mechanism of the anaphylaxis reaction in the dog and that in the guineapig, which could be inferred from an experiment reported by Manwaring. It has been found that the function of the mechanism through which the antibody-antigen reaction operates to produce the physiologic effect of anaphylactic shock in the guinea-pig (that is, the contractility of smooth muscle tissue) remains intact after recovery from such shock. This was early demonstrated by the observation that a guinea-pig which has been sensitized simultaneously with two different antigens and has recovered from a severe shock upon the injection of one of these antigens, can immediately be fatally shocked by the injection of the other antigen. Moreover, that function is unaffected by the more gradual process of desensitization, since a sensitized guinea-pig that has been completely desensitized by the suitable administration of the antigen can be, at once, resensitized.2

288

Manwaring,163 in an experiment that has received little attention in the literature, showed that immediately after a passively sensitized dog has been shocked by the injection of the antigen it cannot be resensitized, even by the direct transfusion of a large quantity (500 cc.) of blood from the same sensitized dog which had served as donor for the first passive sensitization. The possibility that the failure to become resensitized was due to the presence, in the blood of the experimental dog, of some of the previously injected antigen, was excluded by perfusing the animal with about 650 cc. of normal dog's blood between the first injection of the antigen and the second perfusion with the sensitized dog's blood. Manwaring concluded, from this experiment, that the reacting tissues of the dog, which he showed to be situated in the liver, are exhausted by the shock of anaphylaxis.

The question raised by this observation is similar to that presented by the relations that have been found to exist between peptone poisoning and anaphylactic shock in the dog.

Biedl and Kraus,30 who first noted the similarity of the symptoms of anaphylactic shock and of peptone poisoning in dogs-a fall of blood pressure and a condition of incoagulability of the blood occurring in both instances-found that the sensitized dog is refractory to the reinjection of the antigen if this is undertaken soon after the animal has recovered from an intoxicating dose of peptone. They reported, also, that a previously induced anaphylactic shock in the dog protects the animal somewhat against the effect of peptone injection.

These authors interpret their results as indicating that the shock of the anaphylaxis reaction is caused by the formation of toxic digestion products identical with those that are present in commercial peptone.

Weil,237 on the other hand, saw, in the mutually protective action of peptone poisoning and anaphylactic shock, merely the operation of the principle suggested by Manwaring in explanation of the failure of resensitization in desensitized dogs, namely, an exhaustion of the cellular mechanism (in the liver) in which the primary anaphylaxis reaction takes place.

The phenomena just discussed might be thought to be susceptible of another explanation.

The dog's liver, after the shock of peptone poisoning and of anaphylaxis, is probably more or less excluded from the blood circulation on account of the intense hepatic congestion, which is the chief pathologic change in both of these conditions,237 the obstruction to the blood flow through the liver being indicated by the marked stasis in the portal tributaries. It is conceivable that this partial, temporary exclusion from the circulation of the demonstrated site of reaction, in anaphylaxis and peptone poisoning following either of these processes, might place the sensitive tissues out of immediate reach of the subsequently injected antigen or peptone, whereby a local resistance would be only simulated. Against this explanation, however, stand the observations of Voegtlin and Bertheim 220 and of Denecke, showing that the complete exclusion of the liver from the portal circulation in a sensitized dog by means of the Eck fistula does not prevent the subsequent production of anaphylactic shock by the injection of the antigen, the latter reaching the liver in sufficient amount through the hepatic artery.

66

Moreover, it is stated by Biedl and Kraus 30 that the insensitiveness of sensitized dogs to the injection of the antigen after peptone "shock" persists for (at least) 24 hours, after which interval the blood pressure has returned to normal.

The conclusion of Manwaring and of Weil seems therefore justified, namely, that, different from the sensitive tissues in the lungs of the guinea-pig, those in the liver of the dog, after sustaining anaphylactic shock, lose, for a time, at least, their power of responding, in the typical manner, to the antigen-antibody reaction of anaphylaxis.

Anaphylaxis in Human Beings. It has become the almost universal practice of clinical writers in this field to designate as expressions of anaphylaxis the various clinical phenomena which will be considered under the subject of allergy-such phenomena as hay-fever, serum disease, food idiosyncrasy and even drug idiosyncrasy.

Before the question of the occurrence of anaphylaxis in the human being can be discussed, it is necessary to define the meaning of the term anaphylaxis in this connection. The existence in human beings of the sensitizing reagins (which are not capable of mediating anaphylactic sensitiveness in the lower animals) and of other antibodies (precipitin) which are capable of sensitizing the lower animals, dictates the definition of anaphylaxis as "hypersensitiveness mediated by anaphylactic antibodies."

This definition excludes those allergies that are due to the mediation of the sensitizing reagins, such as many cases of hay-fever, asthma, eczema and other related conditions.

It could possibly include some instances of immediate reaction following the reinjection of foreign serum in the human being, but proof is still wanting that, in the few cases of this kind on record, the clinical reaction was actually due to the presence of anaphylactic antibodies, or indeed of any antibodies.

The number of instances in which the intravenous reinjection of serum has been followed by no immediate symptoms, or by no symptoms at any time, cannot be surmised; such instances are, naturally, not reported.

It could be anticipated that some light might be thrown upon the question of the existence of anaphylaxis in human beings by experiments with apes. Such experiments have been carried out 224, 250 by a number of different investigators, some of whom have not published their negative results. Auer made several large intravenous injections at the usual intervals (about one week) into a number of monkeys, studying the effect with several physiologic criteria (respiration, blood pressure and others) without observing the slightest abnormality which could be attributed to a specific increased reactivity. Zinsser 251 alone, in one instance, noted changes in the face suggestive of serum disease. In no other instance, however, did symptoms follow the intravenous reinjection of as much as 5 cc. of serum into the apes.

In conclusion it may be said that, while it is not possible to assert that anaphylaxis does not occur in human beings, it is a fact that the existence of the condition of anaphylaxis in human individuals has not been demonstrated; in other words, no human pathologic change has yet been shown to be the result of an anaphylactic antibody-antigen reaction.

LOCAL ANAPHYLAXIS

The first observation of a local reaction in anaphylaxis was made in the rabbit by Arthus,7, 8, 9 who noted the edema and necrosis that occur at the site of the third or fourth and succeeding subcutaneous injections of the antigen. This local reaction has been designated separately from the general symptoms of anaphylactic shock in the rabbit as "the phenomenon of Arthus.

The typical local reaction of the sensitized guinea-pig consists of edema without necrosis; however, necrosis has been reported in a few cases by P. A. Lewis.1

150

Opie has called attention to the importance of the local reactions of anaphylaxis, which had been neglected in the earlier studies. He considers them as varying degrees of inflammation dependent upon the degree of the hypersensitiveness, that is, concentration of the antibodies in the animal's tissues.

Against this view in its unqualified form stands, for the present, E. F. Grove's observation that under the course of treatment prescribed by her (the Grove series) a large percentage of rabbits become sensitive to the intravenous test injection. However, some rabbits treated in this way, even when they have developed a high titered precipitating and sensitizing power of their serum, do not exhibit necrosis when the antigen is injected subcutaneously instead of intravenously. Furthermore, as Grove has found, rabbits that exhibit the Arthus phenomenon after the Arthus series of subcutaneous injections are usually not susceptible to the intravenous test injection.

Thus it would seem that other factors besides anaphylactic antibody production play a rôle in the development of the Arthus necrosis.

These observations indicate that the character of the cutaneous anaphylactic reaction depends not so much upon the degree of antibody production as upon some other alteration of the reacting tissue. A similar situation seems to exist in the allergies in human beings. Levine and Coca 148 have shown that an increased tolerance to both clinical and parenteral contact with the excitant of hay-fever can be induced without a reduction in the sensitizing reagins and, indeed, sometimes even when the reagins are increased in the blood.

ANTIANAPHYLAXIS

Animals that have been actively or passively sensitized may, by various means, be made relatively more or less resistant to the usual test injection of anaphylaxis. Such resistance is quantitatively or qualitatively different according to the manner and means of its production; the differences shown are due to differences in the underlying mechanism, and in this respect it is necessary to distinguish nonspecific and specific forms of the phenomenon. The need of a term to apply to all forms of the phenomenon of resistance to, or interference with, the injurious effects of "reinjection" is met with the familiar designation "antianaphylaxis," which was invented by Besredka.

Nonspecific Antianaphylaxis.-The fact has already been mentioned that if all of a group of guinea-pigs of nearly equal weight are sensitized, either actively or passively, with quantitatively the same preliminary injection, it will be found that the minimal lethal dose of the anaphylactogen will be very nearly the same for all of these animals. If, however, a short time previous to the test injection, some of the group of sensitized guinea-pigs receive an injection of some material other than the corresponding anaphylactogen, such as inorganic salts, peptone, foreign serum or other proteid substances, such individuals may not then succumb to the injection of the minimal lethal quantity of the anaphylactogen as determined for the other members of the group.

Such nonspecific interference-antianaphylaxis-is not absolute; that is, anaphylactic shock in full measure can be elicited in such resistant animals merely by the administration of a small multiple of the determined minimal lethal dose of the antigen. The mechanism of nonspecific antianaphylaxis is not understood. It has been a pitfall in the way of some investigators of anaphylaxis, but it possesses, at present, no other importance.

Specific Antianaphylaxis.-Two forms of specific antianaphylaxis are distinguishable both as to the manner of production and as to the mechanism of the interference. One of these forms has been designated with the term desensitization; the other form may be referred to as "masked anaphylaxis."

DESENSITIZATION.-If a guinea-pig that has been actively or passively sensitized be given, by subcutaneous injection, a suitable quantity of the respective anaphylactogen, the animal will be found, after an interval

of several hours, to have lost completely its previous hypersensitiveness, so that now its immediate behavior toward injections of the anaphylactogen is in all respects the same as that of a normal guinea-pig.

The difference between this state and that of nonspecific antianaphylaxis is clear, and it points to a difference in the mechanism of the two phenomena. This specific form of antianaphylaxis has been shown to be due solely to a specific neutralization of the antibodies situated in the sensitive tissues. This is demonstrated by the fact that after a sensitized guinea-pig has been rendered specifically quite insensitive it can at once be passively resensitized with nearly the same minimal quantity of immune serum that is required for the passive sensitization of a normal animal.

The mechanism of this form of antianaphylaxis is best designated with the term desensitization, which signifies a complete removal of the previous state of hypersensitiveness.

The specificity of desensitization may be demonstrated, in guineapigs that have been sensitized against more than one anaphylactogen, by desensitizing the animal to one of the anaphylactogens. The animal thus treated will be found to be still hypersensitive to the other anaphylactogens, though in a slightly lesser degree (nonspecific interference), as shown by a somewhat increased minimal lethal dose of the other anaphylactogens.

Desensitization can be accomplished not only by subcutaneous injection but also through the peritoneal or venous routes. It was early noted that animals which had recovered from anaphylactic shock following intravenous injections were insensitive to subsequent intraperitoneal injection of the usual intoxicating doses of the antigen. Later it was found that, by the slow injection of the dilute antigen, complete desensitization can be attained through the venous route without the production of anaphylactic shock.

The obvious explanation of the absence of symptoms during the process of desensitization has been recognized in the slowness with which the antibodies are saturated, the cause of the shock of anaphylaxis being seen, conversely, in the sudden meeting of a completely or nearly completely neutralizing amount of antigen with the antibody in the sensitive tissues.

Masked Anaphylaxis. It was early observed that if guinea-pigs which had received an actively sensitizing injection of antigen were injected a second time, shortly before the establishment of the expected hypersensitiveness, with the same material, they did not become susceptible to the usual (intraperitoneal) test injection until after a much longer period than after a single injection. This form of antianaphylaxis was particularly puzzling on account of the fact that the blood of such resistant animals contained demonstrable antibodies in greater amount than that of ordinarily sensitized animals.

The resistance of the "immune" guinea-pig was ascribed by Weil 231 to the antibodies present in the blood, the injected antigen being prevented from reaching the sensitive cells by its reaction with the cir