Nobel Lecture, December 12, 1934

The Development of Liver Therapy in Pernicious Anemia

The idea that something in food might be of advantage to patients with pernicious anemia was in my mind in 1912, when I was a house officer at the Massachusetts General Hospital, as is noted in certain case records there. Ever since my student days, when I had the opportunity, in my father’s wards at the Massachusetts General Hospital, to distinguish from pernicious anemia two cases of chronic hereditary hemolytic jaundice, I have taken a deep interest in this disease. I watched many patients pass through relapses and remissions, and observed that, despite treatment with arsenic, blood transfusions, splenectomy, and other procedures, all eventually died. Prolonged observation permitted me to become acquainted with the multiple variations and many aspects of the disease, and to realize that from a few cases it was difficult to determine the effect of therapeutic procedures.

The study of the patients’ diets was begun in 1915 in an attempt to determine if some sort ‘of dietary deficiency could be found. The similarity of certain symptoms and signs of pernicious anemia to those in pellagra, sprue, and beriberi was appreciated, as was the fact that certain sorts of anemia were occasionally associated with a faulty diet. Elders, among others, suggested in 1922 that such a state of affairs existed in pernicious anemia. Furthermore, the almost constant occurrence of achlorhydria in pernicious anemia, which appears usually long before the anemia and remains in spite of liver therapy, led me to wonder if this disorder of the digestive system had something to do with the condition which might be in the nature of a dietary deficiency disease. Indeed, Fenwick, about 1880, suggested the possible primary role of the stomach, but it remained for Castle, in 1928, to demonstrate the part this organ plays in the causation of the disease.

The possibility of an excess of fats in the diet leading to excessive blood destruction was considered, but therapy with low-fat diets was futile. Later the effects of high-caloric: feeding with an excess of protein, derived especially from meat, were studied at about the same time that other physicians, such as Barker, reported some benefit from this treatment. The results were not impressive ‘but were perhaps suggestive.

Although Pepper in 1875 and Cohnheim in 1876 recognized that the bone marrow was abnormal, there was a prevailing opinion in the early part of this century that abnormal blood destruction played an important or primary role in the production of the disease. Nevertheless it was believed by many physicians, as I was taught, that the production of blood by the bone marrow was also deeply implicated. About 1919 the late Dr. James Homer Wright taught me to appreciate the character of the abnormality of the bone marrow in pernicious anemia, which led me to believe firmly that something was needed to make the primitive red cells that crowd the bone marrow in relapse grow to normal cells; and that it was of no particular value to aim treatment towards stopping what has been called excessive blood destruction in this disease. In 1922 Whipple suggested that in pernicious anemia there might be a scarcity of material from which the stroma of the red blood cells was formed, or that there existed a disease of the stroma-forming cells of the bone marrow. This concept fitted with the idea that there was a deficiency of something in the body and that dysfunction of pigments metabolism was resultant, or of secondary importance.

‘For centuries the concept that food bore a relationship to anemia had been vaguely expressed in the literature. It had been shown that liver and kidneys, rich in complete proteins, promoted the growth of animals, and that substances in liver could enhance cell division. It was likewise recognized that liver-feeding could benefit patients with sprue (Manson, 1883) and pellagra. These were among the reasons that led to the choice of liver as a substance likely to enhance blood formation. Of invaluable importance was Whipple’s fundamental and classical work on hemoglobin regeneration by means of liver and other foods in anemia due to blood loss in dogs. He has now placed upon a secure quantitative basis the influence of food upon anemia.

A few patients were fed relatively small amounts of liver during 1924 and early 1925. Although these first patients did better than expected, the results permitted no more than speculations. Then Dr. Murphy joined in the work and we pursued the study of these and subsequent cases. Liver had been fed by Gibson and Howard and other individuals to pernicious anemia patients but without persistence or definite results. It seemed to us that to accomplish our object a large weighed amount of liver should be fed daily with regularity. Likewise to determine the effect it was considered essential that data should be obtained in a large number of cases to be appropriately compared with controls. By May, 1926, we had fed liver intensively and daily to 45 patients. In many of these patients symptomatic improvement was obvious within about a week. Soon they craved food, and color appeared in their faces. Tongue and digestive symptoms rapidly lessened. Within about 60 days the red blood cells counts had risen on the average from low levels to approximately normal. Dr. Murphy will describe to you in more detail the improvement that takes place. I wish especially to call to your attention the fact that an objective measure of the effects upon blood production was the chief basis of our conclusions that by feeding liver, significant improvement had been obtained. I refer especially to counts of new adult and young red blood cells (reticulocytes) appearing, as Peabody’s studies demonstrated later, as a result of the maturation of the immature cells crowding the bone marrow.

The next step naturally was to attempt to determine the nature of the constituent in liver responsible for the effects, and to learn if an extract for therapeutic use could be obtained. Dr. Edwin J. Cohn, of the Department of Physical Chemistry in the Laboratories of Physiology of the Harvard Medical School, soon made a potent extract suitable for oral use. We tested on patients the preparations he prepared in an attempt to isolate the active principle. Although unsuccessful in this objective, as time passed we demonstrated (1929) that the potent material could be given intravenously, and produced maximal effects in very small quantities (0.15 g). These small experimental preparations were not practical for regular use, and it remained for Gänsslen in Germany to produce the first practical extract for parenteral therapy. Since then numerous individuals have prepared and studied such preparations. Extracts for parenteral use can be easily made by dissolving in water a powdered extract (Fraction G of Cohn) commonly used in America. The exact nature of the potent substance remains unknown, but especially from the studies of Cohn and West it appears to be a relatively small nitrogenous compound.

Extract given parenterally is at least 50 and perhaps 100 times as potent as extract given by mouth and it assures the individual of receiving into his body proper the material he lacks. It thus permits one to give large amounts easily: e.g., a few cubic centimeters of fluid a week, instead of about 1,500 to 3,000 grams of liver. The physician must know what a given amount of a given preparation may be expected to accomplish under usual circumstances. He must recognize that extraction causes loss of potency and that the potency of a good preparation is usually no more than 70 per cent of that of raw liver. Because one preparation is three times as concentrated as another, it by no means follows that the amount of potent material contained is three times as great. It may actually be less potent. Active material is not confined to liver. It has been found to occur in stomach tissue by Sturgis and Isaacs, and is found in kidney, brain, placenta, and probably other organs. Castle has shown that a reaction between an unidentified substance in normal gastric juice and material in certain foods rich in the vitamin B complex yields potent material. By utilization of the principle described by Castle, Reimann has shown that liver may have its potency increased by incubation with gastric juice. This has resulted in the practical use of combinations of gastric and liver tissue.

Since the presentation of our original work, the regular beneficial results of liver therapy have been confirmed in many parts of the world; for example, in Sweden by Strandell and in Copenhagen by Meulengracht. There have been differences of opinion regarding the value of one or another preparation, but these differences can largely be explained if a proper comparison of dosage is made. There are cases requiring much more potent material than others and in some cases there is difficulty of absorption, making parenteral therapy essential. As in other deficiency disorders there are factors, such as infections, arteriosclerosis and serious damage to vital organs, which inhibit the action of “liver extract” and when they are present it is often necessary to give unusually large amounts of the active principle. The results of treatment will be essentially the same if, irrespective of its source for the given case, enough potent material enters the body throughout life. Indeed, pernicious anemia, like other deficient states, should be treated on a quantitative basis. Failure of liver therapy in a case diagnosed pernicious anemia implies inadequate treatment, an incorrect diagnosis, or the existence of complication sufficiently serious in itself to be disastrous for the patient.

The treatment should not consist in supplying enough material to remove only one symptom, such as anemia, but enough both to supply indefinitely all demands of the body for the substance and to fill it with an adequate reserve supply. The disease affects the gastro-intestinal and neural systems as well as the hemopoietic tissue. Probably more material is required to improve or inhibit the progress or development of neural lesions than to permit blood element to be maintained. With proper dosage, symptoms due to neural lesions usually lessen, sometimes strikingly, as Dr. Murphy and I originally noted. Unfavourable reports concerning the effects on the neural system can be attributed to failure to realize that there is no standard dose of “liver extract”, and that there is great variation in the potency of different preparations; also that sepsis and other complications can inhibit the action of liver on the nervous tissue, as it does on blood formation. For proof of the effect of liver on neural lesions one must not so much study the degree of improvement, as seek for evidence of the arrest of the process. In my clinic this has been done especially by Dr. Strauss. There have been observed critically over from two to three years about 100 cases, including 26 cases with advanced combined system-disease, treated parenterally with liver extract. In no instance did a single objective neurologic sign become more marked nor did an abnormal sign develop. Abnormal signs sometimes decreased in intensity or became normal. There was thus objective arrest of the neural lesions in every case and subjective improvement in all. Indeed, certain patients who were originally unable to walk became sufficiently improved to return to their occupations. Failure to arrest the progress of neural lesions, like failure to restore the blood to normal and keep it there, usually means that more liver is required. Sometimes such severe infection or the like is present, that arrest cannot be achieved. One should not interpret parenteral therapy as acting in any fundamentally different manner from oral therapy. Certain cases treated for over eight years orally have had no progress of their neural lesions. Many cases can be adequately treated orally, but parenteral therapy is simple and permits suitable amounts of active principle easily and assuredly to enter the patient’s body.

Liver therapy has been shown to be of value in other anemias than the idiopathic pernicious anemia originally described by Addison in 1849. There are other substances in liver and in certain liver extracts potent for pernicious anemia besides the principle active in pernicious anemia. Dr. Whipple has demonstrated a factor in liver aiding blood regeneration in “secondary a anemia. This factor can affect blood formation in certain clinical cases but as yet has received little critical study in man. Liver extract potent for pernicious anemia is of value in such other macrocytic anemias as arise in sprue (Bloomfield and Wycoff, Ashford and others), pregnancy (Wills, Strauss, and Castle), coeliac disease (Vaughan et al.) and the like. Partly because we recognized in pernicious anemia that, in addition to improvement of the blood, the alimentary tract and nervous systems were benefited, a still wider application has resulted. The possibility of controlling the gastro-intestinal symptoms in sprue by the use of parenteral therapy with liver extract has been demonstrated by Castle and Rhoads. For this purpose it is of great value. Likewise in pellagra the gastro-intestinal symptoms are responsive. The improvement of the skin lesions in this disease, which also occurs, suggests the possibilities of a broader application to other conditions. It was suggested, in 1913, by Vogel and McCurdy, that determinations of the numbers of the young red blood cells (reticulocytes) might measure within a brief time the effect of therapy. Although I had studied these cells from my student days in many patients, it was not until Dr. Murphy’s and my observations were well under way, that the significance of the reticulocyte reaction as an index of the effect of potent material was fully appreciated. It then became an important aid in the evaluation of therapy and in the subsequent development of effective liver extracts. It was soon recognized that, following the administration of effective doses of liver, the reticulocytes in pernicious anemia are increased for a few days, during which time their numbers follow a distinctive course. This reticulocyte reaction is orderly and simulates the curve for growth and death of organisms and their cells. Critical daily observation of these reactions and their proper interpretation thus gives useful information concerning both the state of the patient and the potency of the material administered.

A reaction that does not conform in character to that due to liver and in proper time relation to the administration of the substance under test, should always be looked upon with suspicion. The reticulocyte reaction induced by liver indicates a bone marrow reaction to a physiological need and not to a stimulant: a normal reaction to the existing anemia. A reticulocyte response does not necessarily mean that specific material which the body lacks has been supplied, for there are other substances and conditions that cause reticulocytosis in pernicious anemia, which do not regularly promote normal blood formation. For example, responses may be induced by potassium arsenite (Fowler’s solution), but they usually differ very much from the physiological responses to liver. Non-specific responses, however, may closely resemble those due to supplying the deficient substance. One must be particularly critical of the nature of responses caused by substances given parenterally, because non-specific responses of various sorts can arise more readily from parenteral injection than from feeding.

It is important that the potency of products used in life-saving procedures be assured. The reticulocyte reaction is serviceable in yielding information regarding the strength of liver extracts and potent substitutes employed clinically. The factors which physiologically influence the reticulocyte reaction, also influence the rate of total red cell regeneration; but, as noted in 1927, the amount of material necessary to induce maximal reticulocyte rises is often less than the amount necessary for the maximal rate of increase of red cells. In testing products one must aim to give an amount expected to yield less than a maximal reticulocyte response, since otherwise considerable losses in potency may occur and remain undetected. The percentage of reticulocytes at the peak of their rise for a given red cell level has been used in comparative data. It is, of course, difficult to determine whether one agent is more effective than another, unless tests have been made in the same way. Very different types of curves for reticulocytes will be obtained from daily administration and from one relatively large parenteral injection, so that reticulocyte peaks cannot be compared directly. In either instance, however, the total number of young red cells poured forth during the reaction will be similar.

Because of the considerable variation in the reactivity of cases, more information can undoubtedly be obtained by comparative tests of known and unknown material in the same patient. This can be accomplished by observing the reticulocyte responses in successive ten-day periods of uniform daily administration of each substance. The dosage of the substance of known potency given first must be such that a submaximal reticulocyte response will result. If so much material has been given that a maximal reticulocyte response has occurred, an increase of potent material in the second period cannot induce a second reticulocyte rise. In a properly conducted test the occurrence of any orderly reticulocyte response to a second substance means that it is of greater potency than the first material given.

There is need for an animal or laboratory procedure to test the potency of products and to aid in determining the nature of the substance or substances effective in pernicious anemia. Numerous studies of this sort have been made; for example, by Vaughan on pigeons. Jacobson’s recent report concerning the use of reticulocyte reactions in guinea pigs is suggestive of fruitful results.

In the early days, as has been noted, it was thought that pernicious anemia might be a dietary deficiency disease. The demonstration of the effectiveness of liver therapy indicated that this was very probably the case and thus provided the proper orientation towards a study of its cause. Castle in my clinic, from work entirely his own, has shown that it is a deficiency disease of a special sort, one which may be spoken of as a conditioned dietary deficiency disease. He and his associates have demonstrated that because of a specific defect of the gastric secretion the patient with pernicious anemia is unable to carry out an essential reaction with certain constituents of food. In the normal individual this reaction is necessary for the production of the supply of “liver extract” and thus for the prevention of the disease. The material is absorbed and stored in the liver and certain other tissues, so that when such animal tissues are fed to human beings, they receive the material needed in pernicious anemia without the necessity of the special reaction within their stomachs. In pernicious anemia it has been shown (Ivy and others) that “liver extract” is absent in the patient’s liver. The factor in the normal gastric secretion responsible for the reaction has not been identified with any of its recognized constituents. Meulengracht has shown from what portion of the stomach the gastric factor is formed in pigs. Strauss and Castle have demonstrated that the food factor is associated with a number of natural sources of vitamin B, although, from the work of Wills, Lassen and others, it is probable that it is not a portion of the vitamin B complex.

As a logical consequence of this work Castle and his associates have postulated the significance of defects of the dietary factor, the gastric factor, and of difficulty in the absorption or utilization of substances promoting blood formation in the production of other types of macrocytic anemia responding to liver extract. Thus, in certain instances of the tropical macrocytic anemia of sprue (Castle and Rhoads) and in the anemia of pregnancy studied by Wills in India, a dietary defect seems of primary importance. The macro-cytic anemia resulting from total ablation of the stomach or its destruction by’ cancer, like Addisonian pernicious anemia, is especially associated with loss of the gastric factor. In the macrocytic anemia of certain cases of late sprue, of coehac disease and of rare instances of intestinal stenosis or multiple anastomoses, difficulty in absorption plays a major role. In the anemia of fish-tapeworm infestation these factors may also be involved. Theoretically there could occur a disorder of the internal metabolism of “liver extract”, and an occasional case of macrocytic anemia might be explained on such an assumption. Today we realize also that the participation of these factors may be variable and temporary, as occurs in the pernicious anemia of pregnancy of the temperate zone (Strauss and Castle). During pregnancy the gastric factor may be absent, only to reappear after delivery. Macrocytic anemia in animals has been produced by Wills, and by Rhoads and Miller as a result of dietary defects, and Bence has shown that the liver of gastrectomized pigs becomes deficient in “liver extract”.

In man, however, dietary deficiency is seldom confined strictly to one factor, nor are the results of disturbances of gastric secretion, of defects of intestinal absorption or of utilization necessarily concerned in only one type of metabolic process. Clearly such disturbances are involved in the production of iron deficiency resulting in certain types of hypochromic anemia. We have noted that combined deficiency of iron and “liver extract” is not rare in the same individual. Other double and even multiple deficiencies occur. In one rare and striking case I have seen the tongue and skin lesions of pellagra subside upon the oral administration of yeast, the edema from protein deficiency vanish when beefsteak was fed and finally the macrocytic anemia disappear rapidly when liver extract was given orally.

Castle and his associates have shown that the gastric factor may return toward normal in pernicious anemia after treatment with liver extract. It is probable that the gastric reaction proceeds somewhat according to the law of mass action, so that very little of the intrinsic factor of the stomach might produce with a large amount of the extrinsic factor of the food, and vice versa, material for absorption sufficient to meet to a significant degree the demands of the body. Such a state of affairs could explain the responses in pernicious anemia obtained by feeding large amounts of autolyzed yeast-extract, as shown, for example, by Ungley. Here it is probable that traces of the intrinsic factor were present; and Isaacs, Goldhamer, and Sturgis have shown that traces, rather than complete absence, of this factor are apt to occur in Addisonian pernicious anemia. In this disease, however, it is almost always the gastric factor that is grossly deficient rather than the dietary factor; but both substances may be involved. We have seen patients who formerly had satisfactory diets, develop pernicious anemia lasting many months or a few years after partaking of distinctly undesirable diets. In such instances the poor diet probably precipitated the onset of the disease. It is thus not difficult, in the light of modern knowledge, to understand the probable causation of the “spontaneous” remissions and relapses in pernicious anemia. Probably in the natural course of the disease the gastric factor slowly declines but fluctuates in quantity from time to time. The extrinsic factor will also vary, and certain foods and rest may stimulate an increase of the intrinsic factor. Infection and severe gastritis could be responsible for temporary decrease of the gastric factor. The occasional patient who, after a relapse, remains without liver therapy in a state of complete remission for years, probably represents an instance of a temporary marked decline in the gastric factor, unless grossly deficient diet is shown to have existed.

Patients undergoing liver therapy for pernicious anemia sometimes, of their own accord, omit treatment and may remain in apparent good health for many months, but many more of such individuals soon find they are sick again. The former perhaps have accumulated a considerable reserve supply of “liver extract” in their bodies and have probably been enabled by treatment to manufacture some potent material from an ordinary diet. Sooner or later a very large number of these patients will relapse, if proper treatment is not resumed. The grave error in treatment is to prescribe too little liver extract or potent substitute. When there is doubt, more rather than less should be given. It is essential that the individual receive into his body indefinitely and with regularity enough potent material for his given case. The physician, however, must do more for his patient than prescribe a proper amount of liver, stomach, or the like; he should attend to all aspects of the case and not neglect attention to the individual’s manifold problems of thought and action.

I have attempted to outline Dr. Murphy’s and my contribution to the work for which the Caroline Institute has honored us. I have pointed out to you, also, some of the studies that have been made as the result of demonstrating with Dr. Murphy that liver feeding is dramatically effective for pernicious anemia patients. It seems to me that one may expect in the future more information to be obtained which, directly or indirectly, will follow as the result of these observations. Thus, upon the foundations laid by previous investigators, do medical art and science build a structure which will in its turn be the foundation of future knowledge.