came down with a fever which ran the typical course of typhoid and was so diagnosed. The other two members of the party at the same time had mild symptoms, lasting one and three days, respectively, and consisting of head- ache, fever, malaise, and a sensation of flushing, which they regarded as abortive typhoid. Neither of the latter two persons gave any other history of typhoid fever. When the skin reaction was applied to the two cases that had run the typical typhoid course and to the two that had passed through the ‘‘abortive’’ attacks, all gave positive results. This well-controlled, natural experiment would seem to prove that the supposedly normal cases that react to typhoidin may well be those that have had abortive or undiagnosed typhoid fever in the past. Our experience with the skin reaction in those who have been vaccinated with various typhoid vaccines, including our own, has shown us that the majority of those who have been treated within a year and a half or two years may be expected to give a positive reaction, provided always that the last injection has 44 HARVEY SOCIETY been given at least a month previously. After two years the reaction is less likely to be positive. This experience with the skin test would correspond very closely with what has been found clinically to be the usual duration of artificial typhoid immunity. We have felt justified, then, in recommending to students who have taken the typhoid vaccine previously that they should have a skin test applied subsequently at intervals and that if it turned out to be negative they should repeat the treatment. Since our last communication on the typhoidin test, we have been led to modify the preparations of the solutions employed. We found that some of the original preparation of typhoidin after four or five months failed to produce a positive reaction in cases in which we had been led to expect it. That this failure was due to a deterioration in the product was evidenced by subse- quent positive results obtained in six such cases by means of a new preparation. Although such deterioration does not occur for a considerable time with tuberculin, it is known to take place in mallein and abortin. In the latter two cases and in tuberculin as well, better reactions have been obtained with preparations purified by alcohol precipitation and dried (compare Haring * and Meyer and Hardenbergh **). A similar preparation has been made by precipitating the original typhoidin with twenty volumes of absolute alcohol, filtering, washing with absolute alcohol and ether, and then drying on porcelain plates over a vacuum with sulphuric acid. The control solution, 5 per cent. glycerin bouillon evaporated to one-tenth volume, was treated in asimilarmanner. Ten c.c. of the original typhoidin yielded 0.78 gm. of dried powder when the culture had been grown for five days before evaporating. The dried powder from the same amount of a control solution gave only 0.5 gm. There is every reason to believe that this dried typhoidin will keep its potency undiminished for at least a considerable time. Our observations, however, extend only to its trial over a period of two months. We dissolve a small amount in carbolated saline equivalent to the original volume of concentrated typhoidin or even to a double concentrated solution, and find that when kept in a cool IMMUNIZATION AGAINST TYPHOID FEVER 45 place it gives very good reaction in typhoid immunes for at least a month. Our suggestion and growing belief that this skin reaction with typhoidin is a real measure of the protection that the person enjoys against typhoid fever is strengthened by observations on our immunized rabbits. We have already shown that the agglu- tinin titer is no indication of the resistance of a given animal to infection, and observations on the Widal reaction in man tend to the same belief (Ruediger and Hulbert **). The typhoidin reac- tion, on the other hand, is positive in that category of persons who are known to be protected against typhoid fever, namely, typhoid recoveries; it does not occur in persons who give no his- tory of the disease, except in a small percentage that may reason- ably be suspected of having had an abortive attack. The reac- tion further occurs in the majority of those persons who have been vaccinated against typhoid within the last two years and then gradually disappears. We had scarcely hoped to show differences in typhoidin reaction between incompletely and per- fectly immunized rabbits as tested by our method of infection, which is, after all, a violent one, but our results in this respect have exceeded our anticipations. It is, however, easy to demon- strate fundamental differences between normal and immunized rabbits by the intradermal reactions. Cutaneous reactions were first tried, but abandoned as unsatisfactory. With the intradermal reaction, performed by producing a tiny bleb under the skin by a short needle carrying typhoidin, a distinctive reaction is produced in rabbits that have been arti- ficially immunized, but not in controls. In each case a patch on the abdominal surface is shaved the day before the test and two blebs produced under the epidermis by injecting in the one concentrated glycerin bouillon and in the other typhoidin. In the normal animal there is little or no difference between control and typhoidin spot. Both are usually surrounded by a red zone of a few millimetres, and there may be slight induration. In the animal that has previously been treated by typhoid vac- eines there is a sharp difference between the two spots. The typhoidin spot, as early as five hours, but with increasing regu- 46 HARVEY SOCIETY larity and intensity to twenty-four hours, becomes surrounded by a red areola, is indurated, and quickly forms a firm, hard nodule with a yellowish, slightly softened centre, from 2 to 5 mm. in diameter. This nodule persists at least for several days and may last two or three weeks. It is characterized histologically by an infiltration of lymphoid cells with which are mixed a few polymorphonuclear leucocytes. Five normal rabbits tested in this manner gave negative reactions. Of 35 more or less perfectly immunized rabbits, 26 (74 per cent.) reacted positively. The differences in reaction between animals effectively immunized against our method of producing carriers and those which are not so fully protected is a relative matter when viewed in the aggregate, but in individ- ual experiments is more striking. Thus, dividing our treated animals in respect to a negative or positive intradermal test and then comparing the results with their carrier condition or recovery on infecting the following day, we find that of nine with a negative intradermal test, seven, or 78 per cent., became carriers; of 24 with a positive intradermal test, 13, or 54 per cent., became carriers, a distinct though only relative indication of agreement between the test and the absolute protection. In Table 4 is given the comparative immunizing value of four different vaccine preparations. Before the test inoculation, the intradermal test was applied to these animals with the following results as compared with the eventual carrier conditions: Group 1, vaccinated with whole unsensitized vaccine (three animals) : intradermal test, 3 negative; carriers produced, 2; recovered, 1. Group 2, vaccinated with unsensitized sediment (five animals) : intra- dermal test, 4 positive; carriers produced, 3; recovered, 2. Group 3, vaccinated with whole sensitized vaccine (three animals) : intradermal test, 3 positive; acute death (anaphylaxis), 1; recovered, 2. Group 4, vaccinated with sensitized sediment (four animals): intra- dermal test, 4 positive; carriers produced, 1; recovered, 3. These results show a distinct relation between a positive intradermal test and recovery and between a negative intra- dermal test and establishment of the carrier state. Mechanism of the Typhoidin Reaction.—We do not intend, at IMMUNIZATION AGAINST TYPHOID FEVER 47 this time, to discuss fully the experimental evidence and the hypotheses that have been brought forth to explain the general- ized and localized reactions that follow the injection or appli- cation of bacterial extractives like typhoidin in infected or im- munized individuals. Of these reactions, the one to tuberculin has naturally been most studied. Some of the most interesting and debatable questions that seem to have arisen in connection with the tuberculin reaction are, first, whether or not it is in reality a reaction exactly similar to the anaphylactic reaction with serum, and, second, whether it is due to an interaction of antigen and antibody accompanied, it may be, by alexin fixation (Wassermann and Bruck). Levy, Blumenthal and Marxer: Cent. f. Bakt. (Orig.), 1906, xlii, 265; 1908, xlvii, 289. %® Von Ruck, K.: Med. Record, 1912. * Deycke, G., Much, H.: Beit. z. klin. c. Tub., 1910, xv, 277. * Wells, H. G., and Corper, H. J.: Journ. Infect. Dis., 1912, xi, 388. * Kendall, A. I., Walker, A. W., and Day, A. A.: Journ. Infect. Dis., 1914, xv, 443-471. “ Opie, E. L.: Journ. Exper. Med., 1908, x, 645. “ Manwaring, W. H.: Journ. Exper. Med., 1912, xv, 1; 1913, xvii, 1. * Roemer, P. H.: Deutsche med. Woch., 1914, xl, 533. 48 Much, H.: Deutsche med. Woch., 1914, xl, 554. “Manwaring and Bronfenbrenner: Journ. Exper. Med., 1913, xviii, 601. *Ruppel, W. G., and Rickman, W.: Zeits. f. Immunitatsforch (Orig.), 1910, vi, 344. “ Bartel, J., et. al.: Cent. f. Bakt. (Orig.), 1906, xl, 723; ibid, 1909, xlviii, 159. “Opie, E. L.: Journ. Exper. Med., 1909, xi, 686. “Manwaring: loc. cit. “ Kling, C.: Zeits. f. Immun., 1910, vii, 1. © Bergel, S.: Zeits. f. Tub., 1914, xxii, 343. 5 Webb, G. B., Williams, W. W., Basinger, A. F.: Trans. National Assoc. S. and P. Tub., 1910, Sixth Arnual Report, p. 279. % Lewis, P. A., and Margot, A. G.: Journ. Exp. Med., 1914, xix, 187. * Murphy, J. B., and Ellis, A. W. M.: Journ. Exp. Med., 1914, xx, 397. “ Hamburger, F.: Beitriig. z. klinik. d. Tuber., 1909, xii, 259. 5 Roemer, P. H.: Beitriig. z. klinik. d. Tuber., 1910, xvii. 5 Rabinowitsch, L.: Deutsch. med. Woch., 1913, xxxix, 105. % Orth, J.: Berlin. kl. Woch., 1913, 1, 429. * Roemer, P.: Beit. z. klin. d. Tub., 1910, xvii, 287. ® Smith, Theobald: Harvey Lectures, 1906, i, 272. © Bruck and Steinberg: Zeitsch. f. Hyg., 1912, Ixxi, 177. * Hillenberg: Zeitsch. f. Hyg., 1914, lxxvii, 101. THE MORE RECENT DEVELOPMENTS IN THE STUDY OF ANAPHYLACTIC PHENOMENA * PROFESSOR HANS ZINSSER Columbia University I T is a fundamental biological truth that the systematic treat- ment of an animal with a foreign protein, if this is adminis- tered by any route other than that of the alimentary canal, induces profound physiological changes. These changes are primarily recognizable by the appearance in the circulating blood of substances which superficially react with the injected protein. For convenience of discussion we speak of these reac- tion products as antibodies, and of the injected substances, which possess this power of inducing their formation, as antigens. Antigens, then, are all substances which, injected into the animal body, induce specific antibody formation. They form a large group in nature and are chemically proteins; indeed, we may say that all known proteins may act as antigens. Whether or not this term may also include lipoid-protein combinations, lipoids or the higher protein derivatives is as yet uncertain and need not in the present connection concern us. We may divide antigenic substances into two main classes. One of these comprises all of those substances of bacterial, animal or vegetable origin which, injected into the animal body, give rise to specific neutralizing or antttoxic properties in the blood of the injected animal. These are the bacterial exotoxins, the snake venoms, some powerful vegetable poisons and pro- teolytic and other enzymes of animals and plants. They are all substances which are powerfully active—some of them strongly toxie to the living animal, others true enzymes or ferments. Indeed, all of them possess properties which at least suggest our placing them into the class of enzymes in general. The number * Delivered January 30, 1915. 12 177 178 HARVEY SOCIETY of such substances known is limited. The reaction they call forth in the animal body seems aimed directly at the specific neu- tralization of their respective activities, and is so unique and different from that induced by other antigens that it would be convenient had we another term like ‘‘antitoxinogen’’ to set them apart by themselves. The other class of antigens comprises all proteins which are inactive, showing in themselves neither toxic nor enzyme-like properties. Introduced into the animal body parenterally, they call forth a response of a nature entirely unlike that of the anti- toxins, and which as far as we can fathom its purpose seems aimed merely at the assimilation or the removal of the injected substance. For the cells of the animal cannot utilize the foreign protein as such, and thus it is only foreign proteins injected into an animal that act antigenically and no antibodies are formed when homologous material is injected. This large group composed of all formed and unformed substances in nature in which a protein structure is involved, does not induce the formation of anything like the neutralizing antitoxins spoken of above. The antibodies appearing in animals treated with such substances have been spoken of as cytolysins or cytotoxins—precipitins—and in the case of formed antigens like bacteria or blood-cells—agglutinins and opsonins. It is our opinion that all these various antibodies are identical in struc- ture and significance. The probable identity of agglutinins and precipitins was suggested long ago by Paltauf, and the identity of precipitins with the antibodies which sensitize for- eign proteins to the action of alexin or complement has been rendered more probable, we believe, by our own experiments. The terms agglutinin—lysin—precipitin and opsonin are all descriptive of effects produced when an antigen meets its specific antibody. These effects will differ according to the physical condition of the antigen. We believe that it is most likely, both from a study of the work of others and our own experiments aimed at this point directly, that the visible agglutination or precipitation are secondary effects incidental to the colloidal nature of the reacting bodies and to the quantitative proportions ANAPHYLACTIC PHENOMENA 179 in which the reactions occur, the essential process being the union of antigen and antibody, by which the former is rendered amenable to the action of complement (alexin) or leucocyte, as the case may be. It is not necessary, at any rate, to assume that functionally there is more than one variety of antibody, this one being the specific sensitizer. However this may be, the definite fact remains that the injection of antigens of this second class into animals induces specific reaction bodies or antibodies in the plasma of the treated animal which can be shown to unite with the homologous antigen in vitro, and which probably do so in the body of the animal when the antigen is reinjected into a subject in which antibody formation has taken place. We must not forget, however, that the observation of anti- bodies in the circulating blood is but one of the changes that have taken place in the treated animal. Much has been made of this phase of the problem because serum antibodies are readily studied in vitro; but their origin of course should be sought in the body cell, in which the original and most profound changes must necessarily have taken place during such treatment, changes the nature of which are to a large extent still a mys- tery, but on which ultimately depend the important physiologi- eal differences between treated and untreated animals. For such changes—whether we refer to those immediately under discussion, namely, those of allergy or anaphylaxis, or whether we think of the so-called immunity remaining after attacks of many diseases—remain present long after the circulating anti- bodies have disappeared and must therefore be regarded as associated with profound alterations in the ultimate tissue unit, the body cell. Pasteur’s observation that animals systematically treated with sublethal doses of bacteria became specifically more resist- ant to subsequent infection, carries in it all the principles of the process of which we speak as ‘‘active immunization,’’ and all the modifications and adaptations to special cases which we now recognize are based on this simple truth. The successful transference of such increased powers of resistance to normal animals with the serum of the immunized individual, by Behring 180 HARVEY SOCIETY and his collaborators, gave us ‘‘passive immunization,’’ and these two discoveries are the pillars on which all our compli- cated subsequent development of details has rested. Since with bacteria and their poisons the process implied the protection of the body against disease or death, we have, rather unfor- tunately, come to speak of these procedures as ‘‘immunization,’’ although the reactions of the animal body to injections of bac- teria, reactions on which incidentally the protection depends, are in principle identical with similar reactions resulting from the injection of entirely innocuous substances, such as egg albu- min, blood-serum or blood-corpuscles. It is, therefore, mislead- ing when we speak of the immunization of an animal to, for instance, sheep cells or horse serum. A physiological change takes place in such animals entirely analogous with that which occurs in those receiving bacterial protein, but the substances injected are in the former entirely harmless; and indeed, as we shall see, the animal, while entirely immune to large quanti- ties at the first injection, may be severely injured or even killed by subsequent administrations of the same substance. Thus the animal most ‘‘immune’’ to horse serum is the one that has never received an injection of horse serum. It is necessary, therefore, to emancipate ourselves from the misleading elements in the habitual terminology so that we may avoid confusion in grasping underlying principles. The essential feature common to all antigen injections, there- fore, is that of specific antibody formation. That their produc- tion in the case of living or dead bacteria—harmful in them- selves—protects the animal from invasion and prevents develop- ment and multiplication of the organisms once admitted, though of the greatest practical importance, is purely incidental. It is not impossible that the physiological reaction, indicated among other things by the circulating antibodies, denotes a mechanism aimed at the more effective assimilation and elimina- tion of the body-foreign antigens that have been injected, and this, in the ease of the bacterial cell, which of course represents a foreign protein, has the effect of protection against invasion. This point of view of the significance of antibodies is the so-called ANAPHYLACTIC PHENOMENA 181 theory of ‘‘parenteral digestion’’ of which we will have more to say directly. We must remember at any rate that in all cases in which, clinically or experimentally, we are confronted with the presence of foreign antigens in the blood and tissues, we are dealing with abnormal conditions in which the mechanism avail- able under normal circumstances for the disposal of foreign proteins, which may gain entrance accidentally in extremely minute quantities, is under a strain and abnormally active. The extreme quantitative increase of the antibodies is alone suffi- cient testimony for this, and under the special conditions which we are about to discuss the repeated introduction of such anti- gens into the body of an animal in which specific reaction bodies have been induced, whether these are freely circulating or still parts of the cells which produced them, may have illness or even death as a consequence. This is anaphylaxis. To approach this subject logically without allowing secondary factors to divert our attention from the fundamental principles involved, we should not limit this term to any arbitrarily stated train of symptoms, nor should we attempt too rigidly to limit the definition of what we call anaphylaxis. Indeed, from the point of view of human pathology, it is of quite as much, if not more, importance to study the effects of slow and slight injuries of this category, than it is to observe them only in the extreme and stormy manifestations of acute anaphylactic death. The former are the types of reactions occurring in the ordinary inci- dents of life. The latter are extreme results of experimental procedures and are for this reason of course more likely to reveal the underlying principles. But it would lead to false logic in our deductions were we to mistake a difference in degree for a difference in principle. In the light of our present understanding, therefore, we should broadly define the term as the injury, acute or slow, severe or slight, which under manifold circumstances may follow on the meeting of an antigen with its specifie antibody within the animal body. When such injury fails to result in the case of the spontaneous entrance or the experimental injection of bacteria into an immunized subject it is probably because the 182 HARVEY SOCIETY organisms are disposed of before the amount of foreign protein is sufficient to permit such a harmful reaction. What these circumstances are is the problem before us. In the case of innocuous foreign proteins such as blood serum or cells incapa- ble of multiplication, it is doubtful whether immunity—that is, ability to escape harm on reinjection—ever exists. However, we do know that the animal may be non-sensitive, as on first injection when practically no specific antibodies are present, or it may be hypersusceptible, anaphylactic or (the most com- prehensive term) allergic. II We may discuss briefly the conditions under which so-called anaphylactic shock may be experimentally elicited in animals. Although of relatively recent development, in their details, the observations which underlie the phenomena took root in the early history of serum investigation. Morgenroth? speaks of an observation by Magendie as early as in 1839 in which he describes the sudden death of dogs when repeatedly injected with egg albumin. Flexner reported similar deaths in rabbits repeatedly injected with dog serum, Richet and Hericourt ? in 1898 showed that toxic eel serum injected into dogs would kill at the second injection in far smaller doses than were necessary to kill at the first injection. Similar significance attaches to the work done by Portier and Richet on actinocongestine. Properly be- longing in this group of phenomena are the early observations on hypersensitiveness to toxins in repeatedly injected animals made by Behring and his collaborators. The problem was brought into particular prominence by the observations of Arthus* in 1903, who found that horse serum injected into rabbits at intervals of several days would eventually, in the latter injections, give rise to severe infiltration and cedema, and almost at the same time Theobald Smith noticed the great sus- ceptibility to horse serum acquired by guinea-pigs that had been used for diphtheria antitoxin standardization. Independently and with great clearness of vision von Pirquet* had made similar investigations on clinical material, and in his work on ANAPHYLACTIC PHENOMENA 183 serum sickness appears to have grasped the fundamental signifi- cance of the phenomena with a thoroughness not shared by most of his contemporaries. The historical development of this subject and the experimental conditions under which hyper- susceptibility may appear were the subject of a paper read before this society some years ago by two of the pioneer workers in this subject, Rosenau and Anderson.» The fundamental facts concerning the anaphylactic reaction were worked out almost immediately under the observations of Theobald Smith and Arthus by these workers and by Otto® in Germany. I may be permitted to summarize this early work and the fundamental principles of anaphylaxis very briefly in order that we may not spend our time in detailed consideration of facts entirely familiar to most of us. It is now certain that hypersusceptibility may be produced in human beings, in guinea pigs, in rabbits, in dogs, in sheep and probably in all mammals, if we were to investigate them carefully. The condition may be produced by treatment with any of the sub- stances known to us which have the property of antibody production; in other words, with all substances in nature of which we speak as antigens. The condition is like other antigen-antibody reactions, specific within the limits of specificity recognized for all such reactions. It is certain that in so-called active sensitization, hypersusceptibility develops only after lapse of a definite interval, and this interval depends to a certain extent on the amount administered at the first injection. An animal once sensitized if not reinjected may remain sensitive for a long period; its sensitiveness will disappear immediately after recovery from a non-fatal reinjection or the animal may temporarily be desensi- tized by reinjection of the antigen at a period before hypersusceptibility has developed. Of the greatest theoretical importance, furthermore, is the fact that a normal animal may be rendered sensitive passively, by the injection of blood-serum from an actively sensitized animal, or by the blood-serum of any animal which has been once or repeatedly injected with the antigen; and according to Doerr and Russ and others there is a definite parallelism between the capacity of a serum passively to sensitive an animal, and its contents in specific antibodies. There are many other facts which are of importance, but which for the present we will neglect, since these are the funda- 184 HARVEY SOCIETY mental phenomena on which we may build our discussion. We may also dismiss very briefly the earlier theories of anaphylaxis, like those of Gay and Southard’ and Besredka,*? in which attempts were made to show that the substance which sensitizes is not identical with that which is responsible for the develop- ment of shock in the reinjected animal. We may, indeed, disre- gard as premature theories all those in which the anaphylactic reaction is removed from the sphere of true antigen-antibody reaction. Indeed, von Pirquet and Rosenau and Anderson from the beginning regarded anaphylaxis as the result of the reaction between the reinjected antigen and the antibody formed in response to the first administration; and indeed, this is the essential premise of the still earlier view of Vaughan. We may accept it at present, identifying the anaphylactic antigen with antigens in general, and the anaphylactic antibody with the protein antibody, not distinguishing for this purpose between agglutinins, precipitins, or cytolysins. The symptoms which follow on the reinjection of antigen into sensitive animals may show a wide range of variation according to the degree of sensitiveness and amounts injected. In acute anaphylaxis of guinea pigs, which as you know has been the most thoroughly studied, there is a rapid and severe death which may not occupy more than a fraction of a minute or at most five to ten minutes. The animals repeatedly show restlessness, cough, pass urine and feces, develop severe dyspnea, with infrequent respiration in which there seems to be almost complete immobilization of the chest wall and in which finally only shallow, irregular, spasmodic efforts take place. This, as Auer and Lewis have shown, is due to tetanic con- traction of the small bronchioles, with occlusion of the air passages, practi- cally no air entering the lungs. As the dyspnea develops, there may be at the same time spasmodic twitching of the limbs, retraction of the head and general convulsions. When for some reason or other the reaction is not so severe the animal may show merely general signs of illness, ruffling of the fur, twitching and restlessness, with respiratory difficulty of varying degree, coughing, and evacuation of urine and feces. In rabbits the symptoms are often less rapid in development, but in general principles are similar; in rabbits there is more frequently in the moderate cases a gradual muscular weakness in which the animal lies flat on the ground unable to support itself on its ANAPHYLACTIC PHENOMENA 185 legs, a condition which may proceed for long periods. Death is largely respiratory and the heart may continue to beat for a long time after respiration has completely stopped. There is a sinking of blood-pressure and a depression of temperature. The coagulation time of the blood is lengthened, there is apparently a depression of the leucocytes, and according to a number of investigators, who have been recently confirmed by Behring, there is a disappearance of blood platelets and an increased flow of lymph. Pathologically in an animal dead of anaphylaxis there may be petechial hemorrhages, according to Gay and Southard, in the heart muscle, pleura and intestinal wall and there may be fatty degeneration of the vascular endothelium. In guinea pigs especially there is a marked emphysematous dilatation of the lungs which is very constant, although according to Doerr it is not absolutely characteristic of this condition. Apart from the anatomical changes following acute anaphylaxis, frequently repeated injections of small doses of horse serum or egg white in dogs, cats, rabbits and guinea pigs have been shown by Longcope to produce cell injury in various organs, especially in the liver, myocardium and kidneys. The sudden onset, the nature of the reaction in the animal and the pathological lesions seem to indicate that the injury as occurring in anaphylaxis is due toa poison. It appears, then, that an animal is sensitive to a protein at certain stages at which specific antibodies to the sensitizing protein have been formed, and that under special circumstances the meeting of antigen with antibody within the animal results in a reaction in conse- quence of which the poisonous substance is liberated. This being the logical point of view on the basis of available knowledge, it was quite natural that many investigators were attracted by the theory of parenteral digestion. The curious changes in the coagulation of the blood during the ana- phylaxis have led to an interesting and important theoretical conception, namely, that the meeting of antigen and antibody may not, as otherwise believed, lead directly to the formation of a poison, but that in some way the results of such a union may influence the coagulation processes and that these alterations are the direct cause of shock. The first to give serious attention to such a train of reasoning was probably Nolf, and Doerr has recently called attention to the work of a number of investi- gators (recently confirmed by Moldevan) who observed that freshly de- fibrinated blood, i.e., blood in which the normal coagulation has been interrupted, may be toxic even when reinjected into the same animal. 186 HARVEY SOCIETY The same is true of serum taken from rapidly defibrinated blood. There is at least a possibility, then, that the anaphylactic injury is the result of an alteration in the blood indirectly brought by the union of antigen and antibody. However, the premises for such reasoning are still very vague, and moreover, any view which introduces the various elements which participate in blood coaguiating processes can have no part in such manifestations as those observed on isolated and washed tissues, as in the experiments of Schultz and Dale. III It is one of the earliest premises of Pfeiffer’s conception of bacteriolysis that the cell-dissolving action of immune serum liberates a preformed poisonous substance or endotoxin from the bacterial cell. It may be remembered that early in the his- tory of such researches Pfeiffer and some of his pupils showed that an immune animal could be killed more quickly by large doses of dead bacteria than could a normal animal, an experi- ment from which the conclusion was drawn that the more rapid bacteriolysis in the immunized animal resulted in a more rapid liberation of the endocellular poisons. This point of view has been many times brought forward, and of recent years most clearly by Wolf-EHisner. It is also a point of view represented by the theory of Nicolle, who similarly tried to explain anaphylaxis by the liberation of poisonous substances from the antigen through the action of the cytolysins or ‘‘albuminolysins.’’ As the investigation of antibody formation against foreign proteins of inherently harmless nature progressed, the belief gained strength that antibodies in facilitating the chemical dis- integration of the injected foreign protein represented a sort of emergency apparatus for parenteral digestion and consequent assimilation. Throughout the development of Metchnikoff’s ideas of immunity it is plain that he had tended toward such an interpretation, looking on the process of phagocytosis as a method of facilitating the removal of undissolved foreign sub- stances from the tissues and blood, while the serum antibodies were conceived as more particularly concerned with the un- formed foreign proteins which in the accidents of ordinary life ANAPHYLACTIC PHENOMENA 187 gained entrance. The most clear and thorough exposition of such a point is that which since 1907 has been carefully worked out by Vaughan, and to him belongs the credit for the develop- ment of many of the ideas underlying prevalent opinions on anaphylaxis. Vaughan, as you well know, subjected many different proteins, bacterial and others, to hydrolytic cleavage in absolute alcohol containing 2 per cent. of hydroxid. The protein is covered in flasks with 25 to 30 times its weight of this alkaline alcohol and the mixture boiled at 78° C. for an hour or more. In this way he has succeeded in splitting off from a large number of different proteins the toxic fraction. Since Professor Vaughan® himself has but recently em- bodied his views in a concise treatise, it is quite unnecessary to go into it more than to review briefly his views. He believes that all true proteins contain a poisonous group which is practi- eally the same in all of them. This poison can become free and active when proteins are submitted to various methods of decom- position. Protein sensitization, in other words, is due to the fact that there is developed after the first injection a specific proteolytic ferment, and this on second injection so acts on the reinjected antigen that the toxin fraction is set free and poison- ing results. This, in brief, is Vaughan’s point of view and is supported, first, by the fact that such poisons can be formed by his chemical methods from many different kinds of protein ; and second, that these poisons, whatever the antigen from which they are derived, may produce symptoms which are in many ways identical with those characteristic of anaphylactic shock. Since, as Vaughan states, proteolysis consists in a gradual break- ing up of the protein molecule into simpler and simpler groups, there is an increase of poison liberation up to a certain point in the process; but when it has proceeded beyond this the poison itself is decomposed and ceases to have toxic action. Vaughan believes that anaphylaxis in all its manifestations, whether acute or chronic, is merely an incident in parenteral protein digestion. In the course of this when the relation between circulating an- tigen and the specific enzyme is such that large amounts of the toxic fraction are suddenly liberated, acute shock follows. 188 HARVEY SOCIETY It is hardly necessary to call attention to the attractiveness of such a theory, which so simply explains the apparently mys- terious conditions prevailing in anaphylaxis, and there seemed to be very little doubt as to its correctness when Friedemann *° some years later showed that the action of fresh unheated serum (i.e., alexin or complement) on sensitized red blood-cells will produce a poison that, injected into rabbits, gives rise to anaphy- ‘laxis-like shock. Following him Friedberger** succeeded in producing a similar poison by allowing fresh guinea-pig serum (i.e., complement) to act on both precipitates formed by the union of the serum with its antiserum and on sensitized and unsensitized bacteria. These investigations clearly suggested that the action of the alexin present in the circulating blood, on an antigen sensitized with its specific antibody, might produce protein cleavage in which there was liberated a toxic fraction similar to that produced by Vaughan with his chemical hydro- lytic methods. It is but natural, therefore, that Friedberger, to whom the greatest credit in the more recent development of this point of view belongs, should assume that the poison liber- ated in this way is the toxic factor concerned in anaphylaxis, and name it ‘‘anaphylatoxin.’’ For reasons which will appear directly, we think that a preferable term would be ‘‘proteo- toxins.”’ The technic developed by Friedberger consists, in the case of dissolved proteins, in allowing the antiserum to act on the serum until a precipitate is formed, then subjecting this precipitate to the action of fresh guinea pig serum or complement. After a variable number of hours, the length of time depending on secondary factors, which need not be discussed in describing the process, the centrifugation removes the precipitate, the supernatant guinea pig serum is found to be strongly poisonous, and injected into guinea pigs intravenously in quantities of from 2 to 4 c.c, produces symptoms typical of acute anaphylaxis. With bacteria his technic is similar. At first bacteria sensitized with inactive immune serum were subjected to the action of fresh guinea pig complement for from one to two hours at 37° C. to as long as twelve to twenty-four hours at refrigera- tor temperature. At the end of this time the bacteria is removed by rapid centrifugation, and the supernatant fluid injected into guinea pigs pro- duces again typical symptoms of acute anaphylaxis. ANAPHYLACTIC PHENOMENA 189 The first interpretation applied to these experiments by Friedberger was an entirely natural one if we consider the general views held before this concerning bacteriolytic and bactericidal processes. He assumed that the complement acting on the sensitized bacteria or on the sensitized protein in the precipitate experiment (or later on the unsensitized bacteria), produced proteolytic changes in the course of which the toxic split product was formed. It seemed that the poison was pharmacologically the same whatever the antigen used, and ex- periments also seemed to show that the poison could be produced more rapidly from strongly sensitized than from unsensitized bacteria, and that an excess of sensitization or a too prolonged interaction resulted in non-toxic supernatant fiuids, which was taken to indicate that the protein had been split beyond the toxic stage by too energetic hemolytic action. Here, then, we have a simple and apparently logical explana- tion of anaphylaxis, entirely in accord with Vaughan’s views of parenteral digestion. An antigen is injected into an animal, specific antibodies and enzymes against it develop in the animal; reinjection of this antigen results in relatively rapid proteo- lysis in the course of which poisonous substances, the anaphyla- toxis, are produced and anaphylaxis is the result. This hy- pothesis, although very attractive, does not entirely meet with the facts as they have been developed since Friedberger’s first work. The premises on which it is based assume in the first place that the poison or ‘‘anaphylatoxin’’ is formed out of the matrix of the antigen; further, it is definitely assumed that in the pro- duction of the poison after the antigen and antibody have met, the complement or alexin plays an active part. Friedberger’s hypothesis as stated by him, moreover, seems to assume that the entire process takes place intravascularly, a matter which we will discuss at considerable length in a short time. It is important to note also that Friedberger, with Nathan, was able to show that this anaphylatoxin production could take place within the animal body; that is, within the peritoneum of a guinea pig into which bacteria had been injected. The simplicity of Friedberger’s explanation and the correct- 190 HARVEY SOCIETY ness of his experimental data soon persuaded many investi- gators that, in essence, his hypothesis probably contained the nucleus of the solution of this difficult problem. However, even his own early experiments aroused some misgivings concerning the matrix of the poisons produced, for he found that the poisons could be obtained as well when boiled antigen was used as when the fresh, unheated substances were employed, and the poisons were easily obtained from such organisms as the tubercle bacillus, which is extremely insoluble and unamenable to serum influence. It was also doubted whether one could truly assume the participation of this specific antibody or sensitizer in the production of Friedberger’s poisons, since it soon developed that from bacteria, at least, the poison could be produced when the organisms were directly exposed to the action of fresh guinea- pig serum without the presence of any immune serum. Experiments which soon threw a definite doubt on the assumption that the poisons were produced by a decomposition of the antigen were reported by Keysser and Wassermann.’? These workers substituted insoluble substances like barium sul- phate and kaolin for the antigen; that is, the precipitates or bacteria used in Friedberger’s experiments. They found that if kaolin were treated with horse serum and then exposed to the action of guinea-pig serum or complement, poisons were pro- duced identical in every respect to those produced by Fried- berger’s method. The conelusions they drew were that the poisons were produced, not by action of the complement on the antigen, but by its action on the horse serum absorbed by the kaolin. In other words, they transferred the matrix of the poison from the antigen to constituents in the serum itself, possi- bly the sensitizer or amboceptor. Bordet** also was able to show that poisons similar to those of Friedberger could be produced by the action of fresh guinea-pig serum on agar, and recently Bordet has further shown that this is the case even when the agar has been by special methods deprived entirely of its nitrogenous components and represents simply a complex of carbohydrates. Agar-guinea-pig serum mixtures of this kind show an increase in total non-protein nitrogen which would ANAPHYLACTIC PHENOMENA 191 prove that the proteolytic action of the guinea-pig serum must have been active against its own proteins. An interesting further development of this work has recently appeared in the experiments of Jobling and Peterson.1* They showed that when bacteria are mixed with fresh active serum they absorb the antienzymes normally present in blood. They have shown this experimentally and have proved that similar antienzyme removal can be accomplished by the addition of kaolin or agar, and by treatment with chloroform. Serums so treated become toxic, the actions of the poisons formed showing great similarity to that produced by Friedberger’s anaphyla- toxins. According to them, the poisons are formed because of the fact that antienzymes are absorbed by the antigen, thus setting the normal ferments in the fresh serum free to act on their own serum protein. It should be recalled that Friedemann, who was really the first one to show that the toxic substances could result from the interaction of fresh serum and sensitized antigens, although he succeeded only in doing this with red blood-cells, suggested rather early that the success of such an experiment does not necessarily mean that the antigen furnishes the matrix entirely. He had studied the metabolism in anaphylactic poisoning and with Isaac has shown that the nitrogen output following rein- jection in a sensitized animal is far in excess of that which could be derived solely from the injected antigen, and in this he has been confirmed by many other workers, notably by Vaughan. It would seem to us our present knowledge of this phase of anaphylactic investigation permits us only to conclude that wherever proteolytic changes take place these ‘‘proteotoxins”’ may be formed, That they can be produced from a protein antigen has been shown beyond doubt by Vaughan and his col- laborators for both formed and unformed antigens. Also this is evident from the experiments of many workers and has been confirmed in our own experience with poisons appearing during the autolysis of bacterial emulsions. On the other hand, it is also clear that the antigen need not represent the matrix which furnishes the poison, and that in the reactions as they are gener- 192 HARVEY SOCIETY ally performed both in the test tube and in the animal body, it is more than likely that if an antigen participates at all in furnishing the substratum for the poison, this is probably less important than that furnished by the animal’s own proteins. However, this does not weaken the importance of the knowledge | that the antigen-antibody reaction in the presence of normal serum and certain antigens in the presence of normal serum alone, induce a reaction in the course of which such poisons are formed. And the fact that they can be produced experi- mentally in the peritoneal cavity of a living guinea pig renders their participation in such reactions in the animal body a likely assumption. Our own work* on these substances induces us to believe that proteotoxins so formed are identical with Bail’s aggressins, a point to which we wiil refer later. Granted that such a poison, call it ‘‘proteotoxin’’ or ‘‘ana- phylatoxin’’ or ‘‘serotoxin,’’ as Jobling and Peterson have called it, is formed, it is important of course to determine as closely as possible its nature. Apparently the poison is the same as far as we can determine by pharmacological action when pro- duced by the chemical methods of Vaughan or by the biological methods of Friedberger and others. As obtained by Vaughan it is water-soluble with slightly acid reaction, is freely soluble in alcohol and mineral acids. It is not diffused readily and con- tains no carbohydrates. In its crude state it gives a biuret reaction, although this may mean simply that the poison has not been completely derived. The fact that the injection of Witte peptone into animals may give rise to symptoms very similar to anaphylaxis has been taken by many workers to signify that the anaphylactic intoxication is produced by a poison which is very similar to, or possibly identical with, the active constituents found in this peptone. After peptone injec- tion in normal animals there is a lowering of blood-pressure, a delay in the coagulation of blood and a development of subse- quent tolerance, together with many clinical symptoms which emphasize this similarity. Biedl and Kraus, who have especially studied this condition in dogs, have felt emphatically that the ANAPHYLACTIC PHENOMENA 193 anaphylactic poison is probably very similar to peptone. Re- cently Dale has suggested that B-iminazolylethylamin or his- tamin may be the active principle concerned in anaphylactic shock. Intravenous injection of 0.5 mg. of this substance into large guinea pigs results in typical respiratory difficulties, con- vulsions with death and distention of the lungs typical of ana- phylactic shock. Treatment with atropin diminishes this action, just as Auer and Lewis found this to be the case in true anaphy- laxis, and fall in blood-pressure also occurs. It would seem then that substances representing cleavage products of native proteins of highly complex nature, the result of proteolytic cleavage not very far advanced, are probably concerned in the production of anaphylactic shock. The anaphylatoxins of Friedberger cannot of course be studied chemically by the methods to which Vaughan’s poisons are amenable. IV A further problem which has arisen in connection with the conception of parenteral digestion is that which concerns the participation of complement or alexin in the cleavage process during which the anaphylactic noxious agent is liberated. When bacteria or red blood-cells are sensitized, that is, have been combined with their specific antibodies, we have believed that it is the complement, or active constituents of fresh blood, which then acts on this sensitized complex, either producing hemolysis in the case of sensitized red blood-cells, or the bac- tericidal or bacteriolytic effect in the case of sensitized bacteria. It is also well known to you that this substance, which we call complement or alexin, but about the true nature of which we know nothing, is fixed or bound by dissolved proteins when they have combined, with or without the formation of precipitates by their antibodies. We have ourselves*® shown that such fixation of complement by precipitates (formed when an antigen and its precipitin have united) is bound in exactly the same way as this occurs in the ease of sensitized red blood-cells; that it is not a non-specific physical complement fixation such as that which occurs when complement is fixed by kaolin, yeast cells or 13 194 HARVEY SOCIETY other unsensitized emulsion. From this knowledge there has gradually grown the conception that the complement or alexin may be a necessary, active factor in the cleavage of the antigenic molecule. (This may or may not be so; we may say we think that we have no proof at present that the complement acts as a proteolytic enzyme; on the other hand, it is more than likely that in some way it is connected with such cleavage processes. ) At any rate, since we know that the anaphylactic reaction is the result of the union of an antigen with its antibody, and this together with our knowledge of complement fixation, naturally suggests that the complement may be directly concerned in the mechanism of anaphylaxis. The first method of approaching this problem naturally was the examination of animals with regard to quantitative changes in the complement contents of the blood during anaphylactic shock. It was found by Sleeswijk 17 that animals actively sensi- tized and reinjected showed a very definite diminution of com- plement. However, under such conditions the diminution was neither rapid nor very extreme, facts since confirmed by Fried- berger and Hartoch,!® who found the diminution very much greater in experiments with passive sensitization. In such cases there was a regular and considerable diminution, so that after shock four to eight times as much serum was necessary to pro- duce the alexic effect as before shock. Friedberger even be- lieved that there was a definite parallelism between the intensity of shock and the degree of complement diminution. The ques- tion immediately arises is the loss of complement, which we may now regard as a demonstrated fact, an incidental effect of shock or has it casual relationship to the development of shock? The latter seemed at first to be likely for a number of reasons. It was found, in the first place, that the addition of complement to the circulation of an animal during the anaphylactic experi- ment did not serve to prevent shock. Similar evidence seemed furnished by certain experiments on the complement of birds, by work of Loeffler ?® and by the observation of Hartoch,”° that but slight shock could be produced in guinea-pigs suffering from trypanosomiasis in which, as is well known, complement is ANAPHYLACTIC PHENOMENA 195 greatly reduced. Loeffler also attempted to support this point of view by sensitizing guinea-pigs and then reducing their complement by the injection of sensitized beef blood intraperi- toneally. Such animals showed diminution of reaction when reinjected with the sensitized antigen. Loeffler’s experiments are not conclusive, since the action of the sensitized blood-cells in the peritoneum must surely have induced an intoxication not at all unlike that taking place in true anaphylaxis, and, as we have shown recently together with Dr. Dwyer, such intoxications are followed by non-specific tolerance to the anaphylactic poison. However, another method of approaching this problem was attempted by Friedberger in his well-known salt experiment. It had been shown by a number of workers, among whom we may mention especially Nolf 74 and Hektoen,” that complement is not bound by sensitized complexes in the presence of hyper- tonic salt solution. In fact, hypertonicity seems to inactivate complement, and indeed it is a method of many laboratories to preserve complement for considerable periods by adding hyper- tonic salt solution, in which condition it will last a considerable time and is easily reactivated on dilution to isotonicity with distilled water. Friedberger ?* injected concentrated salt solu- tion into sensitized guinea pigs just before reinjection. It is possible, as he found and as we have found since, to inject 0.3 c.c. or even more of saturated salt solution intravenously into guinea pigs of about 200 grammes weight without killing them. When sensitized guinea pigs were injected in this way and immediately afterwards received a toxic antigen injection, shock was definitely diminished and death averted. This has been one of the strong- est bulwarks of those who have believed in the participation of complement in serum anaphylaxis. And it was assumed that the mechanism of the salt experiment consisted in a prevention of complement action. Recently doubt has been thrown on this because Ritz 24 has shown that salt injection not only prevents anaphylactic shock but will prevent the toxic effects of Witte peptone and of the so-called ‘‘anaphylatoxins.’’ Recently with Dr. Dwyer ?° we have carefully repeated this work and have 196 HARVEY SOCIETY found that when the dose is carefully adjusted there is no ques- tion about the fact that an immediately preceding injection of concentrated salt solution will prevent death or even symptoms in animals injected with proteotoxins. This tends very strongly to diminish the weight of Dr. Friedberger’s interpretation of the salt experiment; it means either that the salt in diminishing anaphylactic shock does so by a mechanism not concerned with the prevention of complement, or else it signifies that the so-called proteotoxin itself is not a finished poison as it has been thought to be but must still be acted on by the active constituents of serum before it becomes active.’ It is true, indeed, that heating serum to a temperature of 56° C. renders it impotent to lead to proteotoxin production when added to antigen in vitro and that this same inactivation destroys the complementary effect on sensitized red cells or bacteria. This, after all, does not prove identity of the sub- stances carrying those activities, but merely establishes an inter- esting parallelism. We must not forget that the substance of which we speak as ‘“alexin’’ or ‘‘complement’’ is not very well understood. We know little of its nature. It has been successfully shown that globulin participation will divide it into two parts, that it will spontaneously reactivate to a slight degree after heat inactiva- tion, that its activity is influenced by concentration, and that it can be inactivated by shaking. We are aware of the fact that we are here, possibly, dealing not with a single substance, but with one of the effects of a complex serum constituent. As to its relation to anaphylaxis we can only say that the diminu- tion of complement during anaphylaxis is perfectly definite. However, we cannot claim with certainty, in spite of the evidence so far advanced, that it plays an active part in the production of anaphylactic shock. 1 With Lieb and Dwyer (Proc. Soc. Exper. Biol. & Med., 1915, xii), we have been able to show that the hypertonicity produced by salt injections protects by decreasing the irritability of smooth muscle. ANAPHYLACTIC PHENOMENA 197 Vv The fact that the hypersusceptible condition can be trans- ferred from a treated to a perfectly normal animal with the blood-serum of the former, was in itself one of the first strong arguments in favor of the antigen-antibody conception of ana- phylaxis. And this point of view was still more clearly defined when Doerr and Russ** subsequently showed that the power of a serum to convey hypersusceptibility was directly propor- tionate to its contents of specific antibodies. A serum which was strongly precipitating for the antigen would passively sensi- tize in quantities far smaller than those necessary for the same purpose in the case of a weakly precipitating serum. The prin- ciple that anaphylaxis depended directly on the meeting of the antigen with its specific antibody has never been seriously ques- tioned since this time. However, from the very beginning of experimentation on passive sensitization it has seemed unlikely that the acute reaction, as seen especially in guinea pigs, could be attributed entirely to the meeting of these two elements in the blood-stream. It was observed by Nicolle, Otto, Friedemann, Gay and Southard and by many others since then, that sharp reactions can be produced with regularity only when a distinct interval was allowed to elapse between the administration of the sensitizing serum and the injection of the antigen. When the two are injected together, mixed or simultaneously, symp- toms may be and usually are entirely absent, whereas severe and unfailing shock results when the antigen injection is de- ferred from twelve to twenty-four hours after that of the sensi- tizing serum. According to Doerr and Russ the interval may be shortened to four hours, but if lessened beyond this, the reaction may fail to appear, or if present at all is weak and indistinct. This observation alone would tend to persuade us that mere contact within the blood-stream of antigen cannot account for the entire train of phenomena and suggests that the character- istic anaphylactic reaction takes place only after the injected antibody has become attached to the body cells in the same manner. The idea in itself is not new. Wassermann had first sug- 198 HARVEY SOCIETY gested it in an attempt to explain the peculiar hypersuscepti- bility to toxin possessed by some of Behring’s toxin-immunized animals. He assumed that in such animals the formation of antitoxins may indeed have been stimulated, but that much of it might still be attached to the generating cells themselves, thereby rendering these proportionately more vulnerable to the injected toxin. Such a conception of ‘‘sessile receptors’’ was applied by Friedberger *7 to anaphylaxis in his first attempts to formulate a hypothesis. He assumed that at the first or sensitizing injection the production of antibodies (precipitins) was stimulated. These, however, were not produced in great quantity and were not discharged into the circulation, possibly owing to the small single dose given for sensitization. They were present at the end of the anaphylactic incubation time as sessile receptors or sessile antibodies (precipitins). On the second injection a reaction occurred between the injection antigen and these sessile precipitins and the cell was injured because the reaction occurred on its substance, a reaction which, it is suggested, might have been harmless had it taken place in the blood-stream. In passive sensitization, conversely, no injury could result until consider- able quantities of the antibody had become united to body cells in the course of several hours. That the antibody injected into passively sensitized animals indeed disappears from the circu- lation with relative speed, has been shown by Doerr and again recently by Weil. Besredka’s § early hypothesis, too, though incorrect in most of its premises, assumed the necessity of the intravention of the body cell in anaphylaxis—an opinion here again largely based on the observed interval in passive sensitization; and the same idea occurs at about this time in the work of Doerr and Russ, who likewise conceived the process as taking place directly on the body cell. It is true as Doerr ** has pointed out in a recent summary of anaphylaxis, that these early hypotheses were for a time rele- gated to the background, yielding the prominent central position te opinions which held that anaphylactic shock was the result ANAPHYLACTIC PHENOMENA 199 of intravascular parenteral digestion. To some degree this is due to the fact that Vaughan’s work on the toxic protein split products and Friedemann and Friedberger’s experiments on the production of similar poisons by purely biological methods, seemed to offer for the time being a field of work promising logical solution of this difficult problem. At the same time there was much evidence in the published work of such investi- gators as Friedemann, Scott, Briot, Biedl and Kraus, and Doerr himself, which seemed to show clearly that the interval in passive sensitization was not an invariable necessity. Conse- quently and very naturally the early purely cellular conceptions were not accepted as telling the whole story, and a few observers allowed the pendulum to swing completely away from this point of view. Nevertheless it is not fair to say that during this time the cellular theories were entirely neglected. We do not believe that von Pirquet ever entirely abandoned his original opinion that there was involved in certain phases of anaphylaxis an ‘fallergie’’ of the tissues. Moreover, it was during this period that those methods of research were first applied to anaphylaxis which furnished in principle and fact all the important premises for the present almost universal cellular point of view. I refer to the transfusion method as employed in anaphylactic dogs by Pearce and Eisenbrey *° and the method of observing isolated _ tissues from anaphylactic animals as used by Schultz *°—work which appeared as early as 1910. Pearce and Eisenbrey work- ing with two normal and one sensitized dog, transfused the blood of one of the normal animals into the sensitized one, trans- ferring the blood of the latter to the normal dog. ‘‘ At the proper moment the normal dog containing the blood of the sensitized animal and the latter containing the blood of the normal dog, each received intravenously the toxic dose of horse serum.”’ The normal dog having the sensitized blood did not react, the sensitized dog having the normal blood showed typical fall of blood-pressure. Pearce and Eisenbrey drew the conclusion ‘‘that the phenomenon of anaphylaxis is due to a reaction in the fixed cells and not either primarily or secondarily in the blood.”’ In the same year Schultz began to work with what is now 200 HARVEY SOCIETY spoken of as the physiological method. He determined that smooth muscle—freshly excised from various animals—will react with contraction when brought into contact with serum. When such muscle was taken from anaphylactic animals after being thoroughly washed free of blood, it would react more energeti- cally and to smaller amounts of the homologous serum, There are many interesting by-products of Schultz’s work, such as the differences between fresh arterial blood and blood-serum in their abilities to stimulate contraction, but this and other points will not be discussed at present. The important and incontrovertible fact established by Schultz is the changed reac- tion-energy or, in truth, ‘‘allergie’’ of the smooth muscle of ana- phylactic animals to the stimulus of the sensitizing antigen. Dale ** has confirmed and extended these observations of Schultz. He removed the uteri from guinea pigs after thoroughly per- fusing them with Ringer’s solution to remove all blood. He then suspended them in baths of Ringer’s solution and by the customary physiological methods measured the contractions fol- lowing the addition of various amounts of foreign protein in the form of—among other things—horse serum and beef serum. He found that the uterus of an animal sensitized to horse serum would react to this substance in dilutions of 1 to 2000 or even 10,000, while the organ taken from a normal guinea pig reached its limit of reaction ability at dilutions often less than 1 to 200. A uterus that had reacted strongly was found to be subsequently desensitized. A normal uterus could not—strangely—be pass- ively sensitized by immersion into a solution containing serum antibodies. This method of investigation has recently, also, been taken up by Richard Weil ** who has fully confirmed the principles laid down by Schultz and Dale. He has incidentally also answered an objection to the conclusions of Dale and Schultz (never indeed a very valid objection), namely, that the reaction of the muscle tissue of a sensitized animal might be in part due to the fact that the blood, 7.e., the antibodies, had not been entirely washed out of the tissue spaces by perfusion. Weil performed the very simple and ingenious experiment of inject- ing a normal guinea pig with large amounts of immune serum ANAPHYLACTIC PHENOMENA 201 (antihorse serum) and after a few minutes killing the animal. He then suspended the uterus in Ringer’s solution in the usual manner without washing it completely free of blood. Contact with the homologous antigen produced no response. We may accept as definitely established by these researches of Schultz, Dale and Weil that the fixed cells of anaphylactic animals possess an increased reactionability toward the antigen which is in no sense secondary to processes involving the circulating anti- bodies. Moreover, the work of Weil seems to indicate that desen- sitization of a passively prepared guinea pig deprives the uterus of its power to respond and that the gradual spontaneous diminu- tion of hypersusceptibility on the part of the guinea pig is accompanied by an entirely parallel loss of reaction-capacity on the part of the isolated uterus. The recent work of Coca,** too, has further fortified the cellu- lar point of view by a method which in principle is similar to that employed by Pearce and Hisenbrey. Coca succeeded in per- fusing activity and passively sensitized guinea pigs with the defibrinated blood of normal guinea pigs in such a way that the original blood of the sensitized animals was reduced to a necessarily slight residue. Animals so treated could be kept alive for as long as six hours after the transfusions and remained delicately hypersusceptible in spite of the blood substitution. Limiting ourselves for the present to the phenomena of ana- phylaxis in which non-cellular antigens are employed, we may safely say that the evidence furnished by the incubation time necessary in passive anaphylaxis by the transfusion experiments of Pearce and Eisenbrey and of Coca, and most conclusively by the work on isolated tissue of Schultz, of Dale and of Weil, shows conclusively that the hypersusceptible state is largely determined by a changed reaction-capacity to the specifie antigen on the part of the fixed tissue cells—an ‘‘allergie’’ which is probably due to the presence of specific antibodies in the sub- stance of the cell protoplasm, and incidentally accounts for such effects as the skin reactions. It is probable that the acute symp- toms and death of anaphylactic guinea pigs (and indeed of other animals) is in most cases of experimental anaphylaxis 202 HARVEY SOCIETY due to the reaction which takes place between the injected antigen and these sessile receptors. So much we must logically accept. However, are we justified in denying all possibility of injury to the animal when antigen and antibody meet in the circulation? This is indeed the claim of a number of workers who are inclined to regard the presence of circulating antibodies not only as incapable of leading to injury, but in fact as a protection, in that the antigen is de- flected by them from the antibodies united with the cells. Per- sonally we believe that this radical cellular interpretation of all phases of the phenomena of anaphylaxis goes too far. It was shown by Friedemann as early as 1909 that typical anaphy- lactic reactions could be produced in rabbits when the antigen (beef serum) was injected simultaneously with or mixed with the serum of passively sensitized rabbits. Indeed Friedemann claimed that by this method severe and fatal reactions could be produced in rabbits more regularly than when an interval was observed. Richet in the same year reported experiments in which immediate symptoms were elicited in dogs injected with mixtures of crepitin and the serum of a crepitin-treated dog, the crepitin in quantities far below that necessary to elicit symptoms in itself. (In this experiment of Richet the crepitin and the serum were left in contact in vitro for 20 minutes, a fact which somewhat detracts from the direct bearing of this work on our present discussion. ) In 1910 Biedl and Kraus ** obtained immediate and severe symptoms in guinea pigs when they injected intravenously mix- tures of horse serum together with the serum of sensitized guinea pigs. Briot *° in the same year obtained reactions in young rabbits into which he had injected mixtures of horse serum and antihorse serum. Gurd ** in a recent publication obtained reac- tions in guinea pigs when he injected intravenously immune rab- bit serum (antisheep serum) and immediately thereafter sheep serum. We ourselves have been able to obtain occasional and distinet results in rabbits and guinea pigs both by simultaneous and immediately consecutive intravenous injections of antigen and antibody, though we did not succeed in attempts to dupli- - ANAPHYLACTIC PHENOMENA 203 cate exactly the experiments of Friedemann and of Biedl and Kraus. We have here a not inconsiderable mass of evidence which points to the conclusion that the whole story of the anaphylactic phenomena cannot be told by the cellular conception alone, and that probably—as in immunity—both cellular and intravascular processes are involved. Few thoughtful workers on hypersus- ceptibility would think of denying at present the probably pre- dominating cellular factor in the ordinary anaphylactic type- experiment. I may say that many of us have never doubted that this element was an undeniably important one in serum anaphylaxis since the time when the experiments of Pearce and Hisenbrey and those of Schultz confirmed the suggestion of this conception forced on us by the incubation time in passive sensi- tization and the studies of von Pirquet. We do not share, how- ever, the exclusively cellular view recently advocated in a recent summary and apparently accepted by Doerr, one of the most capable students of this subject. It is true that almost all of the workers cited above as having obtained passive sensitization, without the interval, admit the irregularity of such results, and Friedemann, Gurd and others eall particular attention to the great importance of the relative amounts of antigen and antibody when these are injected to- gether or in rapid sequence. This has been our own experience and although we have obtained very definite reactions in this way, we feel that in any given experiment success or failure cannot be as regularly foretold as in the experiments in which the interval is allowed. Moreover, the reactions obtained by these methods are often mild—delayed—and are rarely violent or rapidly fatal. We ourselves have never obtained a fatal result. Yet it is idle to say—as has been said—that the reactions so obtained are accidents, probably due to secondary factors, negligible in formulating a conception of anaphylaxis. There is no such thing as an accident in nature, and the observation, though irregular and depending on elements in the experimental procedures not easily amenable to control, has been made too often and independently by a number of different trained 204 HARVEY SOCIETY observers to be thrown out of consideration in a theoretical scheme which is to be just to all the facts. Since we cannot, therefore, deny that under certain circum- stances injury to the animal may result from the meeting of the antigen and antibody within the circulation, how are we going to account for the fact that such reactions are difficult to obtain and cannot be obtained with regularity? This ques- tion is not a simple one but it is our own opinion that a possible explanation may be found in the failure of rapid union of antigen and antibody in the blood-stream. We have already mentioned that all observers who have experimented along these lines have found that very definite proportions between antigen and antibody govern the success of such attempts and that with each lot of serum and antiserum the optimum proportions must be determined by experiment. In Friedemann’s work on rabbits he found that the relative amounts of antigen and antibody which produce reactions in his rabbits if injected together corre- sponded roughly to the proportions which in vitro gave precipi- tates. An excess of one or the other substance would prevent reaction or at least result in a negative experiment. Now it is well known to all who have worked with antiprotein serums that the precipitin reaction can be inhibited by an excess of one or the other reagent. When a constant amount of precipitating serum is used, the most prompt and voluminous precipitation may, for instance, occur when the antigen dilution is 1: 50, and both the speed and the amount of precipitate may diminish not only as this dilution is increased, but also as the concentration is increased. This is a phenomenon which is common to all colloidal reactions and the mutual precipitation of the two colloids is to a large extent dependent on relative proportions. It is a well-known fact (also familiar to many of you) that Linossier and Lemoine,** Eisenberg,?* Ascoli,*® von Dungern *° and others have frequently noticed that animals treated with a foreign protein such as horse serum, for instance, may contain in their blood-serum, as late as six, seven, eight or more days after injection, both the antigen and its antibody ununited ANAPHYLACTIC PHENOMENA 205 and separated. Thus we have often seen ourselves, if we bleed an animal that has been rapidly treated with such a foreign protein, that its serum will precipitate horse serum, and will at the same time be precipitated by antihorse serum taken from another rabbit. It is thus plain that the serum in the case mentioned contains not only horse serum as such (a remnant of that injected) but also antibodies against horse serum which have been formed in response to the injection. It is unques- tionable from the experiments of others and from our own extensive confirmations, that the serum of such an animal may contain side by side free antigen and free antibody. Why have these failed to unite? If such a serum is allowed to stand at room temperature or in the ice-box there will take place a very slow precipitation and a concomitant diminution in the amount of precipitin present. The precipitate thus formed has slight and distinct complement-fixing properties. Slow union, there- fore, is taking place. Another strange fact about such serums is that if two such rabbits are prepared, in each of which both free antigen and antibody can be determined, these serums when mixed will promptly precipitate each other. A number of explanations have been advanced for the simul- taneous presence of antigen and antibody in the same serum without union. Eisenberg and Volk have attempted to explain it by dissociation—that is the antigen and antibody are present united and also dissociated, reacting according to the laws of mass action. This has seemed to us unlikely. For, were this the case, the serum, as taken, should in itself exert definite com- plement-binding properties, since on the basis of this explanation it must contain not only the two reagents separate but a rather large proportion of the antigen-antibody complex united. This is not the case according to our own observations and according to similar ones made by Gay and Rusk. Von Dungern ** has assumed that the state of affairs de- scribed was due to the fact that the antigen might contain a number of different substances, alpha, beta, ete., each of which produces its own specific Tei-priizipitin. He believes it possible 206 HARVEY SOCIETY that at certain stages in the immunization the free antigen present might be, say, an alpha fraction, the free antibody, let us say, a beta precipitin, the two not fitting and therefore unable to react. This has not seemed likely to us although they are clear when taken and remain so for considerable periods, but do eventually precipitate slowly and in the course of days, an observation made not only by us but by Merckel. It has seemed to us most likely that there might be in the circulation of animals an inhibiting agent, somewhat in the nature of a protective colloid, which prevented the union of antigen and antibody, or at least tended to make it an extremely slow process. We may assume in the light of our present knowledge that both the antigen and the antibody are colloidal in nature, and together with Stuart W. Young, we *? have been able to produce an analogy to the condition found in the serums just described by using three colloidal suspensions, that is, arsenic trisulphid, gelatin and gum arabic. Emulsions of gelatin flocculate sus- pensions of arsenic trisulphid; if small amounts of gum arabic are added flocculation is prevented. In order that a protected suspension shall be produced in which no precipitation will occur, very definite proportions between the three suspensions must be arrived at, but a number of quantitatively varying mix- tures of the three can be produced which will hold up without precipitating for a considerable period. Like the serums de- scribed above, two such suspensions in which the relative propor- tions of the three are not the same will precipitate each other when by rapid mixing the quantitative relationship necessary for protection is suddenly disturbed. We have here, then, a complex analogy to the conditions in the serums. Two substances, mutually flocculable, do not pre- cipitate. They are prevented from precipitating by the presence of a third substance which ‘‘protects’’ when certain definite proportions between the three are maintained. Many quantita- tively different mixtures of this kind may be made in which flocculation is in this way prevented. Mix two such protected ANAPHYLACTIC PHENOMENA 207 mixtures, disturb these proportions and flocculation occurs, faster or slower according to the relations arrived at in the mixtures. Moreover Porges ** has shown that the factor of colloidal protection may well play a part in the occurrences taking place in a medium of blood plasma or serum. He has found that fresh native serum will precipitate mastic emulsions. The same serum heated, if used in very small quantities, will protect mastic emulsions against precipitation of the fresh serum. This alone shows what delicate physical changes in the body fluids may make for fundamental changes of reactions. In our own experience these experiments of Porges were in principle confirmed ; small quantities of heated dog serum added to arsenic trisulphid precipitated this suspension; slightly greater quantities again dispersed it. Of similar significance are experiments by Streng on the so-called conglutinins, sub- stances in serum which are supposed to produce an agglutina- tion of blood-corpuscles or bacteria which have been previously treated with fresh serum or alexin. The addition of minute quantities of alexin to typhoid bacilli and agglutinin prevents agglutination, Friedemann,** furthermore, a pioneer in this branch of serum investigation, in studies on the serum reactions has come to the conclusion that certain anticomplementary activities of the serum globulins may be inhibited by the albumins of the same serum. Schmidt * speaks of a similar Schutzwirkung on the albumin of normal serum. When lues serum was mixed with certain lipoid extracts (of human heart, used for Wassermann antigen) precipitation resulted. Such precipitation was brought about also by the globulins of normal serum—but was prevented or ‘‘protected against’’ when the albumin of normal serum was added to the mixtures. Friedemann himself (and Schmidt agrees with him on the main points) thinks that the globulins and albumins of normal serum are in antagonism, the albumins preventing certain reactions (such as complement fixation) in which the former become active as soon as the albumins are removed or diminished. 208 HARVEY SOCIETY We do not have to force analogy to look on such serum reactions as essentially following laws similar to those observed in the case of chemically definable colloids. Apart from the protein character of serum constituents, we know that serum reactions follow quantitative laws analogous to those observed in colloidal reactions (inhibition zones, ete.). We know the importance of the electrolytes in the phenomena, we know that the immune bodies like the colloids diffuse but slowly, and we know from the work of Landsteiner and Pauli *® especially that certain serum hemagglutinins will wander, like other colloidal substances, to one pole or the other when a direct electric cur- rent is passing through solutions containing them, like ampho- teric substances changing the direction of wandering according to the alkalinity or acidity of the menstruum. The points of similarity are too numerous to be exhaustively reviewed in this connection. They are so many and so striking, however, that we should hesitate to apply any explanation to serum phenomena of any kind which is not in accord with the general behavior of colloids. In recent experiments of our own, moreover, we have been able to show that when precipitin reactions are set up in com- parative series, in one case using the globulins of normal rabbit serum, in salt solution, as the diluent for the antigen, and in another series the albumins of the same serum, the reactions in the latter are noticeably slower than in the former—than similar reactions in salt solution or in active or inactive serum. There is apparent inhibition of the reaction by the serum- albumin. Enough has been said to show the justification of any theory which utilizes as a major premise the possibility of the partici- pation of protective colloids in reactions taking place within the vessels of an animal. We suggested some years ago in a paper on this subject that it was such a protective colloidal action in the plasma of animals which prevented the rapid union of antigen and antibody in the blood-stream, and we thought at the time that such an arrangement would indeed constitute an automatic protection of animals against sudden and severe ANAPHYLACTIC PHENOMENA 209 injury when a foreign protein gained entrance to the blood- stream. Our conception of the whole process would therefore be something as follows: The injection of a foreign antigen into the animal body leads it to antibody formation by the tissue cells. These antibodies are in part discharged in the blood- stream and in part sessile on the cells. There is a gradual union between the circulating antigen and antibody and probably between the circulating antigen and the sessile antibodies. Under conditions apt to occur in the course of normal con- ditions the quantity of nitrogen which gains entrance is small and no injury results from such union by which probably a gradual parenteral digestion of the foreign substances is ob- tained. When in the course of abnormal states, infectious dis- ease, etc., a situation arises in which considerable amounts of antibody have been formed and relatively large amounts of antigen are also present, all the conditions are furnished for what we call anaphylactic injury, unless there were some efforts to prevent the rapid union in these animals. In the anaphy- lactic experiment we see that the rapid union of antigen and antibody on the cell will kill. But it is likely that in most cases during immunization the circulating antibodies are far in excess of those still sessile on the cells, and were rapid union between these and the antigen not inhibited in the circulation, the animal would be constantly and severely ill during all pro- cesses of immunization. However, we know that in highly immunized animals antigen and antibody may be present side by side ununited. Is it not necessary to assume that this is evidence of a protective inhibition of union? For the colloidal protection would lead to a very slow union, in which, because of the gradual nature of the process, practically no severe injury of the individual could result. According to this conception we can quite easily explain why the simultaneous injection of antigen and antibody into the normal animal would result ordinarily in slight and delayed symptoms. Accidental success in so balancing the proportions that complete elimination of protection results would account for the occasional acute symp- toms and death observed in such procedures. It is quite clear 14 210 HARVEY SOCIETY that such an ideal experiment cannot be regularly obtained, for the simple reason that the protective element may be subject to variation, and since there are so many secondary factors even in test tube experiments on precipitation which influence such reactions. When the animal is sensitized by the methods of the classical anaphylactic experiment, the union in the cells, violent and stormy, results in death after anaphylactic shock, and what- ever symptoms might have resulted from the union of the two substances in the blood-serum are overshadowed and secondary. It is perfectly clear that there are many gaps in the absolute experimental proof of such a conception. We know, however, that slow, gradual and acute injury may follow on the simul- taneous interaction of antigen and antibody in the animal body. We know from the many experiments of Vaughan, Friedberger and others that in vitro such a meeting in the presence of active serum can result in the production of injurious substances which produce anaphylaxis-like symptoms when injected into the animal. We know from the experiments of Doerr that the injection of formed precipitates will injure. Whatever we may think about the nature of the poison and its mechanism of pro- duction there is little reason to doubt that the noxious agent can be produced without reference to the body cells. And we believe from this, together with the premises on which we have developed our idea of colloidal protection, that such a concep- tion may form a perfectly legitimate explanation for the scat- tered and yet definite observations made since Friedemann by many others and by ourselves, of immediate symptoms after simultaneous injection of antigen and antibody. Va In discussing the probable localization of anaphylactic reac- tions in the preceding paragraphs, we limited ourselves entirely to the phenomena occurring when sensitization is carried out with non-cellular substances such as blood-serum, egg albu- min, ete. When the antigen employed is cellular, consisting of bacteria ANAPHYLACTIC PHENOMENA 211 or red blood-cells, we are confronted with a problem of consider- ably greater complexity. As morphologically compact struc- tures these cells cannot enter into direct chemical relations with the fixed tissue cells until they have been either disinte- grated or at least have given up constituents to solution in the blood plasma. In consequence we must assume two separate phases of all such reactions—one the occurrences within the circulating blood in which the injected cells come in contact with the solvent elements of the plasma and during which the solution of antigenic constituents is brought about, the other the subsequent reactions entered into by these dissolved sub- stances, either within the circulation or on the fixed tissue cells with their respective receptors or antibodies. If therefore Doerr *® and others (Denzer and Weil) claim that anaphylaxis with cellular antigens is entirely similar in principle to that produced with dissolved, unformed antigens, they may well be perfectly right in so far as the second phase of these phenomena is concerned. They found that guinea pigs injected with hemolytic serums reacted to the injection of the blood-cells when, as in passive serum anaphylaxis, a latent period or interval was allowed to elapse between the adminis- tration of the antibodies and that of the nitrogen. This means simply that they failed to obtain acute or marked symptoms (for quantitative reasons possibly) when the cells and anti- bodies met in the blood-stream. Analyzing the phenomena in this way it becomes clear that when we inject cellular material we are merely injecting an antigen—or more probably a group of antigens—enclosed in the morphological structures of the cell, and amenable to reac- tion only after liberation. After this has taken place, subse- quent occurrences should in no important principle differ from those following on the injection of an unformed substance like serum, or we may say for the sake of clearness, a predissolved antigen; and all that we have said about such conditions in our preceding discussion should apply here. Added to this, however, we have in the ease of cellular antigens a process unnecessary when unformed antigens are 212 HARVEY SOCIETY injected, namely, the cytolytic or cytotoxic reaction which pre- cedes the liberation of the cell-constituents, and in the course of which the formed elements are broken up. And we need only compare the slow autolytic disintegration of cells in sterile inactive serum or salt solution with the rapid changes occur- ring in active hemolytic or, in certain cases, in bacteriolytic serums, to be convinced that such disintegration is due to reac- tion with active serum constituents. We may logically accept, then, that by injecting cells, we are for one thing injecting substances which will, in part, soon be liberated and which will call forth all the changes and enter into all the reactions which are associated with the injection of dissolved antigens. In addition to this, however, we are con- fronted with a further problem. Is there injury to the animal body, comparable in broad principles with anaphylaxis, during this intravascular reaction between whole cell and cytolytic antibodies which precedes the liberation of the soluble constit- uents? Is there, in other words, a true ‘‘cell anaphylaxis’’? Since it is probable that the principles of cellular anaphy- laxis are the same whatever the variety of cell employed, we may take red cell hypersusceptibility as a basis for discussion. It is a well-known fact, long recognized, that a serum which is capable of hemolysing the red cells of any species is toxic when injected into an animal of this species. This is true not only of hemolytic serums but also of such normal serums which like, let us say, goat serum and rabbit cells, can hemolyze normally the red cells of another animal. Since occasionally the serum of an individual of one species can so act on the red cells of another individual of the same species, our surgeons call for careful investigation of receptor and donor before performing transfusion. The injection of such a serum intravenously may kill with symptoms not unlike anaphylactic shock. Here it is often difficult, as we shall see (or indeed it may be impossible), to determine, whether such death is truly anaphylactic in nature or whether it is due to clumping of red cells or hemagglutina- tion, a property which is very often an accompaniment of hemolytic power. However, hemagglutinating properties can- ANAPHYLACTIC PHENOMENA 213 not be held responsible for the cedema and localized injury which, as Uhlenhuth and Haendel have shown, may follow the subcutaneous injection of such serums. It thus appears as though the process of hemolysis were accompanied by the libera- tion of injurious products. The first systematic investigation of red cell anaphylaxis was undertaken by Ulrich Friedemann.'! Friedemann injected washed beef cells into rabbits and followed this by a second injection after from seven days to three weeks. Rabbits so treated showed the symptoms ordinarily associated with ana- phylaxis in these animals. Active sensitization seems thus to have been accomplished with beef cells. Schiff and Moore have recently suggested that Friedemann really obtained serum ana- phylaxis, but since Friedemann explicitly states that he worked with washed cells, we can see no just reason for such an assump- tion. Another objection to Friedemann’s results, however, is possible—one which is far less easy to controvert—namely, that the illness of his rabbits may have been due to hemagglutina- tion, which by itself may produce serious illness or even death by mechanical obstruction of blood-vessels. Friedemann, in- deed, takes cognizance of this possibility but makes no attempts to rule it out in his experiments. As a matter of fact we think it unlikely that hemagglutination played a part in his rabbits, but the possibility cannot be excluded. We will revert to this particular question. Passive sensitization was produced by Friedemann against beef cells in rabbits by injecting the specific hemolytic serum. He obtained his best results when he injected serum and cells together, mixed in vitro. However, he also obtained positive experiments when the two were simultaneously injected into opposite veins. His results were inconstant when he allowed an interval to elapse between serum and cell injection—a fact which argued for the direct occurrence of the reaction within the circulation. Most important of all, Friedemann mixed hemolytic serum and cells in test tubes, letting them stand for five minutes in a water bath and then, before any considerable degree of hemol- 214 HARVEY SOCIETY ysis had taken place, he centrifugalized and injected the faintly red supernatant fluid into rabbits. A rabbit so injected be- came extremely ill and many of them died after shorter or longer intervals, with symptoms typical of anaphylaxis in rabbits. Friedemann concluded that when red cells came in contact with hemolytic antiserum, poisonous substances were liberated, even before actual hemolysis had taken place, and that these toxic products were responsible for the subsequent injury to the animal, He identified the anaphylactic antibody with the hemolysin. This view, therefore, is identical in prin- ciple with the one we have discussed as the conception of parenteral digestion. Indeed Friedemann’s experiments fur- nished the point of departure for Friedberger’s subsequent work on the so-called ‘‘anaphylatoxins.’’ Doerr and Moldovan,*’ a little later (1910), studied the effects of the injection of serums hemolytic for guinea pig erythrocytes into guinea pigs, and drew conclusions which substantiated those of Friedemann. They found that the toxic effect was due to the action of the hemolytic serums on the guinea pig erythrocytes. Toxicity could be removed from such serums by absorption with these cells, and the toxic products could be produced by contact of serum and eells in vitro. From these experiments, again, it seemed that the liberation of a toxie substance followed on contact between erythrocytes and specific antibodies, whether this contact took place within the circulation or in the test tube. That the antibodies concerned need not necessarily be identical with hemolysins themselves follows, we think, from the work of Doerr and Moldovan as well as from work of our own on the toxicity of certain normal serums **—experiments which could not be discussed in detail without taking more space than seems justified. Although much irregularity of result has been obtained in the production of active erythrocyte anaphylaxis in both guinea pig and rabbit experiments, nevertheless, it seems clearly established that acute death does follow the repeated injection of such cells when dosage and interval are properly observed. The recent experiments of Schiff and Moore,*® though they ANAPHYLACTIC PHENOMENA 215 clearly illustrate the difficulties of such procedure in guinea pigs, still record a sufficient number of positive results to recon- firm its actual occurrence. From one of these experiments, indeed, as well as from the experience of Friedemann and others with passive sensitization by antierythrocyte serums, it would appear that with red cells the phenomenon requires a procedure differing from that successful with serum anaphy- laxis, in that a considerable concentration of antibodies is needed, 7.€., a condition calling in the active experiment for more than one preparatory injection, or, in the passive sensi- tization, for the injection of a serum of high potency. This, as we know, is the case, also, in bacterial anaphylaxis, in which experiments are usually successful only if many and repeated preparatory injections are made. It is this factor, possibly, which may account for the failure of so many workers to obtain true cell anaphylaxis when they have followed the technic successful in the serum experiments—i.e., that of only one preliminary sensitizing dose—or that, in the passive experiment, many have failed to duplicate Friedemann’s success when both antigen and sensitizer were simultaneously injected. It is more than likely that a weak sensitization and consequently a slow reaction between the cells and the antiserum may be interrupted by prompt phagocytosis of the injected cells, with consequent protection against the further developments of the process. It is true that in many eases of erythrocyte anaphylaxis it may be impossible to say with certainty whether death was due to true shock or whether it was caused by embolic processes due to hemagglutination. This possibility has not been ruled out in many otherwise complete investigations—though in ex- periments like those of Friedemann and Amako °° it seems but a remote possibility. However, in individual instances, such as our own experiments with normally toxic serum, it has been shown that the toxicity may disappear with inactivation, though hemagglutinating properties are retained, and it seems that, to kill acutely hemagglutination must be rapid, powerful and extensive. Moreover, the speed and completeness of recovery showing non-lethal degrees of erythrocyte anaphylaxis argues 216 HARVEY SOCIETY at least against the frequent occurrence of hemagglutinative death by embolism in experiments carried out in this way. The local injury following the subcutaneous injections of normal and immune hemolytic serums must of course occur entirely independent of hemagglutination. Finally, the fact that con- tact of the cells with active serum—as first carried out by Friedemann—produces a poison in vitro which kills acutely with symptoms of anaphylaxis, seems to render fairly certain the assumption that similar contact in the circulation may lead to like result. For we know that the entire process of hemolysis ean take place intravascularly. Whether the antibodies that so react with the cells are the hemolysins themselves, is a question that we hardly have the time to discuss and which moreover is merely an incidental one. After all, hemolysis itself is merely one visible result of a reaction which probably affects profoundly the entire cell struc- ture. About bacterial anaphylaxis our knowledge is still more defective than is that occurring when erythrocytes are used. We do know, however, that active sensitization with bacterial pro- teins and while whole bacteria is possible—though many injec- tions are apparently necessary—the exact procedure being sub- ject to so many fortuitous influences that so far no regularly suc- cessful method can be outlined. We also know that, as with red cells, contact between the bacteria and active serums will result in the production of acutely toxic substances—which we have dis- cussed above as ‘‘ proteotoxins.”’ We may summarize our views on cell anaphylaxis, briefly, as follows: When whole cells are injected into an animal two distinct processes are set in motion. First, the formed cells come into relation with circulating antibodies. During this con- tact toxic substances—‘‘ proteotoxins’’—may be set free if quan- titative relations are suitable and cells sufficiently sensitized. Where the matrix of the poison is found and to what an extent the complement participates—these are in many respects still open questions. This reaction alone, if sufficiently vigorous, may cause acute symptoms and even death. During this reaction antigenic cell constituents are set free ANAPHYLACTIC PHENOMENA Q17 to solution and these then enter into reaction with their respec- tive antibodies or receptors in the blood or on the fixed cells. The last-named reactions are entirely comparable to those of serum anaphylaxis and have been sufficiently discussed. Whether in the first-named process, when the whole cell meets its antibody in the blood-stream, we regard the poison as originating from the matrix of the antigen or from the serum itself by the withdrawal of antienzymes is immaterial. The reaction is subject to so many modifying factors that experi- mental control is made difficult and results cannot at present be so regularly foretold as is the case in serum anaphylaxis. It seems probable from the work of Friedberger and others that a delicately balanced optimum proportion between antigenic cells and antibodies must be obtained. Moreover, unless the process is rapid and harmful effects very sudden, prompt phago- eytosis of the cellular elements may remove the antigen from further reaction possibility. It is plain that such a conception has the greatest importance in the understanding of infectious diseases. When bacteria form the antigen which gains entrance to the animal body, the gradual stimulation of specific antibodies in the animal may eventually lead to such a two-phase reaction. Specific sensi- tizers or amboceptors (cytotoxins) are gradually formed and these may react with the dead and the living micro-organisms. There may be a direct formation of proteotoxins and at the same time a liberation of soluble antigen from the bacteria. It may be, as von Pirquet has suggested, that the sufficient estab- lishment of such reactions between cell and antibody may mark the end of what we speak of as ‘‘incubation time,’’ no noticeable time accruing to the animal body until the antigen-antibody reaction has been initiated. The ‘‘proteotoxins’’ so formed, whatever their matrix, may then, as we have shown with Dr. Dwyer, act as aggressins, lead to a leukopenia, as in typhoid fever, and thereby increase indirectly the invasive capacity of the micro-organisms. The antigenic substances which have gone into solution may at the same time react both on the fixed cells with sessile receptors, and, to a merely incidental degree, 218 HARVEY SOCIETY with their receptive circulating antibodies, adding thereby to the injury sustained by the host. True immunity against dissolved antigens, we have stated in the beginning, probably does not exist, for animals having high antibody contents in their serum may still die suddenly with convulsions after a fourth or fifth injection with foreign serum. In the case of cellular antigens, however, and especially bacteria, true immunity may exist in two forms. On the one hand if the animal possesses a high concentration of antibodies before the micro-organisms have gained entrance, an immediate bactericidal effect may prevent their multiplication, the harm- ful effects resulting from the union of the small initial amounts of antigen and antibody being so slight as to be unnoticeable. Again, after the bacteria have gained entrance, if the quanti- tative relations between antigen and antibody are such that the reaction is either slight or for purely quantitative reasons results in little injury for the time being, then sensitization of the bacteria or other cells by the antibodies leads to rapid phagocytosis. And this process of phagocytosis represents true immunity, a removal of bacteria incidental to which there is, as far as we know, no injury to the host. It is in the process of phagocytic removal, chiefly, in which the reaction to cell injection differs from that taking place in response to the administration of unformed protein. It may be this element which renders it so difficult to obtain sharp anaphylactic reac- tions with cellular antigens. And it is the absence of phagocy- tosis in the latter case which probably prevents the existence of a true immunity. LITERATURE DIRECTLY RELEVANT *Morgenroth: Ehrlich Gesammelte Arbeiten, translation, Wiley & Son, New York, 1906, footnote, p. 332. ? Richet and Héricourt: Compt. rend. Soe. de biol., 1898, xv, 137. *Arthus: Compt. rend. Soc. de biol., 1903, lv, 817. *Von Pirquet and Schick: Die Serumkrankheit, Wien, Deuticke, 1906. *Rosenau and Anderson: Bull. 29, U. S. P. H. S., 1906; Bull. 30, 1906; Bull. 36, 1907; Jour. Med. Research, 1906, xv, 179; ibid., 1907, xvi, 381; Jour. Infect. Dis., 1907, iv, 552; ibid., 1908, v, 85. ANAPHYLACTIC PHENOMENA 219 *Otto: Das Theobald Smitsche Phaenomen, etec., von Leuthold Gedenk- schrift, 1905, i. ™ Gay and Southard: Jour. Med. Research, 1907, xvi, 143. * Besredka: Bull. de l’Inst. Pasteur, 1908, vi, 826. * Vaughan: Protein Split Products, ete., Lea & Febiger, 1913. * Friedemann: Ztschr. f. Immunitiitsforsch. u. exper. Therap., 1909, ii, 591. 4% Friedberger: Berl. klin. Wchnschr., 1910, p. 1490, 1922; Ztschr. f. Immunititsforsch., 1910, iv, 636. * Keysser and Wassermann: Folia serol., 1911, vii; Ztschr. f. Hyg. u. Infectionskrankh., 1911, Ixviii, 535. * Bordet: Compt. rend. Soe. de biol., 1913, Ixxiv, 877. * Jobling and Peterson: Jour. Exper. Med., 1914, xix, 239, 251, 383, 459, 480. * Zinsser and Dwyer: Jour. Exper. Med., 1914, xx, 387, 582. %* Zinsser: Jour. Exper. Med., 1912, xv, 529; 1913, xviii, 219. * Sleeswijk: Ztschr. f. Immunititsforsch., 1909, ii, 133. * Friedberger and Hartoch: Ztschr. f. Immunititsforsch., 1909, iii, 581. Loeffler: Ztschr. f. Immunitiitsforsch., 1910, viii. »° Hartoch and Sirenskij: Ztschr. f. Immunitiitsforsch., 1910, vii. 2 Nolf: Ann. de l’Inst. Pasteur, 1900, xiv. * Hektoen and Ruediger: Jour. Infect. Dis., 1904, i. * Friedberger and Hartoch: Ztschr. f. Immunitiitsforsch., 1909, iii. * Ritz, cited by Doerr: Footnote 29. * Zinsser and Dwyer: To be published. * Doerr and Russ: Ztschr. f. Immunitiitsforsch., 1909, iii, 181, 706. * Friedberger: Ztschr. f. Immunitiitsforsch., 1909, ii, 208, 644. * Doerr: Ergebnisse der Immunitiitsforschung, edited by Weichhardt, Ber- lin, 1914, i, 257. *® Pearce and Eisenbrey: Congr. Am. Phys. and Surg., 1910, viii, 402. ® Schultz: Jour. Pharmacol. and Exper. Therap., 1910, i. * Dale: Jour. Pharmacol. and Exper. Therap., 1913, iv. “Weil, R.: Jour. Med. Research, xxvii, 497, 1913; xxx, 87, 299, 1914; Proc. Soc. Exper. Biol. and Med., xi, 86, 1914. *® Coca: Ztschr. f. Immunititsforsch., 1914, xx, 622. “ Biedl and Kraus: Ztschr. f. Immunitiitsforsch., 1910, iv, 115. * Briot: Compt. rend. Soc. de biol., 1910, Ixviii, 402. * Gurd: Jour. Med. Research, 1914, xxxi, 205. * Linossier and Lemoine: Compt. rend. Soc. de biol., 1902, liv, 85. * Eisenberg, P.: Centralbl. f. Bakteriol. I Abt, Orig., 1903, xxxiv, 259. *® Ascoli, M.: Miinchen. med. Wehnschr., 1902, xlix, 1909. “V. Dungern: Centralbl. f. Bakteriol., I Abt, Orig., 1903, xxxiv, 355. “Von Dungern: Centralbl. f. Bakteriol., I Abt, Orig., 1913, xxxiv, 355. Auch hier handelt es sich nicht um zwei reaktionsfiihige Kérper, deren Verbindung aus irgend Grunden unterbleibt, sondern um Substanzen, 220 HARVEY SOCIETY welche keiner Affinitiit zueinander besitzen. Die betreffenden Kanin- chen haben zu dieser Zeit noch nicht alle méglichen Teilpriizipitine gebildet, sondern nur einzelner derselben. Diese zuniichst produzierten, nur auf bestimmte Gruppen der priizipitablen Eiweisskérper passenden Partial-priizipitine sind es, welche nach der Absiittigung aller zur Verfiigung stehenden zugehérigen Gruppen der priizipitablen Substanz im serum nachweissbar werden. Daneben bleit aber ein anderer Teil der priizipitablen Substanz, der keiner Affinitiit zu dem gebildeten Priizipitin besitzt, bestehen, solange bis ein anderes Partial-priizipitine von den Kaninchenzellen geliefert wird welches sich mit Gruppen der in Lisung gebliebenen Eiseisskérper vereinigen kann. # Zinsser and Young: On the Possible Importance of Colloidal Protection in Certain Phases of the Precipitin Reaction. Jour. Exp. Med., 1913, Xvii, 396. * Porges, O.: In Kraus and Levaditi: Handb. d. Technik u. Methodik der Imm., Jena, 1909, ii, 1146. “Friedmann: Ztschr. f. Hyg., 1910, lxvii, 279. “Schmidt: Ztschr. f. Hyg., 1911, lxix, 513. “ Landsteiner and Pauli: Cited from Landsteiner, “ Colloide u. Lippoide in der Immunitiit,’” from Kolle and Wassermann, Ed. 2, ii, 1244. * Doerr and Moldovan: Ztschr. f. Immunitiitsforsch., 1910, vii, 223. * Zinsser: Jour. Exper. Med., 1911, xiv, 25. 4° Schiff and Moore: Ztschr. f. Immunitiitsforsch., 1914, xxii. 5 Amako: Ztschr. f. Immunitiitsforsch., 1914, xxii. SOME PROBLEMS IN THE PATHOLOGY OF SYPHILIS * PROFESSOR JOHN A. FORDYCE Columbia University, New York T is proposed in the present paper to discuss some of the phases of syphilis which have resulted from the intensive investigations of the past ten years. These problems include the question of immunity, a few of the most striking anatomico- pathological changes met with in the disease, as well as its effects on the cardiovascular and nervous systems, and the les- sons to be deduced from the knowledge which has been gained. IMMUNITY One of the most interesting and perhaps least understood problems in the pathology and biology of lues is the question of immunity. Formeriy it was taught that like many of the other infections one attack of syphilis conferred a lifelong immunity, that a syphilitic mother might transmit an immunity to her child, while, conversely, the mother of a syphilitic child was rendered insusceptible to the disease, as postulated in the well-known laws of Profeta and Colles-Baumés. Since the work of Neisser, Finger and Landsteiner, Levaditi, Uhlenhuth and Mulzer, and others, however, our conception of immunity in syphilis has undergone considerable modification. Investigations along these lines and in microbiology have shown that syphilis is closely allied with the protozoal affec- tions, especially those caused by trypanosomes and _ piropla- somes. These diseases differ from those of bacterial origin in the manner in which they respond to the introduction of organ- isms into the system. While in the great majority of infectious processes acquired immunity is expressed by a protection more or less absolute against a new attack and the disappearance of the causal agent from the body, or its innocuousness to the host * Delivered February 13, 1915. 221 222 HARVEY SOCIETY if it persists, diseases of protozoal origin behave in a different way. Levaditi has demonstrated, for instance, that animals suffering with a spirillary infection are immune to a new inocu- lation. Their serum has a high antibody content, but the blood still harbors parasites and is capable of producing a fresh infec- tion in healthy animals. So with the serum of guinea pigs inocu- lated with Nagana or Surra trypanosomes. This is trypanocidal for these organisms in vitro, but in vivo they have acquired an insensibility to the trypanolytic antibodies, for the blood and tissues of the animals still contain parasites. The same is true of human subjects suffering from sleeping sickness in whose serum trypanolytic, agglutinating and other protective bodies have been demonstrated. Carrying the analogy to syph- ilis we find that an individual may harbor spirochetes for forty or fifty years, while his skin and mucous membranes exhibit an insusceptibility to re-inoculation under natural exposure. However, as soon as he is freed from his infection he is again in as susceptible a state as he was prior to his first attack. It is a well-known clinical observation that after the develop- ment of the initial lesion a syphilitic is immune to a fresh infection. Why have the integument and mucous membranes ceased to react to organisms introduced from without when they are susceptible to their action from within? The opinion has been generally held that for infection to take place one of the first laws of its requirements is fulfilled by a surface covered by squamous epithelium. The earlier experiments on animals seemingly corroborated this view, for inoculations were only successful where scarification of a squamous-celled area was performed. Improved laboratory technic has, however, over- come the difficulty, positive results now being obtained in ani- mals by the introduction of the virus into the testicle, peritoneal cavity, under the skin, or directly into the circulation, especially the heart. In human syphilis, while it is more usual for the general infection to be preceded by the development of a primary lesion, somewhere on the cutaneous or mucous surface, it is not rare for the local reaction to be altogether absent. To illustrate: PROBLEMS IN THE PATHOLOGY OF SYPHILIS 223 a male nurse in my hospital service pricked himself with a needle used for obtaining blood for the Wassermann test directly after he had withdrawn it from the vein of a patient with florid syphilis. He was carefully watched for the development of a chancre at the site of injury, but the first macroscopic mani- festation was a late secondary papule on his foot. Similar cases are reported in the literature. Another instance is exem- plified in a personal communication by Dr. Dade, in which he related the infection of a young girl by blood transfusion. Her brother, whose blood was used for the treatment of her anemia, was at the time of the operation in the second incubation period of the disease. She in due course of time developed a secondary rash. Congenital lues is a further example of syphilis without the development of an initial lesion. Neisser came to the opin- ion, from his many experimental observations, that in certain cases infection might take place through the integument without visible or even microscopic changes, the organisms gaining access to the lymphatic or blood-stream direct. When spirochetes are successfully implanted, what takes place? As soon as they have become acclimated to changed nutritional conditions at the site of inoculation, as shown by Levaditi and Yamanouchi, they multiply, call forth microscopic changes in the cutaneous vessels, and enter the lymph- and blood-stream. This is known as the first incubation period and varies in length in different subjects. Just how far the organisms penetrate beyond the local sore and satellite lymph- nodes in human subjects in this stage is not known, but it has been demonstrated by Neisser that the spleen and marrow of inoculated monkeys contain virus thus early in the infection. Now, as the chancre is forming, a peculiar reaction on the part of the host is also developing. At first it is so feeble that the individual is for some time susceptible to re-inoculation. In animals it has been found that superinfection is possible if made eight days before the evolution of the primary lesion. Not rarely we see patients with multiple chancres. In some in- stances they develop simultaneously as the result of multiple inoculations; in others they appear successively either from 224 HARVEY SOCIETY auto-inoculation or contamination with a fresh virus. Human and animal experiments are at one in showing that superinfec- tion lesions do not develop in the typical manner of those im- planted on a virgin soil. Their incubation period is shorter and their clinical characteristics simulate those of the stage the patient is in at the time; that is, papular if close to the second- ary period and ulcerative if in the tertiary stage. In other words, they are the expression of an altered tissue reaction. The hypothesis has been advanced by Krause and Volk that this immunity has a regional development, beginning at the site of the primary sore and progressively involving the mucous and cutaneous covering. Neisser has suggested that the suscepti- bility to re-infection returns in an inverse manner. During the second incubation period the refractory state continues to develop until it attains its maximum with the outbreak of the secondary rash. Even now, however, it is only relative, for laboratory experiments have yielded positive re- sults if special methods are employed, namely, larger amounts of active virus introduced subepidermically. From a clinical stand-point, however, this refractory state, or anergy—the term usually applied—is complete as far as ordinary infection is concerned, disappearing only with the cure of the disease, when the patient is again in a receptive condition. It has been demonstrated further that the anergy is a specific one, the attempts of Neisser, Baermann, and Halber- stidter to induce it by inoculation with Spirocheta pertenuis, of Hidaka with Spirocheta dutton, and of Uhlenhuth with Try- panosoma lewisii having proved futile. None of these infec- tions protected against that of syphilis, while, on the other hand, luetic animals were all susceptible to frambesia, relapsing fever, and dourine. Little is known of the nature of this state of resistance or how it is brought about, whether by the tissues, cells, the serum, or intermediary organs. Organisms or their toxins when introduced into the body behave as antigens and stimulate the formation of antibodies which are specifie for that particular bacterium. ‘These antibodies are of the nature of antitoxins which neutralize the toxins; opsonins which lead PROBLEMS IN THE PATHOLOGY OF SYPHILIS 225 to the ingestion and destruction of the organisms by the leuco- cytes; agglutinins which cause the organisms to adhere together and probably also bring about their destruction; and bacteriol- ysins which cause their dissolution. From the contributions to this branch of research in lues the conclusions are that sub- stances of a parasiticide nature do not develop in the course of the disease. At the most, says Neisser, we can reckon with complement binding and agglutinating substances of a specific nature. Levaditi, however, argues that since the spirochetes even before the evolution of the primary lesion may penetrate to the hematopoietic organs, they act like true antigens and their products of secretion stimulate the formation of antibodies. He assumes an analogy between these immune bodies and those of related diseases, where bacteriotropic substances and opso- nins ensure destruction by phagocytosis. Pari passu with this antibody formation the skin acquires a refractory state to exogenous spirochetes, the biological changes being sufficiently advanced at the close of the second incubation period to deter- mine the character of the reaction. He has offered two attrac- tive hypotheses to explain why the skin which is resistant to exogenous spirochetes reacts to endogenous organisms, (1) by a generalized eruption, and (2) after a varying latency, by local and destructive lesions. His assumption is that for a time the immune bodies hinder the multiplication of the parasite, but with the enfeeblement or disappearance of the former the virus gains the ascendency, generalizes itself, and produces the secondary eruption. Or the spirochetes are vaccinated, so to speak, against the immune bodies, and, becoming resistant to these defensive substances, they succeed in multiplying and provoking the exanthem. It is a well-established clinical fact that the early cutaneous lesions are superficial and generalized, that with each relapse they show a tendency to localization and grouping with involve- ment of the deeper structures until we reach the tertiary stage with extensive destruction. Since it has been shown that these differences are not inherent in the specific virus but in the tissues of the host brought about by prolonged contact with the causal 15 226 HARVEY SOCIETY agent, the term allergy or, according to Neisser, Umstimmung has been proposed to explain the biological change which has taken place in the tissues. In the primary and secondary stages, when the organisms are numerous, the host is refractory ; in the tertiary stage when they are few in number an anaphy- lactic state exists, the tissues are frail, and more readily undergo necrosis. This is also seen in superinfection in tertiary syphi- litics when the lesion corresponds to that type, and in the luetin reaction, which is especially applicable in congenital and tertiary syphilis, depending for its result on the hypersuscepti- bility of the tissues. In tuberculosis a similar condition is en- countered. Erythema induratum, for instance, is caused by few organisms, and yet clinically presents extensive destruc- tive lesions. Another clinical type is seen in large and deep ulcerations of the extremities in tuberculous subjects, all treat- ment, excepting surgical, being of little avail. The writer has recently had under his care a young man with glandular tuber- culosis in whom the entire right leg was involved, necrosis extending through the musculature and laying bare the bone. The immunity processes as outlined in the foregoing con- siderations are subject to modification under treatment. Where treatment is intensive and yet not sufficient to effect a complete cure it would appear that the anaphylactic stage is hastened, for it has been repeatedly claimed that specific remedies given early in the disease tend to produce relapses of severe local intensity. Herein we find the basis for the contention that salvarsan has changed the course of the disease, and, when not given in sufficient doses to sterilize the patient, tertiary lesions appear much earlier than when the affection is permitted to run its course with the gradual establishment of a defensive mechanism, and consequently a longer refractory state. In so-called malignant syphilis we must also seek the cause in a precocious anaphylactic condition of the tissues rather than in the type of the invading organism, the reason usually evoked for this form of the infection. It has repeatedly been reported that spirochetes are numerically few in such lesions and that successful inoculations are relatively rare. PROBLEMS IN THE PATHOLOGY OF SYPHILIS 227 As accumulated evidence has shown that individual im- munity is only relative, so racial immunity in the strict sense of the word does not exist. Lesser has said that neither race nor climate makes any difference in the receptivity of the virus of syphilis. Furthermore, it has been demonstrated that the children of syphilitic parents do not develop an immunity. Such children may be frankly syphilitic at the time of birth or they may be in the latent stage for years (syphilis hereditaria tardiva). Or if they escape infection and acquire the disease later on, it has never been observed to have a milder course. Gluck reports that although syphilization is almost general in Bosnia, extra-uterine acquired lues is not rare. He noted 10 per cent. of cases of recent syphilis in children under fifteen years, one-third of whom were only five or six months old. While complete statistics and comparative studies on the course of the disease in the different races are not available, the investigations made have revealed some interesting facts. Neisser found that in the Malays primary and secondary syphilis was insignificant. Treatment is therefore deferred and the percentage of cutaneous tertiary manifestations is proportion- ately large. He seldom met with visceral or nerve manifesta- tions, and never encountered tabes or paresis. In Java and the tropics generally Europeans suffer more severely than the natives, which is attributable to unfavorable climatic conditions and changed mode of living. According to Quennac, Euro- peans, Hindoos, and Arabs in Africa suffer severely, while the negroes only have mild attacks. In Central America, Rutschuh reported severe syphilis in the whites and negroes in contra- distinction to the Indians and half-breeds who only suffered mildly. Among the Indians in the United States the infection is very destructive in some localities, as cited by Dr. Hrdlicka, while it runs a mild course in certain other tribes. This author- ity also states that he has never seen tabes or paresis among the Indians, although they suffer from other forms of mental disease. As to the negroes in our country, the writer has a personal communication from Dr. Green, of Milledgeville, Georgia, in which the statistics collected by him show that 5.27 per cent. 228 HARVEY SOCIETY of negroes suffer from paresis as against 2.12 per cent. of whites. Tabes is rare in this race. It is often said that the European pandemic at the end of the fifteenth century was an exceedingly malignant one. This claim is now challenged by syphilographers that the disease was malignant in the present sense of the word, that is, ulcerating lesions in the early secondary period, but rather that many of the so-called cases were ordinary ulcerating tertiary forms. Instead of malignancy per se and a racial susceptibility, other factors are called into account, as the undeveloped treatment, wars, famines, ete. Neisser suggests that spirochetes in their continual passage through human beings lose in virulence, or that the treatment which the patient receives so modifies the virulence that when such modified spirochetes are transmitted they produce a modified disease. HEREDITY The question of the transmission of syphilis to the offspring is a very complicated one, and still far from clear. While the accession of serological knowledge has presented a partial solution, there still remain certain clinical phenomena for which no satisfactory explanation can be given. The opinion is gaining on almost every hand to-day that germinal infection on the side of the father does not take place, but that the mother, infected by the father either before or at the time of conception, transmits the disease to the foetus through the placenta. Is this view tenable on the ground of our clinical and laboratory observations? Disease of the testicles during the secondary stage is not eommon, although it has been demonstrated by animal experi- ments that this organ is one of the seats of election of the syphilitic virus. Later a diffuse orchitis is more frequent, and during the tertiary stage gummata are not uncommon. As to other lesions of the genito-urinary tract which might harbor spirochetes little is known. In the transmission of syphilis the results depend upon the intensity of the infection, which intensity weakens with time. PROBLEMS IN THE PATHOLOGY OF SYPHILIS 229 In the secondary stage it is most active, so that a man is almost certain to infect his wife. If she becomes pregnant early abortion will result. Such a woman, if untreated, may abort later, then have a stillborn child or a living child with active syphilis, and finally, after eight to twelve to twenty years, healthy children. Such women are obviously syphilitic. On the other hand, there is the group of women who have been free from the clinical manifestations of the disease but who have had several miscarriages or children with unmistakable signs of the infection. These women, according to the Colles- Baumés law, have not syphilis but are immune to syphilis. What has taken place that prevents such a woman from develop- ing a primary sore or a secondary rash and still has rendered her insusceptible to the disease? Is it because she has the infection or because Wassermann substances or antibodies are passed through the placenta? In other words, must we con- elude in view of the positive Wassermann reaction that her tissues are harboring living spirochetes which, without treat- ment, are overcome by the defensive mechanism of the patient, or does she simply receive Wassermann-producing substances? Serological research has in the main overthrown Colles’s and Profeta’s laws. It has been demonstrated by a host of investigators that the majority of women apparently healthy who give birth to syphilitic children have a positive Wasser- mann reaction and are therefore in the latent stage of the disease. Furthermore, Baisch, Trinchese, and Weber constantly found spirochetes in the maternal portion of the placenta and in the intervillous spaces even in negatively reacting women. The claim then that Wassermann-producing substances pass over from the fetus to the mother had to be dismissed as the persistence of the reaction after birth was strong presump- tive evidence of the presence of living spirochetes, their demon- stration in the placenta supporting the contention. The trans- mission of organisms to the foetus by way of the placenta finds its analogy in other diseases, such as smallpox, relapsing fever, and tuberculosis. In tuberculosis this takes place only in ad- vanced and florid cases in contradistinction to lues, where no 230 HARVEY SOCIETY clinical evidence of the disease may be present, multiplication of the spirochetes probably taking place through the impetus given by pregnancy. It is interesting to note in passing that Uhlenhuth found typical testicular lesions in the young of a syphilitic mother rabbit, illustrating that the organisms pass through the placenta. Varying degrees of latency are met with. In some women a negative Wassermann in the blood is accom- panied by a positive reaction in the colostrum of the breast, or the placental blood may be stronger than the rest of the serum. The evidence submitted in favor of maternal trans- mission is convincing, but the question as to how the mother received her infection is not so easy of solution. Is she inocu- lated prior to or at the time of conception or later? Do the organisms penetrate the ovum or do they enter the tissues of the mother without the production of the usual sequence of symptoms? It has been shown experimentally that the semen of latent syphilitic men may from time to time contain spiro- cheetes. As many of these patients have had more or less treat- ment, it has been suggested that an attenuated form of the dis- ease is transmitted to the wife. There are two ways in which paternal transmission could take place. In the one we may assume that the parasite is carried in the head of the sperma- tuzodn and with it penetrates the ovum. As the organism is three times the length of the spermhead this possibility is dis- missed by the majority of syphilographers. The other alterna- tive is that the spirochetes are free in the semen and with it enter the egg cell. Such an advent, it is believed, would have an untoward effect on the fertilized ovum, as it has been shown that the disturbance of a single blastomere in lower vertebrates is sufficient to arrest development or cause malformation, and the regenerative powers are greater in these animals than in the higher vertebrates (Weber). LATENCY What is our present conception of latency in syphilis? Dur- ing these periods of apparent quiescence does the defensive mechanism hold the pathogenic agent in abeyance or is some PROBLEMS IN THE PATHOLOGY OF SYPHILIS 231 pathological process insidiously undermining tissue in some part of the body? The application of the Wassermann reaction has materially changed our views since the demonstration in a large number of so-called latent syphilitics of involvement of the aorta and the central nervous system, It has been shown that such patients may have a chronic meningitis for years without producing obtrusive symptoms, and a limited aortitis may persist for a long time without making its presence known either subjectively or objectively. Aside from gummatous in- volvement of the viscera, little is known of the effects of the infection on the various organs. Where must we seek the explanation of persistent Wassermann reactions in cases inten- sively treated in whom involvement of the central nervous system and the heart and aorta can be excluded? A solution will probably be reached when the nature of the Wassermann reaction is fathomed, but on a priori grounds if a focus in the aorta or nervous system is capable of keeping up a positive reaction indefinitely, it is reasonable to assume that a similar process in one of the viscera may be provocative of a like result. It is often asked where the spirochetes are lodged during the latent period. A number of investigators have urged that, during this stage, resting forms of the organisms are harbored by the lymph-nodes or blood-forming organs. Present knowl- edge does not support the theory of a cycle of evolution with the power to produce lesions according to the stage of its de- velopment. Spirochetes recovered from any source whatever— chancre, mucous patch, or gumma—the blood, nervous system, or viscera have all shown similar morphological and practically the same biological characteristics. The relation of trauma to the localization of specific lesions has been so well appreciated by the medical profession that it is given a prominent place as an etiological factor. The query naturally arises as to whether the spirochetes in such cases have remained in loco since their primary deposition or are earried there by the blood after injury to the tissue. Pasini found spirochetes in an atropie and pigmented spot two years after the involution of a papular syphilide, while Hoffman 232 HARVEY SOCIETY demonstrated them in scar tissue of chancres long after their regression. Levaditi and Yamanouchi were able to recover them from the cornea of a rabbit 113 days after the keratitis had healed and at which time a recurrence took place. The so-called chancre redux is interpreted by many as the result of organisms which have escaped destruction rather than as a superinfection with a new strain. Neisser has demonstrated that spirochetes may be present in the skin without producing any macroscopic changes, and more recently Whartin has called attention to their presence in the heart without exciting tissue reaction. Studies on the infectiousness of the blood in the various stages of syphilis have supplied us with the following data: In 19 eases of primary syphilis, Uhlenhuth and Mulzer obtained positive results in 16 rabbits injected with 2 c.c. of defibrinated blood into the testicle; in 36 cases of early secondary syphilis, 27 positive results, and in 15 eases of latent syphilis, 2 success- ful inoculations. One of the latter was with the blood from a woman whose infection was four years old and who eighteen days before had given birth to a syphilitic child. Lieberman also succeeded with the blood from a woman whose primary lesion dated back four years and who six weeks before had given birth to a luetic child. Friihwald reports 2 cases: In one the disease had existed for one year, the Wassermann was posi- tive, and the patient had had two doses of neosalvarsan, each 0.75 Gm.; in the other the infection was one and a half years old, the Wassermann was positive, and the patient had received two doses of neosalvarsan of 0.6 Gm. each. Numerous attempts have been made with the blood of paretics and tabetics, but with little success. Levaditi demon- strated spirochetes in the blood of one patient with paresis, and Graves reported positive results with blood from two pare- tics. Ellis’s experiments to confirm these findings gave con- stantly negative results. The above facts are in accord with the clinical teaching that the blood of Iuetic individuals is infectious during the active primary and secondary stage, and that this diminishes with the PROBLEMS IN THE PATHOLOGY OF SYPHILIS 233 age of the disease. Exceptions to the innocuousness of the blood of latent syphilitics are seen in the occasional infection of a surgeon in his operative work on an individual in this period of the disease. It is to be inferred therefore from clinical and laboratory experience that, under conditions of which we are at present ignorant, spirochetes may reach the circulation from some focus in the body. In certain cases this may precede the cutaneous relapse, in others it is accompanied by no visible manifestations. Kraus has called attention to febrile attacks as the only symptom of latent lues, and suggests the spiro- chetemia as a possible explanation. An analogy is found in trypanosomiasis where the parasites may disappear spontane- ously from the blood and, reappearing, give rise to an eruption and fever. Moreover, a latent period of three years may ensue before the nervous symptoms appear. PATHOLOGICAL ANATOMY The predilection of syphilis for the vascular system is noted almost from the inception of the disease in the involvement of the cutaneous vessels at the point of inoculation. Here the pathological process consists of an endarteritis, later a pan- arteritis, with a characteristic inflammatory infiltration. Ex- cepting the macule, all the lesions succeeding the chancre show a marked affection of the vessels, especially those of the late secondary and tertiary stages. The formation of giant-cells so frequently encountered in the secondary papule, the nodular syphilide and gumma may be traced to a vascular genesis, while the extension of serpigenous lesions may be explained by the progressive thrombosis of the vessels. A study of the pathological anatomy of syphilis shows that fundamentally the reaction is on the part of the fixed connective- tissue elements, the labile constituents coming into play locally only secondarily. The chief cells are the lymphocyte and the plasma cell, the latter believed to be a derivative of the former and the antecedent of the fibroblast. It has been shown by Hazen that a circulatory leucocytosis is present in untreated secondary cases, sometimes as high as 20,000 white cells. The 234 HARVEY SOCIETY neutrophiles are absolutely and relatively increased and the percentage of eosinophiles is higher. Under treatment there is a slight drop in the total count, with a marked relative increase in lymphocytes which under treatment may run as high as 65 per cent. The syphilitic process is essentially a granuloma having its origin in the perivascular lymphatic spaces. In the primary lesion the main changes are found in the cutis, the very earliest of which are in the new formation of capillaries and the group- ing about these and pre-existing vessels of lymphocytes and plasma cells. At first they mantle the vascular structures as a ‘‘coat-sleeve’’ infiltration, but as the lesion grows older, spread out and become diffuse. The endothelium of the capillaries is swollen and proliferated so that the lumen is narrowed or alto- gether occluded, and in the larger vessels with an external coat there is in addition evidence of inflammation and an in- crease in thickness. In some instances giant cells are found. From the newly-formed granulation tissue, connective tissue is formed which later scleroses and leads to induration. Owing to interference with nutrition, regressive metamorphosis sets in. The epidermis presents a varied picture, depending on whether there is pressure from the infiltrate or retrograde changes with erosion and ulceration. In the secondary stage the disease is characterized by a succession of eruptions and a general adenopathy. Ehrmann has said that the distribution of secondary syphilides is gov- erned by the branching of the vascular stems. The roseola or macular syphilide under the microscope shows very few changes, being simply an erythema with dilatation of the vessels of the papillary body and adjoining corium and an infiltration of lymphocytes and plasma cells about them. In the papular or lenticular syphilide we find in the cutis a circumscribed lesion made up of lymphocytes, plasma cells, and proliferated fibroblasts, all in close relation with the vessels which show the characteristic changes. Lichen syphiliticus owes its peculiar features to localization, namely, distribution about the pilo- sebaceous apparatus. Here the process surrounds a hair follicle PROBLEMS IN THE PATHOLOGY OF SYPHILIS 235 and extends deeply into the corium. Its structure is lke that of other lesions, with usually abundant giant cells of vascular - origin. The epidermis in all lesions of this stage shows only secondary changes. It may be thinned from pressure of the granulomatous tissue beneath or there may be cedema with an increase in thickness and scaling. In condylomata papillo- matous overgrowth is a marked feature. Pustular and suppu- rating syphilides usually signify extraneous inoculation with pus-producing organisms. The pigmentary syphilide, or leuco- derma syphiliticum, according to Ehrmann, is due to chromato- phores, the pigment passing from them to the basal layer of the epidermis. As a rule, secondary syphilides undergo spontaneous absorp- tion. Microscopic residua may, however, persist for a long time. Their connection with local relapses has been suggested. The type of lesion of the tertiary period is the gumma. The process consists of an infiltration of lymphocytes, plasma cells, and proliferated connective-tissue cells about the vessels, newly- formed and old, which are the seat of an endarteritis and pan- arteritis, and give rise to giant cells. Caseous degeneration usually begins in the centre of the granuloma, or in some cases it may be fatty or mucoid in character. The last mentioned is most often found in bones. The necrotic areas liquefy and - are absorbed or discharged, the usual procedure being gradual absorption, with formation of cicatricial tissue, the contraction of which leads to deformity. It is believed by many that the fibrosis is not purely a process of repair or due to the irritating action of necrotic tissue, but that there is also a syphilitic ele- ment. The plasma cell thought to be the precursor of the fibro- blast has been credited with being more than a passive partici- pant. The late nodular or tubercular syphilide is in reality a gumma, situated more superficially in the cutis. The ser- pigenous lesions which also belong to this stage are made up of groups of nodules situated about the cutaneous vessels, the thrombosis of which probably explains the progressive char- acter of the lesion, Syphilitic phlebitis is relatively infrequent, occurs usually 236 HARVEY SOCIETY in the veins of the lower extremities, and is of minor impor- tance. Arterial disease, however, is very common and of serious import, the vessels especially attacked being the aorta, the cerebral, pulmonary, subclavian, femoral, and popliteal arteries. Owing to disease of the walls, vasomotor response is impaired or lost, and where there are obliterative changes due to endothe- lial proliferation or secondary thrombosis, interference with the blood supply and impairment of nutrition. While the brain and the heart suffer more severely in this respect, other organs and the vessels of the extremities are not exempt. Several years ago the writer had under observation a case of gangrene of the leg consecutive to a syphilitic thrombosis of the femoral artery (Fig. 1). The patient was a colored girl, in the City Hospital, upon whom several amputations were made, after each of which the gangrenous process developed and spread. A picture closely simulating Raynaud’s disease is also met with as the result of luetic involvement of the vascular supply of the extremities, and in several cases of symmetrical cutaneous atrophy which I have had under my care the syphilitic element was the predominant feature. AORTITIS It is only within the last decennium that the syphilogenic nature of aortitis or mesaortitis has been generally recognized among internists and syphilographers. During this period the development of the Wassermann reaction and rontgenology, as supplementary to the physical examination, have shown us how widely prevalent the disease is among individuals who have had a syphilitic infection. As to its frequency of incidence, Chiari found the condition present in 59 per cent., Fahr in 29 per cent., and Fraenkel 53 times in 102 cases of constitutional syphilis. Stadler, whose statistics cover 256 cases, demonstrated the disease in 82 per eent., and of 211 of this number it was the cause of death in 117. Staub found an aortitis in 82 per cent. of paretics which came to autopsy, Buder in 84.5 per cent., and Alzheimer in 74 per cent. The findings of Rach and Wiesner show that in —| ‘ Courtesy, The American Journal of the Medical Sciences. lrg. 1.—Syphilitie thrombosis of femoral artery. Transverse section of artery, showing thrombus and thickening of arterial coats. Courtesy, The American Journal of the Medical Sciences. Fia. 2.—Syphilitic aortitis, showing characteristic changes in the walls of the vessels, transverse rupture, and dissecting aneurism into coats of the aorta. (Specimen of Dr. John H. Larkin.) PROBLEMS IN THE PATHOLOGY OF SYPHILIS 237 congenital syphilis, changes in the aorta or pulmonary artery are seen in 67.4 per cent. Lenz states that in large cities 25 per cent. of all syphilitics die from aortitis (angina pectoris, aortitic insufficiency, aneurism) as against 3 to 4 per cent. from paresis, 1 to 2 per cent. from tabes, and 10 per cent. from all other syphilitic affections, as of the brain, liver, kidneys, etc. Mesaortitis has been defined by Doehle, Heller, and others as a specific inflammation of the adventitia and media which terminates in cicatricial deformity. As a rule, there is a com- bination of processes so that cicatrization may be present in one part while active inflammatory lesions are found in an- other. The disease shows a decided preference for the ascending portion of the aorta and the arch, the explanation being that the impact of the blood is greater here than in the rest of its extent, following the well-known law that syphilis localizes where trauma has produced a locus minoris resistentiw. In contradistinction, atheroma selects the lower portion of the aorta where the mechanical element is of slighter moment. Stadler could find no relationship between an increased blood- pressure or marked variations in pressure. Striimpell favors a summation of injurious agencies, as the noxious influence of aleohol and tobacco, with consequent lowering of resistance of the vessels. Grossly the aorta gives a characteristic picture (Fig. 2). The inner surface shows isolated or confluent elevated, wrinkled, grayish, and translucent sclerous areas, with puckering or cicatricial pitting, or again radiating ridges with depressed cicatrices between. The intima over the affected areas pre- serves its glistening appearance. The thickness of the artery varies in different portions, so that while indurated and scle- rotic at one point it is thin and translucent at another. De- pending on the extent of involvement, dilatation is present. This may be uniform throughout the length of the thoracic portion or appear as small aneurismal pouchings or a true sac- cular aneurism. All transitions are found. Ordinarily it is not difficult to differentiate this condition from arteriosclerosis in 238 HARVEY SOCIETY the absence of fatty degeneration and calcification. However, in old subjects there may be a combination. In the beginning the lesion is often confined to the vessel in proximity to the aortic valves. Later it extends to them and gives rise to insufficiency, which condition is very frequent. With the extension of the process the ascending aorta and the arch are involved. Not rarely changes are also noted about the mouths of the larger branches of the arch, causing narrowing of the left carotid, innominate, and subclavian arteries, which localization is significant from a clinical point of view. With involvement of the aortic ring the mouths of the coronary arteries are also considerably reduced. The relation of syphilis to aortitis was for a long time dis- puted. On the one hand investigators classed it with para- syphilis because of the difficulty of demonstrating spirochetes, its indurative or fibrotic tendencies, and indifferent results from treatment. On the other, which has the support of recent exam- inations, it was classed with active syphilis, The demonstration of spirochetes in the aortic wall by Reuter, Schmorl, Wright, Richardson, and others in acquired syphilis, and by Wiesner and Rach in congenital, together with the microscopic picture of gummatous lesions, and a positive Wassermann reaction in about 80 per cent. of cases, definitely places the affection in the category of active syphilis. It is difficult to demonstrate the organisms even in comparatively recent and active lesions. This, however, does not militate against their syphilitic nature, as they are equally difficult to demonstrate in cutaneous gummata. In both, the tissue is probably in a state of altered reactivity, so that numerically few organisms are capable of bringing about the necrotic change. The striking changes are in the media, although the disease begins in the adventitia. Miliary gummata undergo necrosis and are replaced by cicatricial tissue. In the adventitia in recent processes an inflammatory infiltrate of lymphocytes and plasma cells is localized about the vasa vasorum (Fig. 3), while in older ones only fibrotic changes may be left (Fig. 4), the nutrient vessels themselves being the seat of an obliterating endarteritis, such as is found in syphilitic Courtesy, The American Journal of the Medical Sciences. Iie. 3.—Early stage of aortitis, showing characteristic lymphocytic and plasma-cell infil- tration about the small vessels in the adventitia. Courtesy, The American Journal of the Medical Sciences. Fia. 4.—Later stage of aortitis, showing advanced sclerosis of the vessel walls. al of the Medical Sciences. The American Journ Courtesy, Fig. Posterior ningitis. gme anyin d. olumns not involve rior nerve root, with accomp c ste po Insular sclerosis 5. American Journal of the Medical Sciences. Phe Courtesy, solumns involved, rior ¢ oste Bee root in tabe ited posterior nerve -Degeners Ita PROBLEMS IN THE PATHOLOGY OF SYPHILIS 239 lesions elsewhere in the body. The elastic tissue, and this is characteristic of luetic disease of the larger vessels, is either fragmented or has entirely disappeared. The intima shows no change at all or various grades of a compensatory thickening. On the question of genesis of the early changes there is some dissonance of opinion. Doehle, Heller, Backhaus, Saathoff, and others trace the beginning of the disease to the vasa vasorum. They claim, primarily, a swelling of the endothelium of these vessels in the middle third of the media to large epitheloid cells, which lead to an obliteration with secondary regressive meta- morphosis of the aortic wail. Marchand, Beneke, Arnsperger, and others not corroborating these findings incline to the view that the medial changes are primary, the virus so injuring the elastic lamelle that they succumb to the pressure of the blood. The difficulty of determining the genesis with positive- ness even early is apparent, for, as Stadler points out, a simple infiltration of the vasa vasorum may also occur under the influence of the toxins of alcohol and other infections. Benda describes two forms of aortic disease. In the one submiliary lesions without necrosis which cicatrize and leave none or only a very slight deformity of the vessel wall, in the other typical gummata, with central necrosis, the scarring of which leads to typical sclerotic changes in the wall. The clinical diagnosis of aortitis is often extremely difficult. Limited and circumscribed changes of the ascending portion without involvement of the ring are, as a rule, not accompanied by subjective or objective manifestations. Such patients may be observed for a long time without eliciting signs pointing to the condition, Réntgen-ray examination also being negative. Only when a greater area is involved, especially more or less of the circumference, are there symptoms suggestive of disease, namely, pain, dyspnea, and tachycardia on slight exertion, in- creased blood-pressure, weakness, and easily-induced fatigue. The pain is often the primary and predominant symptom, although frequently it is lacking or indefinite. It is character- ized as a dull aching or a feeling of pressure under the sternum, with pain radiating to the side of the chest, the back, and the 240 HARVEY, SOCIETY arms. Disease of the innominate, carotid, and subclavian may be attended by a condition simulating intermittent claudication of the lower extremities, with weakness, radiating pain, and sensory disturbance. Periodic or continuous headaches, often with vertigo, accompany involvement of the mouth of the left carotid, and where the coronary arteries are the seat of disease the attacks are anginal in character. There appears to be no satisfactory explanation for the origin of the pain. Thoma referred it to injury of the larger and smaller nerve trunks as well as irritation of the Pacinian bodies in the aortic wall. Huchard attributes it to a peri-aortitis, and the neuralgic pains in the chest and shoulder to the changed relation of the sub- clavian to the brachial plexus. Longeope came to the conclusion that the anginal attacks were due to reflex disturbances set up by the syphilitic process involving the root of the aorta, the paroxysmal dyspneea being regarded as an acute bronchiospasm. When does the aortic disease begin? Obviously this is diffi- cult to affirm, as the disease in the majority of cases is insidious in its onset, the first symptoms appearing only when the affection has made some progress. In assuming that involvement takes place at the close of the primary stage when there is a general dissemination of the spirochetes, we are confronted with the same problem as when we endeavor to establish the time of infection of the central nervous system. Palpitation, arrhyth- mia, tachycardia, and disturbances in the pulse-rate frequently occur in early secondary syphilis, which has been adduced as evidence that involvement takes place at the time the disease is a spirochetemia. The publication of fatal cases shortly after infection also inclines to the view that involvement takes place early. Brooks recorded perforation of an aneurism before the secondary rash had fully appeared, and another case with a fatal aortic lesion within six months of infection. Longeope cites two patients who died from the effects of syphilitic aortitis four years after the appearance of the chancre. Stadler pub- lishes two cases in whom death occurred five years after infection. The first patient, aged twenty-nine years, showed at autopsy a fibrous aortitis and aneurism. The second, aged twenty-seven PROBLEMS IN THE PATHOLOGY OF SYPHILIS 241 years, the objective and subjective signs of an aortitis. He further studied 12 post-mortem cases of severe secondary and tertiary syphilis in patients under thirty years of age. He assumed on account of the youth of the subjects that the infec- tion could not have been of long standing, but in none were changes of syphilitic aortitis demonstrable. The writer has recently investigated, in his private practice, all latent syphilitics, especially those with a persistent positive Wassermann, for the occurrence of cardiovascular disease, and in a large percentage has found the existence of such a lesion. The average age at which the disease appears is forty-seven to forty-nine years, but it is not rare to find it in much younger patients. Aortic disease in general makes slow progress, for Weintraud, Stadler, and others give the average lapse of time as twenty years, the interval vacillating between five and forty years. The disease usually runs a fatal course in about two years after the development of the symptoms. The importance of early diagnosis is, therefore, apparent, especially in view of the fact that the affection is a manifestation of active syphilis and amenable to treatment. In 95 cases of aortitis collected by Benda the termination was as follows: in 9, coronary stenosis ; in 22, aortic insufficiency ; in 37, combined coronary stenosis and aortic insufficiency ; in 27, aneurism. In a total of 248 cases he found aneurism forty-eight times. The relation of cardiovascular affections to tabes has for a long time occupied the attention of clinicians. According to Lesser’s statistics in 96 cases of tabes come to autopsy, aneurism of the aorta was found 18 times. A review of the literature shows that the coincidence of tabes and aortic insufficiency is most frequent. Rogge and Ruttner report the association in 6.5 per cent.; Rogge and Miiller in 10 per cent.; Stintzing in 3.8 per cent., and Stadler in 6.2 per cent. of all cases of well- developed tabes. The latter did not include his cases of incipient tabes; in manifest tabes which came to autopsy, however, he found disease of the aorta in almost all. Rogge and Miiller have called attention to the cause of sudden deaths in tabeties as not 16 242 HARVEY SOCIETY being due to lesions referable to the central nervous system but to sudden cardiac insufficiency or rupture of an aneurism. The so-called cardiac crises are in reality attacks of angina pectoris, due to changes in the coronary arteries. Striimpell has emphasized the surprisingly frequent pres- ence of rudimentary tabes in patients who seek medical advice for symptoms referable to the heart. Many of these cases show only pupillary changes, such as narrowing, irregularity, and fixation to light. Only one pupil may be affected, as in a case cited by him with severe endocarditis, myocarditis, and aortic insufficiency. The reflexes are only slightly altered or the Achilles reflex alone lost and the patellar reflex exaggerated. Observing these cases over a long period of time, weakening and gradual disappearance of the reflex may be noted. In general, symptoms of aortic disease appear later than the earliest tabetic manifestations. Rogge and Miiller are of the opinion that tabetic symptoms average four and a half years earlier than those of the circulatory apparatus. It is probable that it is not a later involvement, only that the condition has a longer latency and therefore manifests itself later. The writer has under treatment at the present time several patients with tabes and concomitant aortitis. One of them, with optic atrophy, has an aneurism. Five others, in whom the tabetic symptoms have existed on an average of eight years, show a marked sclerosis of the aorta. In a paretic, Réntgen- ray examination shows a dilatation of the ascending aorta and in a case of cerebrospinal syphilis, with cardiac symptoms de- veloping four months ago, a small aneurism. In another patient, aged thirty-one years, who had had an attack of hemiplegia four years ago, signs of aortitis have developed within the past few months. The Rontgen-ray plates also show a marked sclerosis. For the following interesting post-mortem examination and also some of the illustrations which accompany this article I am indebted to Dr. John H. Larkin, Director of Pathological Laboratories, Department of Charities, New York City: PROBLEMS IN THE PATHOLOGY OF SYPHILIS 243 J. E., aged forty years. The patient was admitted to the City Hospital (service of Dr. Evan Evans) the latter part of 1913 for sciatica. After a few weeks he was discharged and five months later readmitted, at which time in addition to the sciatic pains he complained of girdle pains. Lumbar puncture showed 38 cells, an excess of globulin, and a strongly positive Wassermann. His serum was also positive. A diagnosis of tabes without clinical evidence, excepting the girdle pains, was made. Further examination, including Réntgen-rays, revealed an aneurism of the ascending aorta. He was placed on salvarsan and mer- curial treatment, which reduced his cell count and reversed the Wasser- mann in the fluid, but had no effect on that of the blood. The patient died suddenly from a rupture of the aneurism into the thoracic cavity. Micro- scopically the aorta showed a characteristic picture with a round-celled infiltration in the adventitia. The interesting feature of the finding in this case was in the cord, which showed an involvement of the pia- arachnoid with a round-celled infiltration about the small vessels and an insular degeneration of the posterior nerve roots (Fig. 5). The posterior columns were unaffected. NERVOUS SYSTEM With the acquisition of more exact knowledge our ideas concerning syphilis of the nervous system have undergone con- siderable revision. The present views based on clinical and laboratory investigations may be formulated as follows: Infec- tion of the central nervous system probably takes place in the early stage of the disease with the generalization of the virus. Accumulated clinical evidence points to meningeal involvement during the first few months, in some cases before the appearance of the cutaneous eruption. Lumbar puncture in the hands of different investigators has elicited varying results. Thus, Ravaut found the spinal fiuid in secondary syphilis abnormal in 67 per cent. of cases, Altmann and Dreyfus in 78 per cent., Nonne in 40 per cent., and Swift and Ellis in 36 per cent. The writer’s findings, based on the examination of cases in the sec- ondary stage with and without cutaneous lesions, showed menin- geal involvement in less than 20 per cent. Fournier’s dictum that patients in whom secondary symptoms are overlooked or very mild later show involvement of the central nervous system has been reiterated again and again. This led Finger to divide syphilitics into two classes: those with manifest cutaneous 244 HARVEY SOCIETY lesions, and those with absent or mild skin lesions and meningeal involvement, which division is too arbitrary, for many cases with severe cutaneous lesions as well as those which have been treated early later develop disease of the nervous system. The observation of patients over long periods of time has shown that a low-grade meningitis may exist for years without produc- ing obtrusive nervous symptoms. Such lesions are comparable to the superficial lingual and palmar syphilides, which persist for years, produce little inconvenience, and are refractory to treatment. There has been a good deal of speculation as to whether the spirochetes producing lesions in the central ner- vous system are the remains of those deposited during the spirochetemia or whether they later reach the nervous system from another focus. Head’s work is perhaps suggestive. In seeking to explain the more frequent involvement of some areas in root affections over others, he found that the roots most com- monly subjected to irritation were those in connection by their visceral afferent fibres with certain organs known to be the seat of active spirochetosis in syphilis. Thus the second and third cervical contain afferent paths from the tonsil; the first, second, third, and fourth thoracic from the aorta, while those from the seventh thoracic to the first lumbar, the most frequently in- volved, carry afferent paths from the liver, kidney, suprarenal, and testicle, organs which are the sites of election of the parasite. Orr and Rows have shown in their investigations on the lym- phogenous infection of the nervous system that organisms and their toxins travel along the perineural lymphatic space which surrounds every spinal nerve and extends along the roots to the pia mater. These spaces are not only in connection with the pia, but through the latter in communication with the adven- titial lymph spaces of the perforating vessels. The réle of these channels as distributors is therefore quite obvious. The sharp distinction formerly drawn by clinicians and laboratory workers between cerebrospinal syphilis and para- syphilis is no longer tenable. Gradually the barrier has given way until we have come to regard these conditions as identical from an etiological stand-point, though differing in reaction, ——s = — Courtesy, The American Journal of the Medical Sciences. Fig. 7—Syphilitic meningitis, showing various stages of obliterating endarteritis. Courtesy, The American Journal of the Medical Sciences. Fig. 8—Mantling infiltration of lymphocytes and plasma cells about pial vessel in brain cortex. Me pcan) ourtesy, The Americ Fia. re an Journal of the Medical Sciences. syphilitic meningitis. sin rating endarteriti ).—Heubner type of oblite Q Medical Sciences y, The American Journal of the Courte ar spaces rivasculs g infiltration of pe showin brain, Small vessel in cortex of the 10.— Via. a cells. lymphocytes and plasm with PROBLEMS IN THE PATHOLOGY OF SYPHILIS 245 which difference depends on localization and the tissue involved. By cerebrospinal syphilis we understand the exudative, vascu- lar, and gummatous processes which involve the coverings of the nervous system and the blood-vessels within them (Figs. 7, 8, 9 and 10). These processes in the great majority of cases remain superficial, but in some cases extend into the essential nervous structures along the pial or adventitial sheaths of the vessels, and give rise to the borderline cases of tabes and paresis, It is proposed by Head and Fearnsides to divide luetic disease of the central nervous system into syphilis meningo- vascularis and syphilis centralis, the latter including all those eases where degeneration of nerve tracts or nuclei shows that the lesion must lie within the structure of the nervous system itself. As meningitis and endarteritis, however, are common to all these pathological states, this classification does not seem to the writer a tenable one. What is the nature of paresis and tabes? In paresis there is a combined meningitis and encephalitis with a typical infiltra- tion of lymphocytes and plasma cells in the adventitial lymph spaces (Figs. 11 and 12), the secondary degenerative changes probably depending on the primary vascular disease. Likewise we have a meningitis in tabes, but it is easier to comprehend the pathological condition and manner of invasion in paresis than the tract degeneration in the cord. The view that the genesis of both affections lies in a chronic meningitis now has a pretty wide acceptance. In tabes, Nonne believes it leads to disease of the roots and secondarily of the posterior columns. Nageotte was the first, in 1894, to describe a low-grade menin- gitis at the junction of the anterior and posterior roots on the proximal side of the ganglion, and in addition a neuritis, as shown by an inflammatory condition in the perineurium and a perivascular infiltration of plasma cells and lymphocytes in the posterior roots. This was confirmed by Dinkler, Dejerine, and others, but their views were not generally accepted, as the findings were interpreted as a combination. Again, in 1906, Schréder showed that collections of lymphocytes and plasma cells occur in the lymph spaces of the vessels in paresis and 246 HARVEY SOCIETY tabes, and that in the latter they are found not only in the pia and connective tissue about the vessels in the peripheral parts of the cord, but in the intramedullary portion of the columns as well. Stargardt’s findings in optic neuritis also have an important bearing. In a number of cases of tabes and paresis he found a true inflammation of the sheath and endoneurium. He also described an active inflammatory process in the joint structures of tabetic arthropathy, as evidenced by plasma cells and lympho- eytes, with endarteritis of the small vessels. Steiner’s studies on peripheral nerves in tabes and paresis also brought him to the conclusion that the process is an inflammatory one. The older view of a primary neuron degeneration in the light of present evidence would seem to have to be abandoned. According to McIntosh, Fildes, Head, and Fearnsides, para- syphilis ‘‘depends upon an anaphylactic reaction in the tissues of the central nervous system which have been rendered hyper- sensitive in accordance with their peculiar lymphatic distribu- tion.’’ Since it has been shown by Marie and Guillain, Orr and Rows, and others that the posterior columns are in direct lymphatic communication with the posterior roots, the afore- mentioned authors have evolved the theory that sensitization takes place during the secondary period by the spirochetes or their toxins ascending the afferent peripheral nerves. When spirochetes again become active in the posterior columns, a violent reaction may ensue, injuring the nerve fibres and pro- ducing proliferation of the neuroglia and wandering cells. The injury of the nerve fibres will lead to degeneration, which will continue as far as the reflection of the neurilemma. In the same way the cerebrum or isolated groups of cells in the anterior horns or other groups of fibres or cells alone or in combination may be affected, the particular region involved determining the clinical picture. Noguchi had previously shown by animal ex- periments that sensitization may be necessary before spirochetes ean infect the brain. In both rabbits and monkeys he obtained negative results when the organisms were inoculated directly into the brain. He therefore injected rabbits with living and ey ~ * as 2 ye \ ~ Sip a é : ‘ » * é 4 Ree. r; ate : a.) alee 2 y a. . % * ¥ _ oy sa] oad c , ty 3 = E * « 5 : * Big, * : ®, : : ast és iy a u poet apt 8 # 5 a eh S. i. =e ae ae . ; nL. atts ee Courtesy, The American Journal of the Medical Sciences. Fie. 11.—Showing the characteristic picture of paresis, with plasma-cell infiltration. | ge : ; oes “4 2 >” Puy 2% ee D xb «] > on tgs ‘a > mong ‘ a2 SSehe a Pe yo we, . IN *) Fe MS ye i ea ¥ 2 Ction 4 fs ea S Ce % $F geet SY * aoe. “ , y Reet “5” "os Brin * met ae" SS ORES PRs oo * . Ri ‘aaere a aS : * ween. a ae & ee ae oe aA , = ; PF Ns Ss A d: “9 - ve 9 4 a, o>. ios) ’ ® es oon we ".. ri — . “ re “e ' ™ . ¥ , S « Fag a 7 poly ‘y “ 2 | PER Syed ere « F | es Se 2 om € e ie ¢. a? | Ron el ee os | " ~ & > 7 erg | (les * = ee oe g ; a seal | Courtesy, The American Journal of the Medical Sciences. Fig. 12.—Paresis, showing the grouping of lymphocytes and plasma cells about small vessels in cerebral cortex. PROBLEMS IN THE PATHOLOGY OF SYPHILIS 247 dead spirochetes intravenously, and after five months inocu- lated them subdurally with a particle of a scrotal syphiloma rich in spirochetes. They all remained well for two months, then became emaciated, and showed a slightly ataxic and spastic gait. The symptoms continued to increase, and after five months the animals were no longer able to jump. Microscop- ically some of these animals showed identical changes with paresis, but the nerve-cells were intact. In three an exudative meningitis was found and in one a unilateral atrophy of the frontal lobe. The control animals without previous sensitization remained well. Analogies are frequently drawn between tabes and paresis and sleeping sickness. Spielmeyer, as the result of extensive investigations, came to the conclusion that in incipient cases the diagnosis between paresis and trypanosomiasis is not pos- sible. He was able to produce in dogs with Trypanosoma brucei an involvement of the central nervous system which was like human tabes. After nine to ten weeks the process was found in the posterior roots with beginning extension to the posterior columns, sensory trigeminus, and optic nerve. In trypano- somiasis, however, the central nervous system rarely escapes, whereas in syphilis only 1 or 2 per cent. develop tabes and 4 to 5 per cent. paresis. CONCLUSIONS Clinical observation and experimental work have shown that no true immunity exists in syphilis, but that an anergy develops during the first incubation period and is complete at the time of the general eruption. This refractory state usually persists as long as the body harbors spirochetes, and when it disappears, with the cure of the disease, the patient is again in a receptive state. ; In the transmission of syphilis to the offspring the theory of spermatic infection has been abandoned in favor of placental infection, as the majority of mothers of congenitally syphilitic children have a positive Wassermann or show spirochetes in the intervillous spaces and maternal portion of the placenta. 248 HARVEY SOCIETY The histopathology of syphilis is a uniform one. In all stages and in all organs the lesion begins in the perivascular lymph spaces as a lymphocytic and plasma-cell infiltration. The distribution of secondary lesions, according to Ehrmann, is due to a branching of the vascular stems. In the tertiary stage the ulcerative and destructive character of the lesions finds its explanation in a changed tissue reaction, the Umstim- mung of Neisser, or allergy, in the sense of an increased suscepti- bility to the action of the organisms. Syphilis produces a characteristic type of aortitis and is very common in individuals who have had the infection. It may be the only lesion present in so-called latent lues with a per- sistent positive Wassermann, and emphasizes the importance of examining all such patients for possible cardiovascular in- volvement. It is a frequent concomitant of paresis and tabes, especially the rudimentary form, occurring in about 80 per cent, of parasyphilitics. Infection of the nervous system probably takes place during the secondary stage. A low-grade meningitis may exist for years or the spirochetes may remain quiescent for years until called into activity. In paresis a mantling infiltration of lymphocytes and plasma cells is found in the adventitial lymph spaces of the meninges and encephalon. From its pathological anatomy, which has been studied in all stages, it is much easier to comprehend this condition than the tract degeneration in tabes, as tabetics seldom die during the early period of the affection. The few pathological examinations which have been made reveal, however, changes in the meninges about the pos- terior roots between the ganglion and the cord. It is not so difficult to understand how an infiltration in this region by pressure could produce degeneration of the afferent fibres extending to the posterior columns and leading to an ascending tract degeneration following the well-known law that a destruc- tion of the neuron is followed by degeneration along its distribu- tion. This explanation is a much more plausible one than a primary degeneration without inflammatory manifestations. PROBLEMS IN THE PATHOLOGY OF SYPHILIS 249 BIBLIOGRAPHY Altmann and Dreyfus: Miinchen. med. Woch., 1913, pp. 464 and 531. Andrews: System of Syphilis, i, 113. Benda: Die Gefiisse, Path. Anat. Aschoff, Bd. ii. Spec. Teil. Benda: Verhandl. der deutsch. path. Gesellsch., 1903, 164. Chiari: Verhandl. der deutsch. path. Gesellsch., 1903, 137. Finger and Landsteiner: Untersuchungen iiber Immunitit bei Syphilis, Verhandl. d. deutsch. dermat. Gesellsch., 1906, 251. Friihwald: Ueber die Infektiositiit des Blutes im latenten Stadium der erworbenen Syphilis, Derm. Woch., 1914, lix, 1319. Hazen: The Leucocytes in Syphilis, Jour. Cut. Dis., 1913, xxxi, 618. Head and Fearnsides: The Clinical Aspects of Syphilis of the Nervous System in the Light of the Wassermann Reaction and Treatment with Neosalvarsan, Brain, 1914, xxxvii, 1. Hidaka: Zur Frage der Beziehungen zwischen Syphilis und Recurrens Immunitit., Zeitschr. f. Immunitiitsforsch., 1913, xvii, 448. Krause and Volk: Untersuchungen tiber Immunitit bei Syphilis, Ver- handl. d. deutsch. dermat. Gesellsch., 1906. Lenz: Med. Klinik, 1913, 939. Levaditi: Zeitschr. f. Immunitiitsforsch., 1910, ii, 277. Longeope: Syphilitic Aortitis: Its Diagnosis and Treatment, Arch. Inter. Med., 1913, ii, 15. McIntosh and Fildes: A Comparison of the Lesions of Syphilis and Para- syphilis, Brain, 1914, xxxvii, 141. McIntosh and Fildes, Head and Fearnsides: Brain, 1913, xxxvi. Neisser: Beitriige zur Pathologie und Therapie der Syphilis, 1911, Julius Springer. Noguchi: Presse Med., 1913, No. 81. Nonne: Die Lues-Paralyze-Frage, Arch. f. Derm. u. Syph., 1914, exix, 215. Orr and Rows: Lymphogenous Infection of the Nervous System, Brain, 1913, xxxvi, 271. Ravaut: Ann. de derm. et de Syph., Quatriéme Serie, 1903, v, 1. Schindler: Die paterne Uebertragung der Syphilis auf die Nachkommen- schaft, Arch. f. Derm. u. Syph., 1912, exiii, 935. Stadler: Die Klinik der Syphilitischen Aortenerkrankungen, Jena, 1912. Striimpell: Ueber die Vereinigung der Tabes Dorsalis mit Erkrankungen des Herzens und der Gefiisse, Deutsch. med. Wochenschr., 1912, 1931. Swift and Ellis: Forschheimer’s Therapeusis of Internal Diseases, v, 401. Uhlenhuth and Mulzer: Beitriige zur experimentallen Path. u. Ther. der Syphilis, Berlin, 1913; Berl. klin. Woch., 1913, No. 17. Weber, F.: Die Syphilis im Lichte der modernen Forschung., S. Karger, Berlin, 1911. STRUCTURE AND RELATIONSHIPS OF THE ISLETS OF LANGERHANS * CRITERIA OF HISTOLOGICAL CONTROL IN EXPERIMENTS ON THE PANCREAS PROFESSOR R. R. BENSLEY University of Chicago HE relationship between the islets of Langerhans and the rest of the pancreatic parenchyma is of great interest to the histologist, not only because of its significance from the stand-point of the localization of the anti-diabetic function of the pancreas, but also because the question of the functional significance of differentiation, as it is understood by histologists, is at stake. If it should prove, as many investigators have claimed, that the appearance of differentiation and specializa- tion, both cytological and histological, presented by the islets of Langerhans when compared with the acinous tissue of the pancreas, is but a mask behind which lurks a substantial funce- tional identity, then differentiation as indicated by structural characters becomes a delusion, and the effort to analyze further the reciprocal and independent functions of the cellular con- stituents of mixed glands a mere waste of energy. Hence, it seems to me to be worth while to review briefly the salient features of research in this field, and to point out some of the sources of the misconceptions and contradictions which have, from time to time, obstructed its progress. Although the islets of Langerhans were discovered as early as 1869, the twenty years which elapsed between this discovery and the announcement of v. Mering and Minkowski’s experi- ments on the effect of pancreatic extirpation are notable only for Kiihne and Lea’s discovery of the peculiar glomerule-like * Delivered February 27, 1915. 250 THE ISLETS OF LANGERHANS 251 capillary net, now a familiar object to every histologist, and for Lewaschew’s experiments, to which reference will be made again, on the effect on the number of islets of excessive activity of the pancreas. , V. Mering and Minkowski’s discovery that complete re- moval of the pancreas was followed by a grave diabetes which soon proved fatal, stimulated renewed interest in the islets and, in the decade which followed the discovery, they were sought and found in the pancreas of representatives of all the vertebrate classes. Their embryonic development was studied in detail, and it was shown conclusively that they were epithelial in nature, derived from the primary pancreatic anlagen, and not, as some had supposed, mesenchymatic or nervous structures. Particularly active in this investigation were Laguesse and Diamare. The former, in a remarkable series of articles ex- tending from 1893 to the present day, not only recorded results of the investigation of the islets in many vertebrate groups, but studied in detail their development in the sheep, and devel- oped, on the basis of his original observations, a theory of their relationship to the other secreting elements of the pancreas, which, while similar to that of Lewaschew’s to the extent that it admitted reciprocal transformations of the islet and acinous tissues respectively, claimed for the first time a specific activity of the islet cells in internal secretion. On this account Laguesse suggested that a search of the pancreas in cases of human diabetes would reveal changes involving the islets of Langerhans. Diamare also supported the theory that the islets were engaged in internal secretion. The grounds for these assumptions by Diamare and Laguesse were purely anatomical. The search for duct connections and for lumina in the cell cords of the islets of mammals had been for the most part unsuccessful. On the contrary, the char- acter of the islet complex and the provisions made for a rich blood supply suggested an important secretory function, the outlet for which could only be provided by the blood stream, since duct connections were apparently not present. Meanwhile, the production of diabetes by total removal of 252 HARVEY SOCIETY the pancreas, now a commonplace of the physiological laboratory, had been confirmed by Hedon, Gley, Léepme, Thiroloix and others, and Hédon, by showing that a small portion of the pancreas detached from the bowel and transplanted under the skin would protect against the onset of glycosuria, but that when this grait was removed a glycosuria equal in degree to that produced by total pancreatectomy immediately followed, had shown that the result was not due to the cutting off of the external secretion and so to the impairment of intestinal diges- tion, but to some direct influence of the pancreas, exerted through the medium of the blood-stream—in other words to an internal secretion. What could be more natural than to attribute this internal secretion to the very elements in the pancreas, which, anatomically considered, were specialized for that very function? The islet theory of diabetes, thus founded, acquired great favor among physiologists, pathologists, and anatomists alike, but efforts to demonstrate it were unsuccessful until Ssobolew and Schulze published the results of their investigations on the changes in the pancreas which resulted from ligation of the ducts. These observers found that, if the outflow of the pan- creatic juice were thus prevented, the acinous tissue degenerated with great rapidity, leaving, after a certain lapse of time, a gland in which the only secreting elements were islets of Langer- hans. The animals showed no glycosuria; therefore, they con- cluded, the islets were able, by themselves, and without the aid of the acinous tissue, to mediate the antiglycosuric function. In 1901 Opie published the results of his interesting and important investigations of the pathological conditions in the pancreas after death from various causes, and reported cases of diabetes in which the chief or only lesion involved the islets of Langerhans. Thus, at this period, it seemed as if the paramount impor- tance of the islets of Langerhans in respect to carbohydrate metabolism of the body was conclusively demonstrated. Contrary results, however, were not long wanting. In- vestigations of cases of human diabetes revealed a large number THE ISLETS OF LANGERHANS 253 in which no lesions of the islet could be detected, and studies of non-diabetic pancreases showed some in which the islets were profoundly changed. Wright and Joslin found islet changes in but two out of nine autopsies on diabetic patients, Herzog studied five cases, in one of which the islets were completely absent, in another degenerated, and in the remaining three affected to some extent. Hansemann, on the other hand, found in 34 cases only six in which the islets were degenerated and even in those six a certain number of normal islets were found. Sauerbeck, in a general summary of the pathological findings of various observers published prior to 1904, found that, in 40 out of 157 cases, the islets were reported as normal. Heiberg, while admitting the frequent occurrence of degenerations, con- siders that the greatest stress should be laid on the quantitative reduction of the islet tissue, a conclusion which he supports by eareful counts of islets in normal and diabetic cases. Weichsel- baum claims that in the cases where other observers have re- ported normal islets, there is, in reality, islet exhaustion evi- denced by diminution of islet granules, or by early or advanced hydropie degeneration. Cecil, in 90 cases, reported that seven- eighths of the total number showed pancreatic lesions, and that wherever lesions of the pancreas accompanied diabetes the islets were involved. Recently Major has reported thirteen cases, in seven of which the islets of Langerhans were normal in every respect. Thus, the hopes which were raised by Opie’s discovery have not been fully realized in subsequent investigations. The questions raised by Ssobolew and Schultze concerning the fate of the several pancreatic tissues after litigation of the duct have met a similar fate. The results obtained by different investigators are so contradictory that Allen in his book on ‘*Glycosuria and Diabetes’’ has divided them into four groups, viz., those who report survival of islet tissue alone, those who report survival of acinous tissue alone, those who report sur- vival of both tissues, and those who report survival of neither. The most serious attacks on the islet hypothesis, however, have been made by certain physiologists who deny to these structures any specificity whatever, and regard them as simply 254 HARVEY SOCIETY temporary variants of the acinous tissue, into which they be- lieve they may be readily transformed. Dale was the first to revive the Lewaschew theory that the number of islets increases during activity of the pancreas by transformation of acini into islets, and again diminishes during a period of rest by the converse process of change of islets into acini. He and the others who followed his lead were wholly undismayed by the difficulties involved in the peculiar blood supply of the islets and in the apparent lack of lumina and duct connections. Dale stimulated the pancreas for several hours by means of intra- venous secretin injections, and compared the organ, when so exhausted, with the normal resting gland. He claimed that islets in large number were produced by this method and that in addition large areas of the pancreas might be converted into islet tissue without rearrangement of the cells or of the duct and vasenlar connections. Since he interpreted the islet cell as an acinous cell which had lost its characteristic zymogen, any process which brought about exhaustion should result in increase of islets. In accord with this idea he found that starvation had the same consequences in this regard as stimulation. Vincent and Thompson confirmed these observations and in- vestigated the effect of starvation, which they reported to be more effective than secretin stimulation in increasing the number of islets. Laguesse reported similar effects of starvation in pigeons. This observer, however, regarded the transformation of acini into islets not as a result of exhaustion but as a regulatory phenomenon, since he maintained stoutly the internal secretory potential of the islet cells. He thought, however, that the internal and external secretory functions of the pancreas were possessed in like measure by both acinous and islet cells, and that the formation of islets was a sort of reversal of polarity of the cell for the purposes of internal secretion. He rested his argument, however, rather on the presence of intermediate types of cell and on the continuity between the two tissues. He supposed that there was some sort of physiological mechanism of control which determined from moment to moment the rela- THE ISLETS OF LANGERHANS 255 tive amounts of the two tissues which were needed for the ex- ternal and internal secretory purposes. Laguesse thus occupied a position intermediate between the opponents and advocates of islet specificity, since he accepted all the facts of the former while maintaining the conclusions of the latter. These views of the non-specificity of the islet tissue found great favor among physiologists, and especially among patholo- gists who, by reason of the character of their material, found no great difficulty in convincing themselves of the reality of transition forms. There was also doubtless a certain prejudice in favor of results obtained by experimental methods as com- pared with the results of anatomical study. Anatomists, how- ever, remained for the most part staunch supporters of the theory of islet specificity. Could any situation be more unsatisfactory? Investiga- tion, instead of illuminating the theme by the discovery of new facts, simply added to the confusion by throwing doubt on the old ones, and the researches of Dale and Vincent and Thompson seemed to question almost the reality of the islets themselves. Undoubtedly the factors which contributed to this confusion were many. Among them may be mentioned the system of experimental controls which had grown up, particularly in bac- teriology, but also in physiology, suitable enough, doubtless, where the question was the nature of the reaction of an animal to a pathogenic organism or to a toxin, or even to nerve stimu- lation or drug effect, but wholly unsuited to the investigation of variations experimentally produced in the relative amounts of the constituent elements of an organ. Here it is necessary, above all, to know what the normal range of variation is and how it is influenced by age, sex, strain, and conditions of rearing. A second factor, which was, doubtless, to a large extent, respon- sible for the favor accorded to Dale’s views, was the deceptive appearance of accuracy assumed by results obtained by quan- titative methods, even though these methods introduced doubtful factors in mathematical computation, and ignored the normal range of variation in the species and in the organ, and the con- 256 HARVEY SOCIETY ditions which influenced it, and involved the actual inspection of an almost infinitesimal fraction of the total tissue involved. The most important factor, however, was unquestionably the vague definition of the islet itself. Since their discovery by Langerhans, the cytological study of islet cells had made little advance. The majority of investigators still commend methods of study which emphasize the negative characters of the islet. Dale’s conception of the islet cell was a cell which did not con- tain the characteristic zymogen granules and basophile substance of the acinous cell, and hence could be distinguished only by location from a duct-cell. Of the host of workers who have con- cerned themselves with this study only Ssobolew, Tschassonikow, Mankowski, Laguesse, Diamare, in Europe, and a small group of recent investigators in America have concerned themselves with the really significant substances of the islet cells, namely their secretory granules. Under these circumstances the task of finding transitions between islet cells and acinous cells was an easy matter, for any reduction in the characteristic substances of the latter would, in the opinion of these workers, be a step in the direction of the other type. These considerations bring us to the question which is of the greatest importance in this controversy, namely, what would constitute a satisfactory definition of the islet cell. The answer to this is obvious, for it is clear that the only fully satisfactory definition would be the recognition of known chemical sub- stances in the cell by microchemical methods. This, however, is not possible, since we know neither what the characteristic chemical substances are nor how to detect them microchemically. The closest approximation to such a microchemical method is, in my opinion, the determination of morphological constituents of the cytoplasm by examination of the surviving tissue in media which are as nearly indifferent as may be, and the comparison of these constituents as regards solubility in reagents, refractive power, etc., with the cells of the ducts and with the acinous cells. Such a study should also furnish us with information on which THE ISLETS OF LANGERHANS 257 a satisfactory staining and fixing technic could be worked up. This would constitute the second phase in the investigation. The next desideratum in investigations on the pancreas is an accurate quantitative method for the determination of the total islet content, smce any method which involves computa- tions from fractional counts introduces so many sources of error, caused by the inherent difficulties of making the counts themselves from sections, the uncertain mathematical factors involved in the computations, and the great variations in local content, on the one hand, and of the relative count, on the other, introduced by the fiuctuations in bulk of the intervening tissue resulting from physiological activity, or from general conditions such as reduction of water content from active diuresis or purg- ing in the course of the experiment. Finally, the definition of the islet cell should be cytologically complete so as to enable us to recognize those more subtle changes which may have deep functional significance without involving a true degeneration of the cell. The first steps in this direction were taken, as already in- dicated, by Ssobolew, Tschassonikow, Mankowski and Laguesse who described the cells of the islet as containing many small granules, and by Diamare, Schultze, and Laguesse and T'schas- sonikow, who recognized that there were two kinds of cells in the islet. The significance of the last observation, however, escaped these observers because, as will appear more clearly later, they interpreted wrongly the second type of cell as an intermediate between islet and acinous. The investigations of Lane on the cytological characters of the islets marked a distinct step forward. Lane studied the islets after fixation in a larger number of different fluids and found that he could define the islet cells of both types by the way in which they behaved in these fluids—that is, by the solubility of their secretory granules in the solutions, and by their staining properties when fixed. In the same way he could distinguish between the islet cells of either type, on the one hand, and the acinous cells,on the other. He also described mor- phological differences between the two types of cells which he 17 258 HARVEY SOCIETY designated respectively A cells and B cells, the latter constitut- ing the great bulk of the islet. Lane’s division is based on the following characters: the zymogen granules for the most part, and all the granules of the B cells, are dissolved in alcohol, while the granules of the A cell are well preserved and may be in- dicated in the cell by the use of Bensley’s neutral gentian stain ; in chrome sublimate solutions the A granules are not seen but the zymogen granules and the granules of the B cells are well pre- served, stain in neutral gentian, but take a differential color; in picric acid fixations the zymogen granules are well preserved but both types of islet granules are lost. Thus Lane’s observations not only give strong grounds for maintaining the composite character of the islets and the specific nature of their secretory products, but also supply a much- needed technic for handling the pancreatic tissues, and for distinguishing their several constituents in sections. With the aid of these technics Cecil and myself independently attacked the questions raised by the claims of Dale, Vincent and Thompson. ‘The results of these investigations will be more fully discussed later, but meanwhile I may state that, although my own results were contrary to the claims advanced by these authors, as indeed were also those of Cecil, I did not feel quite satisfied that the methods gave sufficient information. The work of counting the islets, even though by Lane’s method they were sharply outlined, was so laborious, and the results so incomplete, that they did not seem to me quite as convincing as they should be. Consequently, I cast about for methods which would enable me to determine the whole islet content by direct methods without employing computations at all. The search for transitions also made apparent the necessity for combining in one method the advantages of the several methods of Lane for the demonstration of the islet and acinous constituents, and I therefore instituted experiments in fixing and staining the pan- ereas with this object in view, depending on differential stain- ing, instead of differential fixation, for my results. Further- more, new methods were needed for investigation of the relation of the islets to the ducts since the injection method and the THE ISLETS OF LANGERHANS 259 silver impregnation method had proven unsatisfactory, and since Laguesse claimed duct connections, which were denied by most observers with objective experience, though admitted by the advocates of variability, because they were a necessary con- sequence of their claims. These efforts resulted in the discovery of new methods for the study of the pancreas which, for the first time, in my opinion, give us full experimental control, since they enable us to examine the whole pancreas instead of a part of it, and to determine by direct count without mathematical computations of any sort whatever, and without cutting sections, the real con- tent of the pancreas in islets. I succeeded in securing two methods for staining all of the islets in the surviving pancreas by means of vital stains, thus permitting actual counts of all the islets in a pancreas of moderate size. Two methods were also devised for staining all of the duct system by vital dyes, and of combining in the same pancreas vital staining of the islets by one dye and of the ducts by another. By this means I was able to determine the true relation of the islets to the ducts and to bring to light a new mechanism of much impor- tance in the history of the organ. I was also successful in securing methods of fixing and staining which met the require- ments which I have indicated above. These various methods made it possible to repeat, under conditions which would give a definite answer to the questions involved, the experiments of Dale, Vincent and Thompson and Laguesse, as well as the duct ligation experiments of Ssobolew, Schultze and others. Before proceeding to a statement of the experimental re- sults, however, I shall have something to say about the methods themselves, for it seems probable that they may prove as useful in other experimental work involving histological control of the pancreas as they have been in the experiments mentioned. The first successful method which I discovered for the vital staining of the islets of Langerhans was by means of janus green. This dye, when injected in the form of a solution of 1 in 15,000 in normal salt solution into the aorta, promptly stains the whole pancreas blue. On inspection with a hand 260 HARVEY SOCIETY lens of low power, however, the islets may be seen as deeply stained particles in the midst of the pancreatic lobules. If, now, the pancreas be covered up, and air excluded, the tissue rapidly reduces and breaks down the dye to a safranin, which is red in color, but this change proceeds with so much greater rapidity in the acinous tissue than in the islets, that presently the latter are the only blue stained elements in the organ, the rest of the tissue being an intense carmine red color. If now a solution of ammonium molybdate be injected by way of the duct the preparation may be made permanent. By this means an entire pancreas of the guinea pig may be so prepared that every islet in it is deeply stained blue while the surrounding tissue is stained red. The fixation in molybdate of ammonium leaves the tissue still pliable and sufficiently transparent, so that one can easily see every islet in a fragment of pancreas without cutting a section. The preparation may be made a permanent mount by washing in ice-cold distilled water, de- hydrating as rapidly as possible with absolute alcohol, and clearing in toluol. By dividing such a preparation into a large number of small fragments and spreading these on slides, it was possible to count every islet in the pancreas, without count- ing any twice, and without making a section at any point, and to ascertain by weighing the fragments the ratio between number of islets and weight of tissue. The beauty, definiteness, and completeness of these prepara- tions cannot be equalled by any other method, especially since it is possible to superpose on the results of vital staining a vascular injection by means of carmine gelatine, or to combine with this method of staining the islets some of the methods to be described for staining the duct system by vital stains. In the majority of the observations and experiments, how- ever, I employed for the same purpose neutral red, which has the great advantage of being restored by direct oxidation from the air if the reduction should proceed too far. The neutral red solutions are prepared and injected in the same way as the janus green solutions but I have not succeeded in finding any way of making them permanent. THE ISLETS OF LANGERHANS 261 The character of the preparations made by the neutral red method is sufficiently indicated by the drawing I have had prepared from one of my preparations. The ease with which counting can be accomplished in these preparations is, however, better illustrated by the photographs which were exhibited when the lecture was given. These photographs were made from preparations of guinea-pig pancreas stained intravitam in neutral red and simply spread on a slide and covered without either teasing or section cutting. I wish to emphasize here also the fact that my reports of counts in guinea pigs are actually what they purport to be and not the result of computations of the total content made from partial counts. In one or two instances they have been quoted by other authors as estimates. Nevertheless, I have found that if a sufficient amount of tissue from each segment of the organ be counted by this method and the tissue weighed and the total computed from the ratio of number to weight, the results will correspond within one or two thousand of the actual numbers, in a guinea-pig’s entire pan- creas. The relations of the duct system were similarly studied by means of new post-vital staining methods which were just as specific for the ducts as Janus green and neutral red were for the islets. The best results were obtained by the use of pyronin injected in the form of a 1 in 1000 solution in normal salt solution, but excellent results were also obtained by methylene blue used in the same way as neutral red. Methylene blue, as Ehrlich pointed out, stains the islets, too, but the latter reduce the dye more quickly than the duct cells, and the duct cells take up the dye more readily than the islet cells, so by careful use it can be employed to confirm the results obtained by pyronin staining. I have since found that acridin red is just as effective as pyronin for this purpose. By means of these stains every duct cell in the pancreas may be stained even to the last centro- acinous cell. The pyronin preparations cannot be preserved, but methylene blue preparations may be made permanent by the usual molybdate method. The difficulties of keeping the proper content of oxygen and controlling leaching prevent these 262 HARVEY SOCIETY permanent methylene blue preparations from being as perfect as the fresh ones. They have been of much service to me, how- ever, in the study of the relationship of islets to neighboring acini. The use of the vital staining methods just described has the great advantage that they permit a complete survey of the whole pancreas, which should prove of great advantage in experi- mental work, since the effect on metabolism of any experi- mental procedure is determined by the condition of the whole pancreas, and not by that of a part. The antiglycosuric power of the pancreas is the sum of the powers of all the units in it which are physiologically active, and unless one can see and estimate all of these units he must necessarily have but a partial appreciation of their potencies. For example, the islet tissue in the remains of a pancreas of which the duct is higated may be so dispersed among fat that a section, or even a series of sections, may give the impression that the total quantity of the islet tissue is extremely small, or by unfortunate chance it may be missed altogether in the tissue taken for examination. But in a preparation in toto of the whole pancreas, as may be seen, for example, in some of the wall plates which I have prepared, the efficiency and completeness of the regulatory processes in the pancreas may be recognized at a glance. These methods, however, although they tell us at once whether, or to what extent, islet tissue is present in the pancreas, do not enable us to estimate the nature or extent of changes in the acinous tissue itself, or to answer definitely those interesting questions concerning the presence or absence of true transitions between the islet tissue on the one hand and the acinous tissue on the other. For this purpose it is necessary to resort to section methods. Lane has shown us how to handle individually the different secreting elements and how to preserve and stain the secretion antecedents. To the methods devised by him, how- ever, must be added a method which does all of this in one preparation, relying on differential staining capacity instead of differential solubilities of the cell contents in the process of fixation. This means, of course, that the fixing fluids em- THE ISLETS OF LANGERHANS 268 ployed must fix all of the several secretion antecedents, and that the stains employed must stain them different colors. Furthermore, the fixation must not introduce into the section precipitates which will confuse the interpretation of the cell structures. These conditions I have found to be fulfilled by a fixing solution containing osmic acid, bichromate of potassium and a minimum of acetic acid, the formula of which is given in my article on the pancreas of the guinea pig published in The American Journal of Anatomy. This fixing fluid has the dis- advantage of slow penetration and, as with all osmic acid solutions, the area of good fixation is very thin, but by care in preparing the pieces and by using sufficient fluid, and, above all, by removing as much as possible of the fat from the tissue, or by increasing the concentration of the osmic acid and the ratio between volume of tissue and volume of solution in proportion to the fat content, these difficulties can be overcome. For staining these sections a combination of aniline acid fuchsin and methyl green is recommended, by means of which the granules of the A cell are stained red, those of the B cell lilac. The details of these methods may be found in the article quoted above. The results are indicated diagrammatically in the drawing of the islet of the guinea pig shown, and ad naturam in the drawing of an islet with the surrounding acinous tissue from the cat, which has been kindly loaned me by Dr. John Homans for this lecture.* It must not be supposed that these technics can be handled successfully by investigators without effort. They require skill in the manipulation of tissue and of stains which, unfortunately, is not often acquired in the routine work of the histological and pathological laboratories, which is concerned almost ex- clusively with the simple and reliable but, unfortunately, cyto- logically not very useful hematoxylin and eosin methods. They require also an appreciation of the rapidity with which tissues undergo post-mortem change. In this connection it mny be mentioned ‘that, provided proper 1 (ae deaehin bible! at ‘the lecture are not reproduced here. 264 HARVEY SOCIETY care be taken in the fixation of the material, and if a fixing solu- tion be employed which actually preserves the secretion ante- cedents, the beautiful staining technics devised by Professor Mallory may be employed with advantage. The preparations and counts made by the vital staining methods, the quantitative results of which are published else- where, at once bring out certain considerations which are of the utmost importance in explaining the contradictory results of observation and experiment. The enormous quantity of the islet tissue becomes at once apparent, and the inaccuracy of estimates made on the examination of sections can be perceived and computed. The number of islets per cubic millimetre of pancreas, in guinea pigs between 300 and 600 grammes weight, averaged 22.28, 19.5 times as great as Dewitt’s estimate and 22 times as great as Laguesse’s. When the error of observation is of such colossal dimensions, what reliance can be placed on estimates made on the basis of a few sections from the splenic end of the pancreas? A great number of islets came to light that are ordinarily overlooked, ranging in size from single islet cells, included in the epithelium of an acinus, or of an excretory duct, through groups of a few cells up to islets of the largest dimensions. The vague outlines, insisted on by Vincent and Thompson, Dale, Laguesse, and other advocates of the variability hypo- thesis, vanish, and the islets, whether they be one cell or a large mass, are sharply outlined among the acinous tissue. Acinous cells which have lost their zymogen no longer appear like islet cells, because all of the latter are filled with sharply stained secretion granulations. The islets are found to have a wide range of variation in different guinea pigs, in different parts of the pancreas of the same guinea pig, even in adjacent parts of the same pancreas, thus reducing the so-called quantitative method of estimating the number of islets after experimental procedure by counting the number in similar areas of a few sections from the same portion of the pancreas to the plane of luck and guess work. Similar observations carried on by Clark by means of the THE ISLETS OF LANGERHANS 265 janus green vital staining technic show that the range of vari- ability in number of islets in the human pancreas is as great as in that of the guinea pig. His observations also show that even the most careful computations made by the section method are far below the real number of islets in the pancreas. One of the arguments which has been seriously advanced against the islet hypothesis of diabetes, by persons little familiar with the structures themselves, it is true, is that the quantity of the islet tissue in the pancreas is too small to have any such importance. The fallacy of this argument is at once apparent when we find that the total number of these structures in a guinea-pig pancreas may be as high as 56,000 and that in a new-born guinea pig there may be as many of these islets in one milligramme of tissue as 551. The average number per milli- gramme as well as the average number per animal is far below these numbers, but, nevertheless, considerable. The vital staining with pyronin and with methylene blue brought to light details concerning the relations of the islets to the duct system and acinous tissue which are interesting because they reveal the normal regulatory mechanism of the pancreas as regards these structures, because they explain fully the con- tradictory results of various observers as to whether the islets have a duct connection or not, and a capsule or not, and because they explain the predilection of the islets for the centre of the lobule which has been a matter of so much difficulty to the advocates of variability. They have brought to light a system of epithelial tubules of great complexity which form a web of anastomosing ductules around the main duct and its branches whose obvious object is the formation of new islets and acini. The relations of these tubules are brought out in the Figs. 5, 6 and 7 of my article which are displayed for this lecture as wall plates. Connected with these tubules as well as with the main duct are large numbers of islets in all stages of growth, ranging from single islet cells to islets of a diameter of half a millimetre. In fact, with a few exceptions, all the islets of the pancreas are connected at some place to the duct system. In only a few instances, however, does the lumen of the tubule penetrate the 266 HARVEY SOCIETY mass of the islet, and even then it is everywhere separated from the islet cells by duct cells. The duct connections become, there- fore, merely vestiges showing the origin of the islet, and the position of the islet, and the nature of its relations, or of its in- vestment, is wholly determined by its developmental history. If it has originated from the main duct or one of its branches or from the web of ductules referred to above, its position is primarily interlobular and it will have a more or less well- defined capsule. If, on the other hand, it has originated from several intralobular ductules, it will be in continuity with acini on all surfaces. Furthermore, the fact that the islets originate from the duct system explains fully why they are rarely found on the surface of the lobules, for, as is well known, the ductules ramify in the interior of the lobule and there also we find the islets which have taken origin from them. The contradictions in the results as regards the relations of islets to ducts become thus a matter of efficient technic and of attention. Those who have concentrated their attention on the large primarily interlobular islets have been perfectly right in concluding that they were discontinuous with the acinous tissue and that they had a sort of capsule, while those who have studied islets which were primarily intralobular in origin have been just as correct in insisting on their continuity with acinous tissue. With both groups the question of the relationship to ducts has been more a matter of rhetoric than observation, except in the case of Laguesse, who not only patiently studied these relations in series, and saw, though not so completely as they are displayed in vitally stained preparations, the short ducts which link up the islets to the ducts, but maintained courageously the reality of these connections when everybody else was asserting the contrary. The existence of similar duct connections in the human pan- creas has been shown by Clark, using the vital stain, and by Weichselbaum and Kyrle, using the section method. We are now in a position to consider the question of vari- ability of the islets, that is, the possibility of transforming islets into acini or acini into islets directly. The evidence in support THE ISLETS OF LANGERHANS 267 of the theory of variability consists of the statements of Lewa- schew, Dale, Vincent and Thompson, and Laguesse, who assert that the number of islets may be greatly increased in a few hours by stimulation of the pancreas or in a few days by starva- tion of the animal, and that cells transitional in type may always be found in the border of the islets. These claims are also sup- ported by the observations of many pathologists who have seen similar transition figures. On the other side we have the experi- ments of Opie, Cecil, Dewitt and myself, who agree in stating that neither secretin nor pilocarpine stimulation nor starvation produces any such result. Cecil’s results are particularly ap- posite here because he made free use of the advantages to be gained by the use of Lane’s differential staining methods to identify the elements. His results offer no support either to the idea that the islets can be experimentally varied by these methods, or to the existence of transitions. My own results are of the same order, though obtained by the method of total counts of the islets in the whole pancreas. The results obtained in the animals which were stimulated for hours with secretin and in which the pancreas was so exhausted that the majority of the _ acini contained no zymogen, were all well within the limits of natural variation, and indeed one of the lowest counts in any animal was found in a guinea pig thoroughly exhausted with secretin. Allen also reports a series of experiments on the effect of prolonged starvation in cats, in which the islets of Langer- hans, instead of being increased, really seem to be diminished. In many of these he saw appearances which might easily be mistaken for transitions but which, he says, are obviously not such, because the acinous cells, though exhausted and deprived of their usual contents, never show an islet arrangement, while the islets all show the characteristic arrangements of the cells and relations to the blood-vessels. Allen’s experiments are doubly interesting, because, in addition to starvation, the animals were subjected to injections of various carbohydrates and fats which should, in accord with the hypothesis of Laguesse, stimu- late the new production of islets at the expense of the acini. On the basis of these experiments, we can now dismiss 268 HARVEY SOCIETY definitely from consideration the claim that secretory exhaustion, or exhaustion by inanition, alters in any way the respective relations of acinus and islet, and accept as definitely proven a high degree of functional specificity for the latter. The question of absolute specificity, however, remains for the present un- decided and, as I have pointed out elsewhere, must remain so until positive results are obtained by experimental means. No accumulation of negative evidence, however great, would be suffi- cient to establish the claim of absolute specificity of the islets as compared with the other epithelial elements of the pancreas, for such aclaim presupposes the loss of some of the original potencies, which may be considered to be possessed in like degree by all of the epithelial elements of the pancreatic anlagen, or, as my observa- tions by means of vital staining show, by the epithelium of the ducts, and especially by the web of fine tubules which surround the true ducts and have for their sole function the production, by growth and differentiation, of new islets and new acini. Closely bound up with this question is that of the capacity and method of self-regulation in the pancreas, whether it be in the ordinary process of growth or in the processes of repair and restitution which may follow disease or surgical interference. Meantime, however, I may give a brief account of the grounds on which we have rejected the descriptions of Laguesse, Dale and others of transitions between the acini and the islets. The investigation, as I have already indicated, must rest on the cytological characters of the acinous cells on the one hand, and of the islet cells on the other. Studies carried on by technical methods which were adequate have been published by Lane, Cecil, Homans, Allen and myself. Lane and Cecil used for control of the observations the neutral gentian staining technic which stains particularly well the granules of the B cell and the zymogen granules, so that they were able at once to say whether a given cell contained islet granules or zymogen granules or both. Their results are consistently negative as to a true tran- sition between islet and acinus. Allen used for the most part eosin and methylene blue, but to some extent the neutral gentian technic, with similar results, except that he found large numbers THE ISLETS OF LANGERHANS 269 of cells which might easily be mistaken for transitions but the arrangement of which, in the form of acini, at once revealed as acinous cells. My own observations and those of Homans have been conducted with the Lane methods, and with my new methods of fixation and staining which combine in one prepara- tion differential staining of both kinds of islet cells and of the acinous cells. They agree in their results with those of Lane, Allen, and Cecil. The nature of the results obtained by this method is illustrated by the schematic representation of a guinea-pig islet and acinus, and by the beautiful drawing of an islet and surrounding tissue of the cat, which Dr. John Homans has kindly loaned me for this lecture. The results of the ex- amination of fresh tissues vitally stained are of the same order. Never does one meet, either in the normal or in the exhausted pancreas, cells which cannot on the basis of positive characters be at once diagnosed as acinous cell, islet cell, or duct cell. We must therefore seek another explanation for the claims that have been advanced by various writers of the existence of such transitionals in the pancreas. We may do this by first in- quiring what characters a cell which is intermediate in type be- tween islet and acinous cell ought to present, and then applying this criterion to the transitions which have been described. It is necessary, therefore, to look for a moment at the microscopic characters of the epithelial elements of the pancreas, and to compare them one with another. The characters of the acinous cell are well represented in the figure made from a preparation fixed in acetic osmic bichro- mate solution and stained with aniline acid fuchsin and methyl green. The acinous cells in this figure are seen surrounding a narrow lumen into which extend a pair of centro-acinous cells con- tinued from a small duct. Each acinous cell presents two zones, which are always present, except under pathological or extraor- dinary experimental conditions. The inner zone is occupied by coarse granules, usually called zymogen granules, because it is supposed that they constitute in part at least the mother substance of the enzymes of the pancreatic secretion. The outer zone in the fresh preparations is transparent, but under good 270 HARVEY SOCIETY apochromatic objectives shows a faint radial striation. In the specimen this outer zone is seen to be composed of a homogeneous ground substance which is optically undifferentiated, but chem- ically defined by the fact that it contains a substance which stains with basic stains, the prezymogen of Mouret, prozymogen of Macallum, or, to use a term more general in its scope and less burdened with unsupported interpretations, the chromidial substance of Hertwig. In this homogeneous ground substance are seen, stained strongly with the fuchsin, long filaments, which are responsible for the faint striation of the living cell. These are the mitochondrial filaments, structures common to all proto- plasm but characterized in the acinous cell by their unusual size. The nucleus of the acinous cell is characterized by its richness in chromatin and by its large oxyphile nucleolus. By appropriate methods there may be demonstrated in the cytoplasm, at the central pole of the nucleus, a network of fine channels containing a clear fluid, the canalicular apparatus of Golgi and Holmgren. In another preparation, made by the Altmann method, the same details essentially may be noted, and in addition a number of small fat droplets may be seen in the base of the cell, for the most part enveloped within the mitochondrial filaments. The small centro-acinous cells and the cells of the intra- lobular duct to which the acinus is attached exhibit an optically homogeneous cytoplasm containing many irregularly scattered mitochondrial rods and granules. They are distinguished from the acinous cells by the fact that they lack both zymogen gran- ules and chromidial substance. In the islets of Langerhans we see two kinds of cells, the A cells and B cells, each of which is characterized by the pos- session of secretion antecedents peculiar to itself. I have already given the reasons why we think that these granules are not simply small zymogen granules of the same sort as exist in the acinous cell. The mitochondria of the islet cells resemble more those of the duct cells than those of the acinous cells, since they are minute filaments, scattered throughout the cytoplasm. These cells differ from one another by the solubilities and stain- ing properties of their granules—they differ from acinous cells THE ISLETS OF LANGERHANS 271 by the lack of both zymogen granules and chromidial substance. The nuclei of the islet cells differ from those of the acinous chiefly in not possessing the large oxyphile nucleolus. Summing up these characters, all four types, acinous cell, duct cell and both kinds of islet cells contain mitochondrial filaments but the character and distribution is characteristic for each. The characters which chiefly distinguish the acinous cell are the possession of chromidial substance and zymogen granules. Those which chiefly characterize the islet cells are the possession of their characteristic granules and the absence of chromidial substance. Considering these properties of the cell it is obvious that to establish a transition between them we must find cells which partake of the properties of both, or we must find cells which have lost the characters of the type from which it is developing, and other cells representing all gradations connecting it with that type on the one hand and with the type towards which it is developing on the other. We cannot accept the presence of undifferentiated cells per se as evidence of transition, but must insist on the intermediate phases which demonstrate progress and its direction. Examining on the basis of these criteria the assertions of those authors who describe transitions, we find that they fall into three classes: first, those who state that, as the acinous cell loses its zymogen and chromidial substance, it becomes more like an islet cell; second, those who have lightly and without due consideration interpreted the A cells as transitions because they frequently occur on the surface of the acinus; and, third, those who perceived the logical necessities of the case and attempted to prove a real transition by showing the acquisition by the acinous cells of positive islet characters. The first class requires no discussion because the statements rest upon a negative in- terpretation of the islet cell. The second group has been sufii- ciently answered by the demonstration by Lane and others that the A cell was not an intermediate but a specialized cell with characters peculiar to itself. The third group, consisting of Mankowski and Laguesse, requires a more extended discussion. 272 HARVEY SOCIETY Mankowski, in preparations of the guinea-pig pancreas fixed in Flemming’s fluid and stained with safranin, found acinous cells filled with tiny granules, which he thought to be acinous cells changing into islet cells. Apparently the same granules have been seen by Laguesse, who thought that it was an indica- tion of a new form of internal secretory activity of the pancreas. This appearance is very common in the guinea pig, particu- larly in animals which have been kept for some time on a diet of hay and oats without green food. It is, however, not a normal occurrence but a degenerative change, all stages of which may be conveniently studied. The granules first make their appearance in the outer portion of the cell, and, as they increase in number, the chromidial substance disappears. At the same time the mitochondrial filaments swell, become spherical, and finally dissolve in the cytoplasm. Soon the granules fill the entire cell and the zymogen also disappears. The change in the mitochondria stamps this change as a pathological one. Moreover, the granules are not islet granules, since they are larger, more highly refractive, and are insoluble in any of the reagents used for dissolving the islet granules. The accom- panying illustration,’ from a preparation of a guinea-pig pan- ereas fixed in Hermann’s fluid and stained in neutral gentian, is sufficient evidence of the fact that these granules are not islet granules. In this preparation the Mankowski granules are well preserved, discrete, and intensely stained. The granules of the A cells are fused into a homogeneous mass, but the proto- plasm, as a whole, stains intensely, while the granules of the B cells are dissolved out. These granules therefore have no con- nection with the formation of islet cells. Laguesse at first considered the A cells as transitional types, but in the majority of his descriptions relies more upon position and vague differences for the evidence of transformation. In a recent publication, however, he describes a cell intermediate in position between an islet and an acinous which partakes in a measure of the characters of both. I have searched diligently * Exhibited at the lecture but not reproduced. THE ISLETS OF LANGERHANS 273 in my preparations for such appearances but without success, except where it was obviously a case of overlapping of cells, in which case the characters may be optically combined. In these cases, however, the true character is at once resolved by the study of adjacent sections in the series. Another resort of the advocates of variability is the assertion that the network of wide capillaries described by Ktthne and Lea, and confirmed by many observers, is not, in reality, a point of difference between the islets and the acinous tissue, since the capillaries of the islet communicate freely on all sides with those of the general circulation of the pancreatic lobule, but rather an effect of the withdrawal of the mechanical support of the tissue, resulting from a diminution in size of the constituent cells. The implication is that such a glomerular structure might arise anywhere in the lobule for mechanical reasons. I have indicated in a previous paragraph that it is possible after vital staining with janus green, followed by appropriate reduction of the stain, to inject the pancreatic blood-vessels with carmine gelatine and so obtain a preparation in which every islet large and small is stained, and in which all the blood-vessels are colored. An additional advantage of this method is that, in the process of staining, the arterial walls are stained with the dye, and thus in the preparations the arteries are bluish, the veins carmine red, and the capillaries intermediate in color, thus adding to the advantages of a vital staining of the islets the equivalent of a perfect double injection of the vascular system. After fixation in molybdate of ammonium these preparations may be dehydrated, cleared, and mounted in balsam, or they may be dissected under the binocular microscope to obtain frag- ments suitable for high power study. Preparations made by this method, in which there is never any doubt as to which is artery and which vein, show that the larger islets have a direct arterial branch running into them, and some of the largest interlobular islets may have three or four such branches as a direct arterial supply. The blood-vessels of the islet are neither sinusoids nor venous retia mirabilia. Even the smallest islets, groups of only 18 274 HARVEY SOCIETY a few cells, have in relation with them a capillary loop which is distinguished from its neighbors by its larger calibre. The end of all these observations and experiments brings us back to the situation which existed when Opie succeeded in demonstrating the specific changes in the islets in cases of diabetes in man. The object has been to investigate, by methods which were really discriminative, the attacks which have been made upon the theory of individuality and permanence of the islets of Langerhans and to afford thus a firm anatomical founda- tion upon which experimental inquiry as to the functions per- formed in the body by these structures may proceed. As a result of these studies, we can now state with assurance that the islets of Langerhans are specialized elements of the pancreas, having secretory powers differing from those of the acinous tissue, developing in embryonic life from the undifferentiated epithelium of the pancreatic anlagen and in post-fetal life from the epithelium of the ducts, having a peculiar blood supply char- acterized by its direct arterial source, by the larger calibre of its capillary vessels, and by the close association of the latter with the epithelial cells. These experimental morphological studies give us no information as to the nature of the internal secreting function which is obviously indicated by their struc- ture. For enlightenment on this topic we must look to further experimental work. In this direction, naturally, interest has centred around the phenomena of diabetes, and workers have sought by various experimental methods to show that the work done by the pan- creas in connection with the control of carbohydrate metabolism was the function in particular of the islets of Langerhans. That this is a narrow point of view, whether we regard it from the stand-point of the derangements of metabolism in diabetes or from the stand-point of the internal secretory activity of the pancreas, is obvious, for the fact that the islets are themselves composed of two independent sorts of internal secreting cells would be sufficient to suggest that their internal secretory function is two-fold, and many studies indicate that there is more to diabetes than a deficient utilization of carbohydrate material. THE ISLETS OF LANGERHANS 275 The investigation of the relation of the pancreas to carbohydrate metabolism is only a part of the difficult problem of internal secretion in the pancreas. It has, however, the merit of definite- ness and so serves as a convenient starting point for investiga- tions designed to determine what the internal secreting functions of the pancreas may be, and to what extent they are mediated by the several epithelial elements which constitute the organ. The efforts to solve this problem have been directed towards isolating experimentally the two principal tissues of the pancreas and determining the effect of this isolation on the excretion of dextrose in the urine, observations on the pathological changes in the pancreas in human diabetes, and observations on the effect on the pancreas of diabetogenic drugs and procedures. Most important are the attempts to reduce the pancreas to the level of a pure endocrine gland by ligation of its excretory ducts. The first attempts by Schulze and Ssobolew were most interesting and apparently conclusive, but the results obtained by those who followed them were so contradictory that Allen in his able summary of this work divided the experimenters into groups, the designation of which sufficiently indicates the nature of their results. Allen groups them as follows: (1) authors re- porting preservation of islet tissue only; (2) authors reporting preservation of acinar tissue only; (3) authors reporting pre- servation of neither tissue; (4) authors reporting preservation of both tissues. In order that experiments of this sort may be definite, it is requisite that the result of the operation be the total exclusion of one of the tissues, and that the identification of the remaining tissue be established beyond a vestige of doubt. Accordingly, in the consideration of this matter, the question of the efficiency of the experimental method and of the completeness of the investigation becomes of the utmost importance. Since ex- clusion of one of the tissues is a necessary part of the experiment, conclusions based on a partial exclusion, for example, of the acinous tissue may be at once rejected as irrelevant. Concretely stated, what we seek to know is whether, by the method of duct ligation, the acinous tissue may be wholly destroyed, and, if so, 276 HARVEY SOCIETY what is the nature of the tissue remaining in the pancreas and what its efficiency from the stand-point of carbohydrate meta- bolism. We can now accept as something agreed upon by all workers that the profound changes produced in the pancreas by duct ligation do not produce glycosuria, although there may be fluctuations of carbohydrate tolerance during the duration of the experiments. Although the logical way to treat this matter would be to discuss the changes that take place in the pancreas in the order of the occurrence, the question of prime importance is whether the tissue which is left behind contains islet tissue. We may then proceed afterwards to discuss whether it contains at any period of the experiment anything else of the nature of secretory tissue besides islet tissue. Here again it is necessary to insist on the identification of the islets by positive cytological characters, and not by arrange- ment of cells, or by blood supply, or by lack of the character- istics of an acinous cell. If the question is decided in the affirmative, then we can utilize the observations of individuals who did not employ so exact a criterion in considering the general questions involved. In this regard the investigations of Laguesse, Gontier de la Roche, and Kirkbride are significant. Laguesse used for recognition of the islet cells the character- istic secretion granules which he stained by a method similar in its results to my neutral gentian method. Gontier de la Roche worked in conjunction with Laguesse and employed his methods. Kirkbride, working in the laboratory of Professor W. G. MacCallum, used for the identification of the tissues the neutral gentian technic recommended by Lane. Laguesse re- ported that in a rabbit examined 45 months after ligation of the duct the sole tissue remaining in the pancreas was islet tissue. The animal exhibited no glycosuria. Kirkbride ligated the distal third of the pancreas and examined it 15 months later. She found that it consisted of a eyst-like duct surrounded by masses of fatty tissue in which were imbedded opaque dots or strands, the only remains of the former pancreatic tissue. Studied microscopically by the Lane technic, she could find no THE ISLETS OF LANGERHANS 277 trace of acinous cells, but on the contrary the remains of the pancreatic tissue consisted of small masses of cells which were filled with the fine blue stained granules of the islet cells. These observations establish beyond doubt the survival, whether by direct continuity or by regeneration, of the islet cells. Since they do not involve a serial study of the whole pancreas they do not establish the complete absence of acinous cells, which is necessary to the argument. Tiberti rejected the demonstration of the islet nature of the tissue by Laguesse on the ground that the latter had not demonstrated the character- istic blood supply of islets. This objection, however, is of small moment, since the physiological efficiency of the islet is directly dependent on the islet cells, and only indirectly on the blood supply. MacCallum isolated a portion of the pancreas of a dog, allowed the animal to recover from this operation, then seven months later removed the portion of the pancreas not included in the previous ligation, leaving only the atrophic remains of pancreas isolated at the first operation. After this operation the animal showed a large but rapidly diminishing glycosuria for four days, following which it acquired a considerable carbo- hydrate tolerance. Then, twenty days later a third operation was performed, removing all of the remains of the pancreas. This was immediately followed by grave glycosuria which per- sisted for five days, when the animal was subjected to a total thyroidectomy, sparing two parathyroids. This operation the animal survived for three days, during which time no further sugar was secreted. This experiment shows that the atrophied remnant was capable alone of securing a high carbohydrate tolerance. As the author states, it is unfortunate that the dog was not kept under observation for a longer period without the thyroidectomy, for some doubt still remains as to the influence of the anesthetic in producing the glycosuria. This experiment, if repeated and confirmed, is exactly what is re- quired to establish beyond doubt that it is the remains of the pancreas itself which constitute the antiglycosurie factor and not some vicarious adaptation to the new conditions, outside 278 HARVEY SOCIETY of the pancreas. The examination of the tissue removed in this case revealed only islet tissue and duct tissue. Thus, we can say that ligation of the duct is not followed by glycosuria, that true islets are preserved in the pancreas, that the amount of acinous tissue, if any, evades search, and that the removal of the remains of the pancreas produces at once the effect of a total pancreatectomy performed without previous ligation. The one thing that is lacking to a proof of the physiological efficiency of the islets of Langerhans apart from the acini is to prove that a search of the whole pancreatic remnant by adequate means would be equally unsuccessful in revealing acinous tissue, that is to say, that acinous tissue is actually absent. For this task the methods of vital staining which permit of accurate identification of the islet tissue and a search of the whole pancreas under the binocular microscope for other pancreatic elements offer obvious advantages. In- vestigations by this method have been carried on in my lab- oratory by Mr. Elbert Clark. Before proceeding to a general description of the results obtained by Clark, it may be well to emphasize certain aspects of the experiment which doubtless have been responsible for the differences in the results obtained by ligation of the ducts. In the case of the dog especially, Lombroso found very little degeneration of the acinous tissue after ligation of the ducts, while MacCallum obtained results exactly comparable to those obtained by many workers in the guinea pig and rabbit. Pratt even claims to have secured complete degeneration of all pan- creatic tissue after an operation which involved complete separa- tion of the pancreas from the duodenum and ligation of all possible ducts. It is obvious that the nature and degree of the reaction will be modified by several circumstances concerning which the experimenter must inform himself when the tissue is finally investigated. In this connection it is to be expected that the degree of the primary reaction will be proportional to the injury inflicted on the pancreas, and that an operation, which requires the dissection of the pancreas away from the duodenum, while sparing the blood supply of the latter, will be THE ISLETS OF LANGERHANS 279 followed by a more intense inflammation and a more thorough- going fibrosis of the organ than one which involves the mere finding and ligating of the ducts. It also gives as a result a tissue which is much more difficult to investigate thoroughly. Undoubtedly the chief sources of uncertainty in these experi- ments and ones which should be rigorously controlled are the presence of supernumerary ducts, and the re-establishment of continuity between duct and bowel. The lack of adequate evi- dence on these matters throws doubts on the results of the operation. That there may be supernumerary ducts in the dog has been established by the observations of Hesse, and the fact that Lombroso admits that there may be rapid atrophy after ligation of the ducts in some cases, though he insists on the generally adequate preservation of the acinous tissues in others, justifies the suspicion that the differences may be due, in part, to the presence of supernumerary ducts, and also, in part, to the occasional re-establishment of communication with the bowel. To offset this possibility, in every experiment, the proof of adequate and permanent closure of communication with the bowel should be forthcoming. That the re-establish- ment of communication with the bowel is possible is well illus- trated by one of the cases in Clark’s series, which he permits me to quote. In this animal, a rabbit, the duct was ligated in two places, and cut between the ligatures. When the animal was killed 15 months later it was found that the duct had effected a new communication with the bowel at a distance of about two centimetres from the stump of the duct, which was also found at autopsy. The patency of the new opening was demonstrated by picking up the duct in the body of the pancreas and inject- ing a colored mass through into the bowel. In this case the duct was much larger than normal, but the pancreas, though somewhat reduced in size and somewhat richer in fibrous tissue, was, in other respects, normal. Clark’s observations on the rabbit and guinea-pig pancreas after ligation of the duct are not yet complete, but they show certain interesting facts which must be taken into consideration in connection with experiments of this sort. They have been 280 HARVEY SOCIETY controlled by means of the vital staining technic, which per- mits a complete examination of the pancreas instead of a small fragment of it, as is the case when the section method only is employed. In the guinea pig the primary effect on the acinous tissue is very complete. A majority of the acinous cells degenerate at once; probably, however, not all of them. The examination of the tissue at the end of seven days reveals a large number of cells which, because of their imbrication over the ends of inter- lobular ducts, we regard as acinous cells which have lost their characteristic substances and are now indistinguishable in structure from duct cells. These cells as well as the remains of the duct system show many mitoses, and it may be con- sidered that, in the regenerative processes which go on in the pancreas from this time forward, cells which were parts of the duct epithelium, and these cells which, with Tiberti, we must regard as dedifferentiated acinous cells, participate in equal measure. In this primary effect apparently all those islets which are of small size and closely related to the acinous tissue participate. In the rabbit, the general course’of the change was the same except that in the primary changes not all of the acinous cells were involved. Towards the end of the first month regenerative changes become prominent, resulting in the formation of new acini and new islets of Langerhans from the remains of the duct system. At the same time the original islets not involved in the primary effect, that is to say, the larger islets, begin to be involved in the advance of the sclerotic process. The connective tissue in- vades the tissue of the islets themselves and the cells show to some extent atrophic changes. Thus, both primary tissues are present in the pancreas for a certain period, and throughout this period acini are being constantly formed, pass through a series of developmental states, and, to some extent, undergo atrophy, or become again de- differentiated. At the same time the islet tissue is being progressively increased by the addition of new islets developed THE ISLETS OF LANGERHANS 281 from the duct system with which they remain in continuity, and the original islets for the most part atrophy as a result of the invasion of the connective tissue. At five and one-half months practically all of the islet tissue in the pancreas is new islet tissue which has been developed since the ligation of the duct. The oval and spherical islets of the original pancreas are no longer to be found. Instead one sees branching masses of islet tissue which stain quite normally in neutral red and janus green, or, after fixation, in the granule stains. Never- theless there is present also considerable acinous tissue in the form of bulb-like acini, containing large lumina, and composed of cells which contain abundant zymogen granules. Beyond this period the number of cases which we have investigated is small, but they show unmistakably a progressive increase in the islet tissue, coupled with a tendency for the efforts to reconstitute the acini to diminish. The appearance of the islets at these several stages is represented in the figures which I have had drawn from photographs of the fresh neutral red stained pancreas. These figures show not only the formation of islets from the ducts and the presence of the acini, but also the difference in the morphology of the new islets as compared with those in the normal pancreas. At what period the last vestiges of acinous tissue disappear we have not yet determined, but one case, kept for 533 days after the ligation of the duct, showed a complete absence of acinous tissue. The neutral red preparation is shown in the wall plate. The pancreas tissue consisted almost entirely of islets imbedded in a mass of fat © having the shape and situation of the original pancreas. The islets consisted of a few isolated islets of various shapes, race- mose masses with bulbous protuberances composed wholly of islet cells, and irregular networks of cords composed of islet cells. In addition there were here and there in the pancreas complicated nets of epithelial cords composed in part of islet cells, in part of undifferentiated cells, and a few masses of dendritic tubules composed wholly of undifferentiated cells. The animal showed no glycosuria. In this case the whole pan- 282 HARVEY SOCIETY creas was thoroughly inspected and we can be quite sure of the absence of acinous elements. In view of Tiberti’s objection, another rabbit which had been kept for 21 months after ligation of the duct was injected with janus green and carmine gelatine to show whether the new islets had a characteristic arrangement of the blood-vessels. The conditions found in this animal were identical with those in the preceding, except for the fact that the stump of the duct left in contact with the bowel had regenerated a small mass of pancreatic tissue. The islets were found in this case to have the same characters as in the preceding one, and the usual direct arterial supply and rich capillary net. No acini were found except in the regenerative portion, though the whole pancreas was inspected. These observations of Clark, which show a capacity on the > part of the pancreas to regenerate acini, which persists for several months, lead me to conclude that the experiments along this line, though promising, are as yet not wholly conclusive. The case which I have quoted, where after 533 days had elapsed from the ligation the animal was free from glycosuria and the pancreas free from acini, I think it may be fairly said, is the only case so far reported where the whole pancreas has been investigated, for the presence of acini, by adequate means. With the exception of this case, then, the most we can conclude is that the islet tissue is the predominating tissue in the pancreas if the animal is kept a sufficient length of time after ligation of the duct, and that the length of time necessary for this purpose has been for the most part under-estimated. It should be remarked, here, however, that the failure of one investigator to get complete results in no respect invalidates the results of another who did obtain complete results from ligation, that is to say Lombroso’s results in the dog must simply be placed in the category of unsatisfactory experiments which help in no way to decide the question whether the islets alone mediate the carbohydrate function, and in no respect invalidate the results of MacCallum who did obtain complete results in the dog. MacCallum and Pratt’s results, on the other hand, as THE ISLETS OF LANGERHANS 283 well as Kirkbride’s, Laguesse’s and Gontier de la Roche’s can be challenged, in my opinion, only on the basis of the fact that inevitably by the section method only a small portion of the pancreatic tissue can be inspected, and so the proof of absence of acinous tissue or of islet tissue, as the case may be, cannot be made complete. The difficulties of this line of experiment, as well as the expense of it are very great. The accidents which happen in the operation, such as missing an accessory duct, or during recovery, such as re-establishment of communication with the intestine, or from causes outside of the pancreas which may cause the death of the animals before the experiment has lasted long enough to be conclusive, all add to the difficulty of obtain- ing definite results. It is clear for reasons which Allen has pointed out that these results would better be obtained in the dog, and that two things are necessary which have not been met in any investigation of this sort hitherto, viz., a complete in- vestigation of the remains of the pancreas, without missing any part of it, by means of the vital staining method, followed by confirmatory examination of all portions which remain doubtful after the vital staining, by adequate discriminative section methods, and a complete history of the carbohydrate tolerance throughout the duration of the experiment. Pratt’s results are subject to the same criticism as the results obtained by other workers with the ligation method, that is, the proof of the absence of islet tissue is not adequate. These experiments also should be repeated and the results tested by injection of the animal with neutral red and janus green. Experiments and observations which show a destruction of islet tissue leaving the acinous tissue unimpaired include the observations in human diabetes, and the experiments of Homans and Allen on the changes produced in the pancreas of dogs and cats which have become diabetic as the result of the reduction of pancreatic tissue. The pathological observations in human diabetes are interesting and suggestive but inconclusive, since many authors report cases of grave diabetes in which the islets are apparently unaffected. 284 HARVEY SOCIETY The experiments of Homans and Allen, done independently by these two observers with consistent results, constitute just what is needed to establish the islet hypothesis when the results of the ligation experiments furnish unequivocal results from the other side. These observations show in animals which have lost a sufficient amount of their pancreas to produce a mild glycosuria, in Homans’s observations, an exhaustion of the islets, indicated by a disappearance of their secretory granules, and in two animals in which a grave permanent glycosuria was produced a degeneration of the islet cells. Allen found that removal of nine-tenths (more or less) of the dog’s pancreas, leaving the remnant in communication with a duct, led to a regular and definite result, viz., a diabetes which, in the earlier stages, was relatively mild, but increased with time, till, appar- ently, there was complete inability to utilize dextrose. ‘‘The microscopic, apparently, agree with the clinical observations. At the outset the diabetes is in one sense functional, viz., in the sense that the islets show no visible alteration.’’ Homans’s experiments controlled by my fuchsin methyl green method show at this stage an exhaustion of the islet granules. ‘‘ At later stages, under the typical conditions, a typical condition df the islets appears to be present in every instance, viz., a visible degenerative change involving all the islets and showing regularly the same picture, viz., loss of cells, deficiency of protoplasm in remaining cells, degenerating (generally pyk- notice) nuclei, occasional naked nuclei.’’ Homans’s results show by the anilin fuchsin methyl green technic hydropie degenera- tion of the B cells. In both these series of experiments the acini were well preserved and apparently normal. These experiments constitute a most important contribution to the proof of the islet theory and I think we may confidently predict that the proof on the other side, namely, the competence of the islets per se in the absence of acinous tissue to care for the carbohydrate function, will before long be satisfactorily estab- lished by means of the methods which I have outlined. The experiments of Homans and Allen establish in my opinion the incapacity of the acini alone to mediate this function. I have THE ISLETS OF LANGERHANS 285 had the pleasure of examining Dr. Homans’s preparations and can vouch for the accuracy of his descriptions. He has also loaned to me for the purpose of this lecture the beautiful colored drawings, one a preparation showing a normal islet from the eat, and one from a diabetic cat in an advanced stage of hydropic degeneration. An interesting consequence of his observations also is the fact that the B cells of the islet are mainly concerned in this function. As for the pathological observations on cases of human diabetes, it is my opinion that this method will only lead to satisfactory results in the hands of those who are competent to handle the tissue technically by methods which tell us what the secretory potential of the islet tissue is. Homans’s results show that the first evidence of pancreatic deficiency is accom- panied, not by a gross pathological change, but by exhaustion of the secretory granules of the islet. To the ordinary pathological technic such islets would appear wholly normal. Moreover, the application of adequate histophysiological technic requires a promptness in the performance of autopsies which cannot usually be obtained in human eases. Accordingly, all negative results reported in human cases should be rejected, unless the time at which the autopsy is performed, and the ante-mortem history, and the methods of investigation justify the presump- tion that adequate precautions have been taken to detect by microscopical means changes which would indicate a depression of the functional potential and rate in the islet tissue. One of the most interesting aspects of the work of Allen and Homans is the question they raise as to the capacity of the pancreas for self-regulation, and the conditions which in- fluence this capacity. In neither series of experiments does there seem to be much evidence of a response on the part of the pancreas in the way of compensatory hyperplasia, not- withstanding the fact that in many of the experiments the re- duction of the pancreas was carried to a point where the organ was incompetent to carry on its function of internal secretion adequately. The reaction in these cases was an exhaustion and finally a degeneration of the islet tissue. 286 HARVEY SOCIETY On the contrary, the observations of Tiberti, Gontier de la Roche and Clark indicate a high capacity for restoration of the pancreas after the duct has been ligated. Clark has shown that for a period lasting for at least five and a half months after ligation of the duct in guinea pigs and rabbits, the remains of the duct system are actively mitotic, and new acini and new islets are being constantly produced. There is indirect evidence also that if the duct continuity be established at a sufficiently early period the pancreas may be entirely restored. The case of the rabbit which I have described, where, after ligation of the duct in two places and cutting between the ligatures, a new opening was effected into the intestine, illustrates this. The dilated duct and a certain degree of fibrosis of the pancreas indicated that the occlusion of the duct had persisted for some time before the new opening was re-established, and, when we consider the rapidity with which degenerative changes go on in the pancreas of the rabbit after duct ligation, it seems probable that in this case the acinous tissue had first been almost com- pletely destroyed and subsequently completely restored. A number of similar results were obtained in guinea pigs. Whether our interpretation of the process in these cases is correct or not, the fact remains that the pancreas after ligation of the duct shows a high capacity for regeneration of its parts, while reduction below the point of physiologic competence is followed by very slight indications of this capacity. It is obvious, there- fore, that the magnitude of the physiological demand made on the organ, whatever that may mean, is not the only thing to be considered, but that there is a biological equilibrium in the eells of the organ themselves which influences profoundly the reaction. One of the things which, it may be supposed, in- fluence the reaction is the physiological activity of the cells themselves. In general it may be said that mitotie activity in secreting cells means a suspension of secretory activity, that the two processes are to some extent mutually exclusive. That being the case any condition which would favor mitosis would by so doing inhibit secretion, and conversely any condition which would inhibit secretion would favor mitosis. In other words THE ISLETS OF LANGERHANS 287 regeneration of pancreatic tissue would be favored by experi- mental procedures which would check secretion, and stimulate cell division. The progress and course of regeneration so initiated would depend upon the degree of participation of the several tissues involved, and on the chance of survival, under the conditions, of the several tissues so regenerated. Under the conditions which obtain after ligation of the duct both these needs are realized. Mitosis is favored by the in- flammatory reaction which follows the operation and which involves the whole pancreas, and the external secretion of the pancreas is abruptly brought to an end. Thus regenerative processes are initiated which result in a continuous increase of islet tissue because it is the only tissue which can survive. The new formed acini, as soon as they reach the point where they are able to secrete, become cystic and undergo retrograde changes. The differences between the pancreas as a whole and frag- ments of pancreas in this respect may be simply due to an extension of the process which destroys the acinus in the former case to the islet in the latter. The pancreas as a whole con- tains a great excess of islet tissue, so the work of the original islets is not increased markedly by the operation of tying the ducts. The degeneration of the original islets, though inevitable according to Clark’s experiments by reason of the invasion of the islets themselves by the process of fibrosis, does not proceed rapidly enough, so that when one considers the sum of all islet tissue in the pancreas old and new there is at any point a marked decrease in the excess of islet tissue present, and so the new islets survive and do not fall a prey to the degenerative changes which Homans and Allen have demonstrated to be a result of overwork. The explanation of Allen’s results which show that a smaller fragment of pancreas will protect against glycosuria when the duct is tied may be simply on the one hand that the operation favors mitosis in the pancreatic cells, and on the other that the exclusion of the external secretion from the intestine with its reduction of the efficiency of carbohydrate digestion diminishes the load on the islet cells which are left behind, 288 HARVEY SOCIETY so permitting them to survive, the increased formation of new islet tissue accounting for the gradual raise in carbohydrate tolerance as time goes on. Though the results are suggestive, the matter obviously requires further investigation. The indications of pancreatic hypertrophy after reduction of the pancreas should be tested not only by the increase in size over that observed at operation, but also by careful measurements of the fragment left, and by search for mitoses in the fragment at different intervals. The effect of pancreatic feeding on the rate and amount of regeneration in animals in which the duct has been tied should also be determined, both in cases where the whole pancreas has been left and in those where a large portion of pancreatic tissue has been removed. In the latter case of course it would be necessary to proceed with caution because of the fact that pancreatic feeding increases glycosuria, and according to Kirk’s recent experiments, also, acetonuria in totally diabetie dogs. It has been my purpose in this paper to emphasize the neces- sity in experimental work on the pancreas of full histological control, and to illustrate by reference to the results as well as by means of colored illustrations drawn from actual prepara- tion how this histological control may be accomplished. I have already indicated some of the directions in which further experimentation is desirable, and doubtless many others will occur to those who hear this address or who afterwards read it. One of the most inviting fields, apart from the investigation of diabetes, is undoubtedly the investigation of the interrelation of the various internal secreting organs. There has already been a great deal of stimulating speculation in this field, but as yet very little in the way of objective proof. The influence of experiments on the other endocrine glands on the structure of the pancreas is susceptible of more accurate study perhaps than any other. In the guinea pig, my counts of the islets in 99 guinea pigs, of all ages and of both sexes, would serve as a convenient starting point for such an investigation. The neces- sity of examination of the whole pancreas indicates for these investigations the use of small mammals in which the vital THE ISLETS OF LANGERHANS 289 staining methods work well. The white rat offers many ad- vantages, also, because the statistical studies of Donaldson, Hatai, and Jackson provide information of the normal range of variation which is available to the same extent in no other animal, and also much interesting information concerning the correlations between the various organs, and their plasticity under varying experimental conditions. CARBOHYDRATE UTILIZATION IN DIABETES BASED UPON STUDIES OF THE RESPI- RATION, URINE, AND BLOOD * PROFESSOR ELLIOTT P. JOSLIN Harvard University N the classical work of Naunyn? upon diabetes mellitus occurs the following sentence: ‘‘In general, even in severe diabetes, at least in man, the carbohydrates ingested are not com- pletely excreted in the urine again as sugar. A portion of the starch, as well as of the dextrose, will be burned in the organ- ism.’’ This view was also shared by Kulz. Naunyn, however, refers to a case in which von Mering records an excretion of all the sugar ingested, and attention is called in the report of the cases of Kulz to four instances in which apparently a similar condition existed. Von Noorden? defines diabetes as ‘‘a disease in which the capability of the organism to adequately burn grape sugar is pathologically lowered,’’ and in another place he says: * ‘‘One cannot help thinking that, in man, even when death has resulted from coma, the diabetes has not always been quite ‘complete’— that is to say, the pathological processes which produce diabetes have not developed so far, and the factors which favor the storing up of glycogen have not been so completely destroyed as is the case in a dog whose pancreas has been entirely ablated.”’ * Delivered March 13, 1915. From the Nutrition Laboratory of the Carnegie Institution of Wash- ington, Boston, Mass. I wish to acknowledge my grateful appreciation of the help received from Mr. Emmes, Miss Babcock, Miss Tompkins, Miss Corson and Miss Sandiford, of the Nutrition Laboratory, as also my indebtedness to Mr. Higgins, who controlled several of the experiments with the Tissot apparatus, and to my secretary, Miss Helen Leonard, for cheerful work upon long computations and puzzling charts. 290 CARBOHYDRATE UTILIZATION IN DIABETES 291 Notwithstanding all the work upon diabetes, this question of the utilization of carbohydrates in human diabetes has not been settled. In diabetic dogs evidence has accumulated point- ing to the complete loss of this power to utilize carbohydrate, and the work of Murlin and Cramer‘ has given definite results upon this point, although so recent a writer as Landsberg,’ work- ing from a different point of view with other animals, comes to the opposite conclusion. The present paper is concerned with diabetes in man. At this time I wish to call to your attention certain observations bearing upon this problem which are related to the body weight, the urine, the storage of carbohydrate in the body, the respiratory metabolism of diabetics both fasting and following the administration of food, and the remarkable disappearance of acidosis in diabetics with prolonged fasting, which is associated with a rise in their respiratory quotient. I. THE INFLUENCE OF WEIGHT UPON THE DETERMINATION OF THE UTILIZATION OF CARBOHYDRATES IN DIABETES The changes in weight which occur in a normal individual, following a slight increase of the carbohydrate in the diet, are so striking that one might hastily conclude that a study of the weights of a diabetic patient would give some idea as to his utilization of starch and sugar. A closer scrutiny of the problem, however, reveals many difficulties. In the first place, the diet employed in most cases of diabetes and all severe cases is low in carbohydrates, and seldom reaches 10 per cent. of that of normal individuals. In other words, it amounts to less than 50 grammes carbohydrate—200 calories—per day. The effect of 200 calories upon the weight is possible of determination theoretically, but practically such an experiment is difficult because the protein, fat, and carbohydrate must be kept at uniform levels for a long period. But in a severe case of diabetes some of even this small amount is lost in the urine, which renders the available carbohydrate for increasing the weight still less. There are other complications. A severe case of diabetes with 50 grammes of carbohydrate in the diet usually excretes more than 50 grammes of sugar in the urine, and it is 292 HARVEY SOCIETY difficult to assign in proper proportion this excess of urinary sugar between the carbohydrate ingested and the carbohydrate already stored in the body on the one hand, and the protein simultaneously ingested and the body protein on the other. Remarkable changes in the weight of normal as well as of diabetic patients will also occur, although the caloric value of the diet remains constant, if the proportion of fat to carbo- hydrate is altered. working at the Nutrition Laboratory of the Carnegie Institution, have studied the basal gaseous metabolism for 89 men and 68 women and their average results are shown in Table 9. TABLE 9.—Respimatory Quotient AND ToTat MeETABOoLIsM oF NORMAL InpIvipvaALs aT Rest at Least 12 Hours Arrer THe Last MEAL Individuals H: 6: eas oe hy per RECT teria Aste ate Me iues ke os Average =0.83 25.5 ROP LIION haya: oa: eh caine nie afate Average =0.81 24.9 312 HARVEY SOCIETY I would eall attention to the slight difference existing between the respiratory quotient of men and women—0.83 and 0.81. I have also incorporated the heat production, calculated from the oxygen intake, which was approximately 25 calories per kilogramme per 24 hours. This latter figure is lower than we are apt to consider, but it should be remembered that it is based upon fasting periods when the patient 1s purposely endeavoring to be quiet. It would be absolutely wrong, from such determina- tions covering periods of 15 minutes, or even a few hours, to draw conclusions upon the total heat production for the day. In illustration of the method and at the same time of the diffi- culties of determining the respiratory quotient of normal indi- viduals I give my own chart. TABLE 10.—Normaut Inprvipvat (EB. P. J.) Fastina EXPERIMENT. December 23, 1914. Wexiaut, 64.9 Kitos; Heiaut, 177.8 Cm. Duration Calories per A | kilo per 24 Min. Sec hours 15 6 20.45 14 59 20.53 15 0 20.78 Average =0.78 Naturally I took the greatest possible pains to be quiet and breathe in a normal manner, but it will be seen that, whereas the values for the carbon dioxide of themselves, and of oxygen for themselves, vary to an extremely small degree from period to period, yet they differ sufficiently to make a considerable varia- tion in the respiratory quotient. This experiment emphasizes the possibilities for error in the determination of the respiratory quotient even under most favorable circumstances. The respiratory quotient of individuals fasting for long periods is well exemplified by the studies made by Benedict *° CARBOHYDRATE UTILIZATION IN DIABETES 313 upon a man, who went 31 days without food. These are illus- trated in the following tables: TABLE 11.—TxHe Respiratory Quotient or A Man Dorinea aA Pro- LONGED Fast Calories per Period Time R. Q. pees nine weight Preliminary period....| 4th day before fast......... 0.81 33 3rd day before fast......... 0.89 32 2nd day before fast......... 0.89 29 Ist day before fast......... 0.82 27 Period of fast........ Days 1— 5 of fast... .22....: 0.77 26 (Average) Days 6-31 of fast.......... 0.72 23 (Average) Days 6-31, early a. m....... 0.73 23 (Average) After period.......... 2nd day after breaking fast 1 0.78 25 3rd day after breaking fast ! 0.94 27 1 Twelve hours after food. TABLE 12.—QvuanmTiTIEs oF PROTEIN, CARBOHYDRATE AND Fat OxmpiIzED By Fastinag Man at Nutrition LABORATORY Determined from the Daily Metabolism, the Urinary Nitrogen and the Calculated Non-protein R. Q. R. Q. Quantities oxidized Calories Period of fast Ba Actual | protin | Pat” | Gu | Ga | Reus (CaS sea 0.78 0.76 43 69 135 30 p.0s 0.77 0.74 50 42 142 30 RG CBM ys vio. es aie 0.74 0.74 68 39 130 29 CLONES Gee 0.75 0.71 71 4 136 28 UY OLY a s's.. hiv ss 0.76 0.72 63 15 133 28 6th to 3lst day average....... 0.72 0.70 53 Ql 114 26 1 Actually a total of 32 Gm. carbohydrate were burned during the 6th to 13th day inclusive, and later none, 314 HARVEY SOCIETY It will be seen that, prior to the experiment, the respiratory quotient differed little from that of the group of normal indi- viduals above mentioned. With the withdrawal of all food the respiratory quotient fell, and after the fifth day reached a point which Magnus-Levy 7’ has said theoretically represents the quotient which is obtained when the metabolism consists of 85 per cent. of fat and 15 per cent. of protein, namely, 0.72. In other words, five days of starvation removed the last discern- ible influence of carbohydrate remaining stored in the body, and the individual lived wholly upon body fat and body protein. It is possible to discover how much fat and how much protein take part in the metabolism. Knowing the nitrogen in the urine, one can calculate the amount of oxygen employed by the body for the oxidation of the protein { which it represents, and, correspondingly, the amount of carbon dioxide given off can be determined. If we subtract these computed figures from the total carbon dioxide and oxygen obtained by direct experiment, we have left the carbon dioxide produced by the non-protein metabolism in the body, and the relation of the two gives the non-protein respira- tory quotient. In a useful table constructed by Lusk,”* the per cent. of carbohydrate and of fat for any given non-protein respiratory quotient between 70 and 100 can be calculated. On this basis it was shown that Benedict’s fasting man burned either none or only a trace of carbohydrate after the sixth day. When the respiratory quotient of this man was 0.73 on the seventh day, it represented a non-protein respiratory quotient of 0.70 and no carbohydrate was burned. A respiratory quotient of 0.74 gave a non-protein respiratory quotient of 0.71, which represents the oxidation of 3.8 grammes of carbohydrate; a respiratory quotient of 0.76 gave a non-protein respiratory quotient of 0.72, which is evidence that 15 grammes carbohy- drate were burned. Respiratory Quotient in Normal Individuals after Food.— The respiratory quotient following the ingestion of food is shown well by the fasting man at the Nutrition Laboratory tIn estimating the quantity of body protein burned from nitrogen in the urine the equivalent 6.00 is employed instead of 6.25. CARBOHYDRATE UTILIZATION IN DIABETES 315 for the periods before fasting commenced. It will be seen that 12 hours after food it varied from 0.81 to 0.89 in the four days. Similarly, following the termination of the fast, the respiratory quotient rose, indicating the combustion of large quantities of carbohydrate. An experiment was performed upon myself which was com- parable to those later carried on with the diabetic patients when tests were made of the influence of food upon their metabolism. The changes in my own respiratory quotient following the inges- tion of 60 grammes of carbohydrate in the form of oatmeal are given in Table 13. TABLE 13.—Metaspouism oF A Normat_ INDIVIDUAL Arter Foop. Weiacut, 64.9 Kitos; Heicut, 177.8 Cm. Os Date, 1914 Condition per min. | per min.} R. Q ior c.c. c.c. hours Sept. 9...| 1-2 hours after breakfast...) 205 241 0.85 26 Sept. 10...| 1-2 hours after breakfast...| 192 237 0.81 25 Sept. 30...| 9 a. M., fasting..-.......... 159 194 | 0.82 21 10.30 a. m., after 60 Gm. carbohydrate as oatmeal 189 212 | 0.90 23 Pecado. .1 6 A.M, fasting............ 152 194 | 0.78 21 Rey REs TARGUS 5 2/25/51 01'<. «yarn < «1 15) 196 | 0.77 21 ww a It will be seen that the respiratory quotient within an hour rose some 8 points after eating 60 grammes of carbohydrate in the form of oatmeal. It has been calculated that if 48 grammes carbohydrate are burned in 24 hours at the rate of 2 grammes of carbohydrate each hour continuously for the 24 hours, the respiratory quotient would rise 3 points; in other words, would be about 0.75 instead of 0.72, which is a fat-protein quotient. I wish to emphasize the change in respiratory quotient of only 3 points when approximately 48 grammes of earbohydrate are burned at the rate of 2 grammes of carbohydrate per hour per day, and the rise of 8 points following directly upon the inges- tion of 60 grammes carbohydrate. The continuous combustion of small portions of carbohydrate amounts to the combustion of a considerable quantity of carbohydrate during the whole 316 HARVEY SOCIETY day, and yet it will raise the respiratory quotient very little. The combustion of 24 grammes of carbohydrate at the rate of 1 gramme per hour could scarcely be detected with our present methods, and yet a tolerance for 24 grammes of carbohydrate is relatively a high tolerance when one is dealing with serious cases of diabetes. The Respiratory Quotient in Diabetes——In mild eases of diabetes, when the urine is free from sugar and the carbohydrate in the diet large, the respiratory quotient differs little from that of normal individuals. The respiratory quotient of these same mild cases of diabetes will be lowered by fasting or by the with- drawal of carbohydrate, as just shown in the case of the normal fasting man. Undoubtedly the limited quantity of carbohydrate in the diet in cases of severe diabetes is responsible to a large degree for the low respiratory quotient which such patients show. Magnus-Levy called attention to this, and so have other observers. It is well exemplified by the change in the respira- tory quotient of Case No. 714. This patient, with only moderate acidosis, became sugar-free upon April 16, 1914, following 14 days of treatment. On December 3, 1914, she re-entered the hospital with 4.4 per cent. of sugar, but under fasting treat- ment became sugar-free after the omission of four meals. The respiratory quotient on successive days is shown in Table 14. TABLE 14.—I.uustrRaTION oF FALL IN RESPIRATORY QUOTIENT OF MILD Diasetic. Case No. 714. FEMALE Diet } Date R. Q. Urine sugar hear Protein Fat Alcohol Gm. Gm. Gm. Gm. IDECP Nee. sot La 4.4 per cent. -- ++ + + 0 Dec. 94=/ 52 y ees 20 Gm.? + ao + 10 Dec; '5— 6.5...) 107s 0 0 0 0 25 Dec. 6- 7....| 0.75 0 15 40 45 10 Dec. 7- 8....| 0.75 0 15 45 60 a Dec. 10-11....| 0.73 0) 15 55 100 9 1 Approximate. ?In14hours. Tests were made fasting at 8 A. M., which was one hour after the collection of the 24-hour urine. CARBOHYDRATE UTILIZATION IN DIABETES 317 It will be seen that, whereas the respiratory quotient was 0.78 on entrance, due undoubtedly to the oxidation of some of the carbohydrate ingested, though much at the same time was being lost in the urine, this rapidly decreased to 0.73 under starvation followed by a low carbohydrate diet. Yet this woman could not be considered a severe case of diabetes. The quotient was low simply because she was living almost exclusively upon a fat protein diet. The problem of drawing inferences from the respiratory quotient in diabetes is complicated by the fact that much of even the little carbohydrate which is given to a diabetic patient is lost in the urine. The patient really approaches the condition of the fasting man in that he is living exclusively on fat and pro- tein, although in this case not that of his own body. If all the carbohydrate ingested is lost in the urine, his respiratory quo- tient would be 0.72. But there are other complications. Occa- sionally cases of diabetes are seen where the sugar in the urine exceeds that of the diet, and speculation at once arises as to the source of this excess of sugar. Magnus-Levy *® has pointed out that if the sugar in the urine amounted to 60 grammes and the protein in the diet to 100 grammes, the additional quantity of oxygen which would be demanded to form this amount of sugar out of protein would lower the respiratory quotient to 0.70. The situation is still further complicated by the presence of unoxidized acid bodies in the urine, amounting frequently to 20 to 40 grammes and occasionally to 60 grammes calculated as beta-oxybutyric acid. The amount of oxygen consumed in the formation of these bodies—for beta-oxybutyric acid is far more rich in oxygen than are protein and fat—would again lower the quotient, and it has been ealeulated by Magnus-Levy that the respiratory quotient of a case of diabetes presenting 60 grammes of sugar in the urine for 100 grammes of protein in the diet, and excreting 20 grammes of beta-oxybutyric acid, would fall as low as 0.69. For convenience, these figures are summarized. The respiratory quotient of the fasting man at the Nutrition Laboratory was 0.72. The calculated respiratory quotient of a 318 HARVEY SOCIETY normal individual who is burning 15 per cent. protein and 85 per cent. fat is 0.72. The theoretical respiratory quotient of a diabetic individual excreting all the carbohydrate in the diet, and 60 grammes of glucose for each 100 grammes of protein in the diet, is 0.70. The theoretical respiratory quotient of the diabetic individual excreting 60 grammes of glucose for 100 grammes of protein and 20 grammes of beta-oxybutyric acid as well, is 0.69. These calculations presuppose that the sugar and beta-oxybutyric acid excreted were formed during the same 24 hours, but who knows whether this is the case? The theoretical non-protein respiratory quotient of a case of diabetes living upon fat and the non-carbohydrate part of the protein molecule, as calculated by Lusk, is also 0.69. TABLE 15.—THEoRETICAL RESPIRATORY QUOTIENTS (From Magnus-Levy) Diet Calories RQ: Protein. 100 Gin, (100X451 =AdO) oa sce aon as cea Carbohydrate 567 Gm. (567 X4.1 =2325)..........05. \2o735 | 0.97 Pratem: 100i (1004 140s ok. (Le \orss | 072 Fat, 250 Gm. (250%9.3 =2825).........ssss cose ee. , Loss in Urine Surar, 60) Gm. (60 X40 246) oi sie he ey ye cies 2489 0.70 Loss in Urine Sugar, 60:Gm. «..22.2(60 KA. 22246). sc dee see ! B-Oxy. acid, 20:Gm. (20 X4:7 = 94). oie cnes kinetic 2395 0.69 Total loss =340 ) Table 15 shows the theoretical respiratory quotient, which should be reached under varying conditions of diet for a normal individual, and the changes which theoretically are present under special conditions in diabetes. Figures of this type have dominated the discussions of the metabolism in diabetes from the start, and whenever experiments have not produced figures comparable with these, they have often been considered erron- eous. We are taught to believe that diabetic patients are not severe unless the respiratory quotient is 0.69. It is questionable, however, whether the experimental data at our disposal enable us to say that our theories are backed up by the results which CARBOHYDRATE UTILIZATION IN DIABETES 319 we obtain. If one looks over the lists of respiratory quotients obtained in successive periods with any variety of respiratory apparatus or calorimeter, he will be shocked at the discrepancy and is forced to the belief that any argument based on a change in the respiratory quotient of one point is unjustifiable, and any argument which is based on a change in the respiratory quotient of two points really rests on a very slender thread. A change of three points is, however, deserving of consideration, but modesty should rule in conclusions which are to be drawn from any given set of experiments. It is appropriate to discuss here what constitutes a severe diabetes. At the outset it can be said for our own encouragement that Naunyn did not pretend to be able to distinguish accurately between the various types. Usually by severe diabetes is under- stood those cases in which, to quote von Noorden, ‘‘notwith- standing a prolonged, extreme carbohydrate-free diet, the urine contains sugar.’’ By an extreme carbohydrate diet von Noorden undoubtedly meant one containing protein, fat, and a few green vegetables, in other words, a diet with 10 grammes of carbo- hydrate, more or less—not a strictly fat-protein diet. The definition is also open to objection, because one frequently meets with cases of diabetes of long duration who excrete in the urine but a small per cent. of the carbohydrate intake, yet this persists for many days upon an extreme carbohydrate-free diet, but the patient could not be classed as severe. Another method of classification is adopted by Lusk, who considers cases severe which, when put upon a protein-fat diet have a dextrose-nitrogen ratio of 3.65:1. By this he means that 3.65 grammes of dextrose appear in the urine for 1 gramme of nitrogen, or the 6 grammes of protein which it represents. In other words, 60 per cent. (actually 3.65 ~ 6.25 = 58.4) of the protein burned by the body appears in the urine in the form of sugar. Lusk considers that this is the greatest possible amount of sugar which can appear in the urine on a carbohydrate-free diet, and he assumes that it comes wholly from protein. This conclusion has been reached with many observations upon dogs, 320 HARVEY SOCIETY following injections of phloridzin, and by one case of diabetes coming under his personal observation, and he refers to other eases selected from the literature. Unfortunately, or perhaps fortunately, neither of these methods of classification at the present time is wholly satisfactory, because, thanks to Dr. Allen, our patients now become sugar-free very readily. It is possible that fasting will not remove the sugar from the urine of all diabetic patients, but this has been my experience with every case when I have followed Dr. Allen’s directions, and my experience coincides with that of many others. It may be that recent cases of diabetes have been of a different type from those hitherto encountered, but this is hardly possible. Consequently we cannot adopt the definition of von Noorden, and it is embarrassing to adopt the precise definition of Lusk. The dextrose-nitrogen ratio vanishes with fasting, and the clinician does not wish to expose his patient before beginning fasting to the dangers of a protein-fat diet simply to determine his severity. I am hoping that, with the added knowledge of diabetes which the introduction of fasting has brought about, Professor Lusk will pursue his studies further and give us definite rules for testing the severity of the disease. Perhaps definite quantities of protein alone or some special form of pro- tein or derivative of protein could be administered to these patients, and the amount of sugar in the urine determined. Should this method not furnish satisfactory results, another series could be carried out in which varying quantities of fat as well as protein could be added, and, if a third factor were necessary, the calories per kilo could be standardized. But we ean trust Professor Lusk to give us help. Of course dextrose- nitrogen ratios are of little significance without simultaneous reports of the blood sugar. In the data which will follow, consideration will be taken of both von Noorden’s and Lusk’s classifications, but also the severity of the cases will be indicated by a statement of the time intervening between the period of observation and death in coma. It would seem as if the severity of the type of diabetes which resulted in death in coma should challenge criticism. CARBOHYDRATE UTILIZATION IN DIABETES 321 As the periods of observation before death in coma are of importance, the intervals between the determination of the respiratory quotient of the patient and death are given. See Table 17, which will later be discussed more in detail. This appears far more rational than to give the duration of the course of the disease, for many patients present a mild type of diabetes for many years, changing over to a severe type at a compara- tively short period before death. The following table summarizes the respiratory quotients of cases of diabetes considered severe by their observers: TABLE 16.—ReEsPImRATORY QUOTIENT IN SEVERE DIABETES Year g Observers R.Q. 5 1894...... 1 | Weintraud and Laves: Ztsch. f. Physiol. Chemie., 1894, VOM ERER ANON OOssse i RE Re okie Sart yale eiate tere tae acters Wee ike 2 | Nehring-Schmoll: Ztsch. f. klin. Med., 1897, vol. xiii, Dane OLA CA 2 la aes CR ek Pe Png eee Gee ts 0.72 OOS sie. 2 | Magnus-Levy: Ztsch. f. klin. Med., 1905, vol. lvi, p. 86 | 0.71 TOD Ter ats: a org ; Ztsch. f. Exp. Path. u. Therap. ., 1907, vol. iv, CLA CINE ANSAM At ak ica MA oe ot BRN Te leciaiabee Os ts 0.72 1908-1911. |19 Banedict and Joslin: Carnegie Inst. of Washington, Publications 136 and 176, 1910, 1912............... 0.73 r+) ee 8 | Rolly: Deut. Archiv. f. klin. Med., 1912, vol. ev, p. 494 | 0.74 See ah. 3 | Grafe and Wolf: Deut. Archiv. f. klin. Med., 1912, vol. erNa N= OAD 805 Lee as a tata ean ak 8 EE Ue Leake aed od ob 0.74 1912-1914.| 7 | Benedict and Joslin: 1914-15....................6. 0.73 Total. . .43 Average........ 0.73 It will be seen that there is surprising unanimity of agree- ment among the different groups. It should be stated that Leimdorfer *° has obtained much lower quotients, varying be- tween 0.64 and 0.68, with five cases which he considered severe. His figures, however, have not been generally accepted. One of the cases which he considered mild at no time showed a respira- tory quotient above 0.70. According to the computations given above from Magnus-Levy, it was shown that, theoretically, in a diabetic patient with 60 grammes of sugar in the urine for each 100 grammes of protein in the diet—in other words, ap- 21 322 HARVEY SOCIETY proximately the Lusk dextrose-nitrogen ratio—and with 20 grammes of beta-oxybutyric acid, the respiratory quotient would not go below 0.69, and he further points out that, in order for the ratio to sink to 0.653, 150 grammes of sugar must be formed from 150 grammes of protein and 40 grammes of beta- oxybutyrie acid must appear in the urine when the patient is upon a diet of 150 grammes protein and 250 grammes fat. A respiratory quotient of 0.653 is a figure so low that it should be entertained with scepticism. The average respiratory quotient of 0.73 for 43 cases of clinically considered severe diabetes is a far safer figure to follow than to pick out one, two or three from the 43 cases and say that these represent severe cases of diabetes and the others do not. The errors of the deter- minations of the quotients are so great that the average figures are safer than the individual ones. These respiratory quotients, as Grafe and Wolf ** pointed out, show that at least some carbo- hydrates were being oxidized by severe diabetic patients. They also pointed out that with the improvement of patients the respiratory quotients increased from 0.743 to 0.817 in a fasting condition. These figures suggest at the first glance that very little carbohydrate was burned in this group of severe cases of dia- betes. The respiratory quotients are identical with the quotients obtained under similar conditions with the fasting man at the Nutrition Laboratory, though his average for the whole day for the fasting period was 0.72. But we must remember that two corrections are to be made in these figures: first, sugar has been lost in the urine which has been formed from protein, and second, there have been varying amounts of beta-oxybutyric acid, diacetic acid and acetone excreted. Both of these processes represent processes of oxidation, and by demanding additional oxygen for which no carbon dioxide is produced tend to lower the respiratory quotient. Therefore, if we grant that the series represents cases of severe diabetes, we must reach the conclusion that these diabetic patients utilized some carbohydrate, and that their respiratory quotients would have been several points above CARBOHYDRATE UTILIZATION IN DIABETES 323 0.73 had they not been lowered to 0.73 by the production of sugar from protein and the formation of acid bodies. Are the cases reported in the above table severe? At least no cases of greater severity have been hitherto published. By von Noorden’s criterion they might be considered severe, for they did not become sugar-free with restricted diet, yet it is true that this restricted diet was not so rigid as is often employed on account of the marked acidosis. If we accept Lusk’s criterion (and I am not ready to do so until a second human case is studied under modern conditions §) they were not severe. Not one of Benedict’s and my cases showed a persistent D: N ratio of 3.65:1. Yet the clinical facts point to severity. Of the first group of 19 cases of diabetes reported in 1908-1912 by Benedict and myself, 18 are dead, and of these 15 died in coma. This fact can be taken as a measure of their severity. I do not believe, however, that this alone justifies us in saying that a diabetic patient is of the severest type. I conceive it possible, for a moderately severe case of diabetes, by sudden changes of diet, to be driven into coma accidentally. This was done years ago, when diabetic patients, who were living on a free diet, upon coming to the hospital were suddenly deprived of carbohydrate and the fat and protein were increased. It appears to me quite probable that most cases of coma in diabetes have occurred long before the disease had reached its greatest severity, and I wish to point out that therein lies great hope for the future. However, it will be of interest to note the respiratory quotient of a group of six cases of diabetes all ending in coma, § By modern conditions I mean (1) exclusive fat-protein diet; (2) under surroundings which make errors in diet impossible; (3) a duration of at least 7 days to exclude the washing out of stored carbohydrate; (4) a constant (not falling) D:N ratio of 3.65:1 for the last 3 of the 7 days; and (5) several daily blood sugar determinations to furnish some proof, inadequate though it be, that the sugar in the urine has not come from that left over in the blood. At present I cannot advocate such a test because of the danger of acidosis, and believe it better to leave the question, in this form, undecided. 324 HARVEY SOCIETY who died within a period of 44 to 14 days from the time of observation, and to compare these with a group of patients whose respiratory quotient was observed at a greater interval from death in coma. This is shown by the following table: TABLE 17.—REsPIRATORY QUOTIENT IN FaTaL AND Livina CasEs OF SEVERE DIABETES COMPARED Fatal cases Living cases No. of D bef C D Coaa pleas iN Rae ae Nooter | Meck i916 | | =o 552 801 0.72 765 125 0.73 786 ag 0.71 806 72 0.70 | Seeewenire) pomeeen | 4 cases 801-72 0.715 i All of these cases were in good condition May 1, 1915, which would add 61 days to the duration since the observations were made. A consideration of this table suggests that with approaching death the respiratory quotient falls. It will be seen that the cases dying within a period of 44 to 14 days from the time of observation gave a quotient of 0.71, as contrasted with a quotient of 0.74 in cases dying in coma at an interval from death of 442 to 70 days. If we had these figures alone, the inference might be justified, but caution is necessary before drawing such a conclusion. Four living cases of diabetes show a respiratory quotient almost as low—0.715. Instead of progression toward death in coma, their general condition has improved. In other words, a falling respiratory quotient does not necessarily mean approaching death in coma. It does mean that these patients have lived for prolonged periods upon an almost exclusively fat-protein diet, and suggests that they are forming carbohy- drate out of protein and producing acid bodies. It should be said that all of these living cases have been treated either by much restricted diets or by fasting as advo- cated by Dr. Allen. When they were first seen they appeared CARBOHYDRATE UTILIZATION IN DIABETES 325 to be quite as severe cases of diabetes as those earlier studied which died in coma. What shall we say of them at present? None of these cases can be considered well, but all lead a com- fortable life at home. The group of cases dying within a period of 44 to 14 days deserves further comment. The average quotient of these cases was 0.71. From four cf these the non-protein respiratory quotient has been reckoned, and it amounted to 0.695. This respiratory quotient implies that much material must have been formed in the course of the metabolism which used a portion of the oxygen. This was especially true of Case 246,°* who had a respiratory quotient of 0.69, which was based upon an average of 29 periods, most of which were fasting. Stimulated by in- quiries from Professor Lusk, I am fortunately able to show the cause of the particularly low quotient in this patient. His diet and urinary analyses will be found in Tables 18 and 19. TABLE 18.—METABOLISM OF A SEVERE DIABETIC WITH A RESPIRATORY QUOTIENT oF 0.69. Case C. No. 246. Mate. Acute ONSET aT 28. DrATH In Coma In 15 Montus. Monta or Disnass, 13 Urine Diet SS ee Sodium Day Gare bicarbo- Vol. | N. | NHs3| B-oxy. Der bohy- nes Fat pier nate fe SE oe a 2935 | 16.3 | 4.8 | 29 72 1b PSone 0 ieee a7 0) 1.3: |.5.0)| 34: 106 98 | 22] 225 | 30 0 LLU Seana» 4370/| 19.6 | 5.5 | 61 134 65 | 100} 200 | 30 60 BE BAO 4035 | 19.4 | 5.4 | 61 107 65 | 100} 200 | 30 60 Wise ie chaise 3330 | 14.7 | 5.6 | 46 100 | 125 | 45] 100 | 30 25 WD Tae es ae eae 3765 | 16.3 | 5.0 | 48 93 65 | 100] 200 | 22 25 August..... 3691 | 16.6 | 5.2 | 46.6 | 102 71 | 65) 165 | 26 28 Carbohydrate in dict. 00... ce 5 es 72 Gm. Dextrose in urine................... 102 Gm. Carbohydrate balance............... 30 Gm D:N Ratio 1.9 : 1.0 Daily protein metabolism estimated at 100 Gm. Total acetone bodies estimated at 60 Gm. 326 HARVEY SOCIETY TABLE 19.—To Suprtement Taste 18. Case C. No. 246 Diet Calories O: CO: RQ: Grammes Litres Litres OCCT siete ees ce 100 X4.1 = 410 96.6 78.2 Aten acc eee 165 X9.3 =1535 Boowk 235.5 Garpohiydraters. 6 ito. foi5 oe 71X4.1= 291 58.9 58.9 IACONO Meares ae coe ce tiens 20 eee 37.9 25.3 2418 | 526.5 397.9 | 0.756 WEXtrOSe ets eek Hebe eee 102 X3.7 = 337 76.1 76.1 Acetone bodies as f-oxyb..... 60 X4.5 = 243 58.1 51.6 620 | 134.2 127.7 1798 | 392.3 270.2 | 0.692 The respiratory quotient found, based on an average of 29 periods, chiefly fasting, was 0.69. The average daily urinary nitrogen for the six days of obser- vation was 16.6 grammes, and it was considered that this repre- sented approximately the metabolism of 100 grammes of protein. The beta-oxybutyric acid was 46.6 grammes daily, and allowing for acetone and diacetic acid the total excretion of acid bodies was assumed to be 60 grammes. The fat in the diet as originally recorded was probably inaccurate, and I believe 165 grammes daily near to the exact quantity. From these tables it will be seen that the daily carbohydrate in the diet was 71 grammes, and the dextrose excreted was 102 grammes, giving a minus balance of 31 grammes. This, with the 16.6 grammes of nitro- gen in the urine, gives a D: N ratio of only 1.9to1. In Table 19 are placed the data from which the respiratory quotient can be calculated from the diet and urine, and they show that after deductions for dextrose and acetone bodies, the theoretical quotient would be 0.692, which it will be remembered was identi- cal with the respiratory quotient found by experiment. These tables are submitted as proof that a quotient of 0.69 does not necessarily mean that the capacity for burning carbohydrate has been totally abolished. Computations of a similar character by Grafe and Wolf * CARBOHYDRATE UTILIZATION IN DIABETES 327 lead to the same conclusion. According to these writers, ‘‘the conception which, on the whole, appears to have the greatest probability is that even the severest diabetic has at his disposal 20 to 30 grammes of glycogen for combustion or synthesis, 13 to 20 hours after a meal containing a minimal amount of carbo- hydrate. Perhaps the complete loss of the power of combustion of sugar is, broadly speaking, no longer consistent with life.’’ Effect of Food upon Utilization of Carbohydrates in Severe Diabetes—A moderate number of experiments upon the effect of food on severe diabetics has been recorded, but the actual number of experiments to determine the effect of carbohydrate upon the respiratory metabolism is very limited. Such experi- ments have been published by Leo,** who considered that the respiratory quotient did increase in two cases of severe diabetes, although this was not uniformly the rule, and he concludes that even in severe diabetes a part of the sugar ingested or formed in the body is utilized. Nehring and Schmoll * tested the effect of carbohydrates also in two severe cases of diabetes, but were unable in either to show an increase in the respiratory quotient. Frequently a fall instead of a rise in the quotient took place. Benedict and Joslin,** in a series of experiments chiefly with bread and dex- trose, state that ‘‘the ingestion of carbohydrate produced no very material alteration in the metabolism as a whole,’’ and later, ‘‘no evidence can be found of a combustion of ecarbohy- drate. .’ Two years later a series of experiments with oatmeal and levulose was reported, but without comment. Schilling, in an Inaugural Dissertation, Leipzig, 1911, tested the effects of various meals upon the respiratory quotient of 1 severe, 1 mild, and 2 moderately severe cases of diabetes, and demonstrated no specificity for oatmeal. With the severe case the results were inconstant, but usually tended to show a slight increase. Rolly,*” in a series of experiments, tested the comparative effects of oats, rye, wheat, lentils and green corn-meal upon diabetic patients. Unfortunately, few of the experiments were preceded by control periods. Two of his cases he considers severe. In Case I, at 3, 5 and 6 hours after 70 grammes of oatmeal were administered, the respiratory quotient was 0.73. 328 HARVEY SOCIETY After 70 grammes of wheat meal it was 0.76. The respiratory quotient of his Case V after 80 grammes of oatmeal was 0.71, after 80 grammes of rye meal was 0.73, and after 80 grammes of wheat meal was 0.71. Two of his other cases were only moder- ately severe, and the other only a light case, and all showed an increase in the respiratory quotient after their meals reaching up to 0.83, 0.85 and 0.84 respectively. It will be noted, further- more, that of the two severe cases, in the first the quotient follow- ing administration of wheat meal which was given after oat- meal reached 0.76. Roth ** records slight increase of the respiratory quotient following the administration of carbohydrate. The experiments, however, lose much of their value because of the absence of fast- ing controls upon the day the carbohydrates were given. Falta *® has mentioned several experiments designed to show the effect of the oatmeal cure upon the respiratory quotient. The data of the experiments are not given, but he states that with one moderately severe diabetic the respiratory quotient rose only on the third day of an oatmeal cure, in which 400 grammes had been given daily. Despite this enormous quantity of oatmeal, no glycosuria was observed. It is unfortunate that I have not been able to find a later publication, which was prom- ised. He furthermore makes the interesting statement, which is so remarkable as to invite confirmation, that a similar result was encountered with a normal individual, whose carbohydrate depots had been robbed by living upon a diet poor in carbo- hydrates for a long time. It would appear that only after these depleted carbohydrate stores were replenished, the normal individual, like the diabetic, began to burn carbohydrate. His results are in striking contrast to the changes in respiratory quotient which were shown by the fasting man at the Nutrition Laboratory. At the end of his fast of 31 days he ate food almost exclusively in the form of carbohydrate and the quotient promptly rose to 0.79 and 0.96 on the second and third days respectively. Falta emphasized the fact that in a mixed diet car- bohydrates are burned much earlier. He further states that on a meat diet or on a diet with a moderate amount of carbohydrate the diabetic patient seldom shows a quotient above 0.74, and CARBOHYDRATE UTILIZATION IN DIABETES 329 he also noted the fact, to which attention has been called by Nehring and Schmoll, which is also borne out by our own series of cases, that, following the administration of carbohydrate, a considerable quantity of carbohydrate not only remains in the body, but the respiratory quotient remains low. Intravenous injections of sugar (30 grammes) given by Falta to severe diabetics, who had eaten 300 grammes of oatmeal for three days without glycosuria, brought about an evident glycosuria, but the respiratory quctient rose proportionately little. In the case of a severe diabetic there was no increase, but a still further lowering of the already low respiratory quotient. The present series of experiments with foods which I have to report represent a part of the experiments upon diabetics whose metabolism following the administration of food was studied at the Nutrition Laboratory since 1910. TABLE 20.—Errect or LevuLose Uron A SEVERE DriasBetic. Case No. 332. Fremaue, AGE 37; Weicut, 40 Kitos Date, 1911 Conditi co re) Calories per ate ondition oi ae eer aay R. Q. less Here tnee March 31 Fasting c.c. c.c. Se a 151 205 | 0.74 35 | PASTE. S55 o6 eid alpen 145 211 | 0.69 36 100 Gm. levulose 1D, & Le cena OANA eae ee 172 | 271 | 0.63 46 TE ET SW SS blk Sree a ere eee 184 261 | 0.70 44 EOTech) cis) sg 'okac!ecore os, caig'e-s acess 180 | 246 | 0.73 42 3 Ey ola pA Se Ae, 2a) | O73 40 AWN Ps tbs Ed ails av avays ried o tor ote 171 250 | 0.68 42 fet Fey a Sat S| ee 166 240 | 0.69 40 April 2 “7 SUITE cee Sic os oad ws, ime 148 199 | 0.75 34 ay eats ai [TT ha eal 2 ak NA 151 203 | 0.74 35 SLL, hake ry SETS eee 154 we QIN O72 36 Oatmeal =70 Gm. carb., 38 Gm. butter LIS US aie) Ug aie fe AR a A SS 163 234 | 0.70 39 UES Se on AD OES ee gah ee A 8 a 167 228 0.73 39 UMM Niches, bat Se sl Me ales 177 238 | 0.75 40 ERTS Aalv heresies Scie avin at Sale St ayes 170 | 230 | 0.74 39 MPA Fete ON tie Sera s als bie Dalene Meee es 154 206 | 0.75 35 SPE SA UME ideas Sha cic/ala wher bie ob chosen elds 163 209 | 0.78 36 330 HARVEY SOCIETY TABLE 21.—Errect or Levutose Upon RESPIRATORY QUOTIENT OF D1aBEtic PATIENTS ‘ Carbohy- s i Case | Duration| Month arate pre Levulose ee R. Q. months | observed ceding day 24 hours Dead Gm. Gm. Gm. Before After 332 28 23 100+ 100 120 0.72 0.69 Alive 552 32 18 30 100 3 0.72 0.76 785 23 20 20 sl 7 No increase 181 Gm. levulose and later, 9 Gm. carb. as vegetables. 90 Gm. total. Three experiments have been conducted with levulose. Case No. 332 was given 100 grammes of levulose when fasting. This patient was a severe diabetic, weight 40 kilogrammes in the twenty-fourth month of her illness, and died five months later. The respiratory quotient before the levulose was adminis- tered was 0.72, and following the levulose the quotient was determined in six different periods during the following three hours and showed an average of 0.69. Despite the fall in the respiratory quotient, the total metabolism increased markedly, although apparently most of the levulose was excreted in the urine. Unfortunately, it is impossible to state how much of the 120 grammes of sugar in the urine for this 24 hours came from the levulose and how much from carbohydrates of the preceding day. Our records indicate that the patient was on a diet containing approximately 100 grammes of carbohydrate. This fact is of interest in comparison with the next two cases, to whom levulose was also given. Case 552, age thirty-seven, weight 40 kilogrammes, in the twenty-third month of her illness, received also 100 grammes of levulose, but this was given after a prolonged period of low carbohydrate feeding. Upon the day previous to the experi- ment the carbohydrates in the diet amounted to 30 grammes. The quantity of sugar in the urine in this case during the 24 hours of the experiment was 3 grammes. The respiratory quo- tient rose four points, namely, from 0.72 to 0.76 after the levulose. CARBOHYDRATE UTILIZATION IN DIABETES 331 The third case, No. 785, was that of a boy of 16 years of age with severe diabetes of 20 months’ duration, weight 42 kilo- grammes. He had been made sugar-free by prolonged fasting and had been kept upon a diet low in carbohydrate and protein, as well as fat. During the 24 hours of the test, the urine con- tained but 7 grammes of sugar. Notwithstanding this fact, the respiratory quotient showed no increase, but a fall of two points. The actual figures are not published now, but the comparative values may be considered trustworthy. The evidence in these three cases, therefore, points to no utilization of the levulose in two of the cases. In one of these most of the levulose was probably excreted, but in the other only a negligible quantity. In the third case there was an increase of three points in the respiratory quotient, indicating a slight utilization of the levu- lose and there was no excretion of levulose of account. It was possible to determine the effect of the administration of potato in two cases. In the first case the experiment was complicated in that the patient was given a small quantity of oatmeal at the start, but on account of her dislike of the same it was stopped and potato substituted. In this case, No. 765, no TABLE 22.—Errect oF Potato Upon. RESPIRATORY QUOTIENT oF SEVERE DIABETICS Carbohydrate intake een Case Duration | Month |———_——————_ : RO: No. months | observed | preceding Test 24 Nome “ day day Gm. Gm. Gm. Before After 63! 765 7 3 15 22 29 0.74 0.73 85 60? 806 6 3 10 6 3 0.68 0.70 66 148 Gm. carb. as potato, 10 Gm. carb. as oatmeal, 5 Gm. carb. as cream, —total 63 Gm. Later in day, 22 Gm. carb. as potato and vegetables. Also 1 egg and 30 Gm. of butter. Se Gm. carb. as potato. Later in day, 1 egg, butter, 6 Gm. carb. as vege- tables. 332 HARVEY SOCIETY TABLE 23.—Errect or Potato Upon THE RESPIRATORY QUOTIENT OF A SEVERE DIABETIC. Date, 1914 December 22 Condition dartiin. pan eah c.c. €.c. RUG) oh al a 156 223 Fasting 150 224 Pet SA seas 155 228 Potato =60 Gm.carb. cA a INA Ki Fi 257 ite TER Aa 168 252 IRE Bech ietes eae 172 250 ica tta A cite 170 233 POM wre. ae 157 227 Evageluda pis dita: 166 231 Case No. 806. Matz, Weicut 62 Kinos ls. per RQ: me Dee 24 ee hours Per cent 0 24 0.67 +0.68| 24 24 8 25 PEN SAVER EA Prey AN 0.14 SOA eatery as a otc 0.18 0.71 28 0.67 70.69 27 hor 0.69 27 0.73 26 0r0)0.7 25 25 0.72 25 Pe WAN SOY HRSA A 0.19 change in the respiratory quotient took place, but in the second, Case No. 806, a slight increase was noted, and apparently rather more than would be accounted for by the limits of error. TABLE 24.—Errect oF OaTMEAL UPON THE RESPIRATORY QUOTIENT OF A SEVERE Diasetic. Case No. 773. FEMALE, WeicHt 40 Kinos Date, 1914 * Condition PIMBASHIN Gece (rete eietale carb. carb. eae we (eG © we ré Tee ay 6 8 @ Oatmeal =42 Gm. co per min./per min. c.C. 146 178 138 O Cals. per Blood sugar Per cent. aR Sasi c.c. 212 | 0.69 | 36 249 | 0.72 | 43 189 | 0.73 | 33 Sea wog'l aig’ ade” 237 | 0.70 | 40 e 8b). 8 oe e 6 0.40} 0 @ 6 06 o/s eee @ 6 oe Diet contained 15 Gm. carb. Oct. 9 and Oct. 18. CARBOHYDRATE UTILIZATION IN DIABETES 333 TABLE 25.—Errect oF OaTMEAL Upon THE RESPIRATORY QUOTIENT OF SEVERE DiABETICS 246 281 332 336 441 561 591 773 746 786 Date 31 | Sept. 22 wee Duration Onset to | Mo. coma of months test 34 15 iui 13 19 17 28 13 24 26 132 127 11 9 10 33 23 50 44 20 18 22 18 prior to March, 1915 17 14 prior to March, 1915 24 Aug. 9 Oct. 29 30 25-31 Dec. 1 2 3 May 19 26 Apr. 2 June 2 May 18 21 Sept. 29 Oct. 9 Feb. 7 Apr. Oct. 8 Ocha Nov. Carbohy- drates ignited Day pre- | Before ced-| test ing Gm.| Gm PSs 15 |100+ Gio ee 50| 40 65| 60 UPA halle Ai LAER pa LH Gasol 135| 29 45 0 100 | 25+ GON tats 2 | 52 ? 48 DAD) | iene ces 45| 25 LG i LEM te GO vase 60| 116 185 | 200 PAU Ue ye igh eee SO ver 15| 80 165} 80 Al) i|ektae Tis FO} 15) 47 SON teee 15| 80 65| 26 15| 50 UG 5s aye 165| 80 15} 60 Ot After Fasting} oat- ea ON G4E Ween ares 0.71 0.71 ged ge 0.71 0.67 0.68 | 0.70 EZ Litah fey bres O69) Poet). OETA ie & Sel ents 0.76 OSG) Meas Savoteyats 0.73 (Oras ett 0.74 | 0.74 0.71 0.69 Osierellteesve ARLE te 0.75 0.70 0.71 Lie 0.69 DSP AST IIe sae Ae 0.71 0.74 0.72 0.72 O76) | ee OVA eras Onoullan a 0.70 0.70 0.73 0.69 OLGO Rae. Peat tins 0.70 0.69 | 0.72 Widest | eeaciak os 0.70 | 0.70 satebies vane 0.73 0.73 0.71 ee ee 0.74 0.69 0.74 +69 +62 ‘One line under a figure indicates that the R. Q. was taken following an oatmeal day, and two lines that it was subsequent to two oatmeal days. 334 HARVEY SOCIETY Eleven experiments have been carried out upon severe dia- betics with oatmeal. These were arranged in some cases to determine the immediate effect of the administration of oatmeal, and in other cases to determine the effect of the prolonged administration of oatmeal. It will be seen from a study of the charts that as a rule the respiratory quotient remained station- ary or fell, in one case it rose four points, and in two other cases it rose one point. It will be noted further that the respiratory quotient, when taken fasting upon the morning following an oatmeal day, amounted in three cases to 0.73, 0.72 and 0.73 respectively, and that upon the morning following a second oatmeal day was 0.69 and 0.76. The respiratory quotient also determined in three experiments after the administration of carbohydrate on the second oatmeal day was 0.72. If one looks at the table as a whole, it will be seen that little change in the respiratory quotient took place; in fact, none of any account except upon the morning following the second oatmeal day. The sum total of the results following the feeding of levu- lose, potato and oatmeal to severe diabetics affords little evi- dence from the respiratory quotient that the carbohydrate was burned, save in the case of one of the experiments with levulose, one with potato, and one with oatmeal. These results corre- spond closely with what has been recorded in the literature. Personally, I believe that before a final decision upon this point ean be reached from this particular line of study, further experi- ments must be performed. Unfortunately, in the experiments recorded, no stated agree- ment was noted between changes in respiratory quotient and variations in the quantity of blood sugar. From Table 7 it is evident that there is a general tendency for the respiratory quotient to rise with an increase in blood sugar, but this may be accidental. Studies now in progress will soon throw light upon this phase of the question. Vv. ACIDOSIS AS A MEASURE OF THE UTILIZATION OF CARBOHYDRATES It has been generally accepted that acidosis will appear when carbohydrate food is either withdrawn from the diet or CARBOHYDRATE UTILIZATION IN DIABETES 335 excreted in the urine. It has been unquestionably the universal clinical experience that the patient who excretes quantities of sugar in the urine equal to or in excess of that in the diet exhibits acidosis, and that patients do not show acidosis who are able to utilize approximately 70 grammes of carbohydrate, or large quantities of protein from which carbohydrate may be formed. This statement cannot be so unqualifiedly made, be- cause I have under observation a woman who, in her sixth month of pregnancy, showed over 6 per cent. of sugar, and later under a careful diet became sugar-free, acquired a tolerance for approximately 100 grammes of carbohydrate, and yet a slight acidosis persisted. Nevertheless, the general mass of evidence points to the disappearance of acidosis when carbohydrates are burned, and upon this general concept arguments have been based for and against the utilization of carbohydrate in severe diabetes. During von Noorden’s oatmeal treatment a considerable quantity of carbohydrate ingested is usually retained or burned in the body, and the decrease of acidosis at the same time is usually considered evidence of the latter supposition being correct, but occasionally the acidosis persists although the carbo- hydrates are not excreted. I doubt if we are in a position to accurately explain the disappearance or non-disappearance of acidosis under these conditions. Oatmeal and other carbo- hydrates are better retained in the body following starvation, and it is quite possible that with the retention of carbohydrate goes hand in hand a mechanical retention of acid bodies. Magnus-Levy pointed out long ago that these were seldom ex- ereted in concentration of more than 1.5 per cent., and that the fall in acidosis during an oatmeal cure may be simply apparent, because the volume of urine excreted has diminished. The in- fluence of preceding fasting is also important, because this undoubtedly regulates to some extent the storage of carbohy- drate. Despite these possibilities, which lessen any argument for combustion of carbohydrate based on the decrease of acidosis following the ingestion of the same, the slight amount of acidosis which is usually found when diabetic patients are on a full 336 HARVEY SOCIETY carbohydrate diet points strongly to the fact that some carbo- hydrate is burned. The increase in respiratory quotient on the last days of an oatmeal cure, which Falta observed and we also have noted, is confirmatory of this position. Various writers have observed that the acidosis in diabetics decreases upon a vegetable day or fasting day, but it remained for Allen to conclusively demonstrate the remarkable fact that acidosis vanished in practically all severe cases of diabetes under these conditions, and that in the remainder, if carbohydrates to a moderate extent are allowed temporarily, the acidosis wholly clears up. If anormal individual fasts, it has been the universal experience of observers that acidosis appears. In other words, the normal fasting individual corresponds with the concept that when carbohydrates are withdrawn from the diet (and this implies carbohydrates which might be formed from protein) acidosis appears. Thus, in the fasting man at the Nutrition Laboratory, acidosis appeared upon the second day and con- tinued until the fast was terminated. How can we reconcile these two opposing facts: the one that fasting dissipates acidosis in diabetes, but produces it in normal individuals. Must the prevalent conception be given up that carbohydrate oxidation and acidosis are unrelated and must we acknowledge that here is an instance where the absence of the burning of carbohy- drates does not lead to acidosis? Such a conclusion appeared unavoidable until observations at the Nutrition Laboratory upon severe diabetics during prolonged fasting began to accu- mulate, showing that, whereas at the beginning of the fast the respiratory quotient was the ordinary respiratory quotient of severe diabetes, 0.72, with a continuance of the fast this had a tendency to rise several points, occasionally even to the neigh- borhood of 0.80. Later experiments, as yet unpublished, at the Russell Sage Laboratory made under the direction of Dr. DuBois and Professor Lusk upon one of Dr. Allen’s patients suggested a similar condition. In other words, whereas the normal individ- ual showing acidosis exhibits a respiratory quotient based upon the combustion of protein and fat alone, the severe diabetic during fasting shows a respiratory quotient which could only CARBOHYDRATE UTILIZATION IN DIABETES 387 be accounted for by the combustion of notable quantities of material other than fat and protein. That this material was not protein was evident, because the amounts of nitrogen in the urine excreted during these periods were not abnormal. The explanation why the severe diabetic shows no acidosis when fasted, in contradistinction to the normal individual, is found in this increase in the respiratory quotient. Several explanations for this increase in the respiratory quotient of fasting diabetics are available. During fasting the diabetic may be able to draw upon sources of carbohydrate in the body which the normal individual cannot. Furthermore, the diabetic has in the body undoubtedly more carbohydrate stored than we have hitherto supposed, and the supposition must be entertained that the diabetic really may actually have more carbohydrate in some form in the body than exists in the normal individual. foi ters y oY ee % a Sree: fis ‘ = ih AY ng 8 aaah a im ar C B k * i, 4 ) f ¢ i a 1 “ R Harvey Society, New York att The Harvey lectures H33 ser, 10 “Ssological ds Medical Seria PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY | \} itt ri hh) ante HAL ARETE 4 Hi i HINT al , y ie AN ANN i i ea Ly PUUAHANNRUALAYNADEVUTANLHYHUAR GHA AGYE UTLEY EAE A / Oh i / i HAH ii ii a Win i | VANDENVNADELH SS od —— ———— : == = —— — = —— —————— rn en — —— - : = =——— —————— ————— a ——————— a — — ; | | NNT i th } PRUE a