A critical history of idiopathic anaphylaxis
The evolution of a perplexing disease entity
Almost from the moment that the phenomenon was given a name, idiopathic anaphylaxis (IA) has been a “debated entity.” Reference [Lieberman P. Idiopathic anaphylaxis. Selected articles from the recent literature 2007. American Academy of Allergy Asthma & Immunology web site. 9 August 2007]. This article will provide a review of the medical literature related to IA and will attempt to explain why it may have been viewed that way.
1978: First IA paper published
While I once found the term idiopathic anaphylaxis used in an American Academy of Allergy Asthma & Immunology committee roster from the mid-1970s, the first medical journal publication on the subject was not published until 1978.
In that year, Edita Bacal, Roy Patterson, and C. Raymond Zeiss, all working out of the Section of Allergy — Immunology Department of Northwestern University Medical School in Chicago, submitted a paper to the journal Clinical Allergy, titled, “Evaluation of severe (anaphylactic) reactions.” It was in this paper, published in the May 1978 issue, that we first encounter the term idiopathic anaphylaxis in the medical literature. Reference [Bacal E, Patterson R, Zeiss CR. Evaluation of severe (anaphylactic) reactions. Clin Allergy. 1978; 8(3):295–304].
The authors offer case reports of 21 patients who “presented with life-threatening symptoms, appearing to be allergic in aetiology [cause of disease].” Reference [Op. cit., p. 295]. The authors do not state whether this was a random collection of patients or a group of patients who had been referred to the Allergy Section of the Immunology Department of Northwestern University Medical School because their cases were unusual or difficult to properly diagnose or treat.
In either event, the authors worked up all 21 patients and in ten cases they were able to establish a diagnosis of anaphylaxis due to drug allergy (aspirin, penicillin), food allergy (peanut, legume), systemic mast cell disease, complement C1q activation, underlying malignancy, self-inflicted nut anaphylaxis (psychiatric etiology), or "Munchausen’s stridor" (also a psychiatric etiology, in the authors’ view).
For a more contemporary take on Munchausen's stridor, see Sticky ideas in my personal blog.
From the vantage point of today, the battery of tests run on these patients seems rather minimal: medical history, physical exam, CBC [complete blood count], urinalysis, SMA 12 [blood chemistry tests for albumin, alkaline phosphatase, total bilirubin, BUN (blood urea nitrogen), calcium, cholesterol, creatinine, glucose, phosphorus, SGOT (serum glutamic-oxaloacetic transaminase), total protein, and uric acid], chest x-ray, and skin tests for inhalant and food allergens.
And what about the remaining eleven patients? They were given a diagnosis of IA, a new clinical entity, which they described like this: “In this report idiopathic anaphylaxis refers to a reaction of sudden onset which is of such severity that emergency treatment is required, or an immediate-type reaction involving more than one organ system.” Reference [Op. cit., p. 295].
If I had been writing this paper, I think I would have attempted to encourage interest in this new clinical entity, with an eye toward fostering research that could improve our understanding of its etiology. However, instead the authors adopt a strangely blasé attitude.
Right up front in the paper’s abstract, the investigators report, “In those in whom no aetiology could be found, there was usually a spontaneous subsidence of the frequency and severity of attacks or spontaneous remission. No fatalities occurred.” Reference [Op. cit., p. 295].
To me, that sounds as though they were saying, “Well, we’ve delineated this new condition, but it’s not really a big deal. It usually goes away on its own, and it hasn’t killed anyone.”
In the final paragraph of the paper’s Discussion section, the authors repeat and amplify the attitude expressed at the beginning of the paper:
Despite the often dramatic presentation of those of our series of patients with idiopathic anaphylaxis, and the enigmatic nature of the problem, the clinical course of the group was relatively benign. The majority experienced either significant diminution or complete disappearance of their symptoms. Of the eleven patients reported in whom no cause for anaphylaxis was found, five had a complete remission and are off all medication, four have had control of symptoms with medication and two continue to have some symptoms, but of decreasing severity and frequency on no medication. One was lost on follow-up. There were no fatalities in any of our patients. Thus, from this series, idiopathic anaphylaxis appears compatible with a benign prognosis, if the acute episodes are controlled until an aetiology is found or spontaneous remission occurs. Reference [Op. cit., p. 303].
That last sentence is especially interesting. It suggests that in the case of IA, either a cause for the anaphylaxis will later be identified (in which case the original diagnosis of IA was made in error) or the condition will go into a "spontaneous remission." This would suggest that the authors are of the opinion that all cases, which are truly IA, will enjoy "spontaneous remission."
It amazes me that after working up a mere eleven patients and following them for no more than a year (based on information in the included case reports), the authors could express such utter confidence that IA would always cure itself. Given that the authors had no idea what causes IA and that at that point no longitudinal studies had been done, on what evidence could they base such a prognosis?
The authors' only comment relating to a possible mechanism for IA was this: “It is conceivable that some unknown agent acts directly upon the mast cells in these patients, without intervening immunological mechanisms.” Reference [Op. cit., p. 303]. Of course, if that were the case, how could they know that this "unknown agent" would disappear within a year?
I should note that the prophylactic [preventive] treatment that worked so well for these patients consisted of prednisone [glucocorticoid, anti-inflammatory drug], hydroxyzine [1st-generation H1 antihistamine, aka Atarax], and/or ephedrine [relieves nasal congestion]. The eleven IA patients ranged in age between 11 and 51 (mean age of just over 31), and five were male, six female. Only one of the 11 patients was African American.
1979: Second IA report
The next paper published on the topic of IA came from a different group. Phil Lieberman and William W. Taylor were both affiliated with the Division of Allergy-Immunology at the University of Tennessee Center for the Health Sciences, Department of Medicine, in Memphis.
Their paper was accepted for publication in April of 1979 and came out in the September issue of Archives of Internal Medicine. Reference [Lieberman P, Taylor WW. Recurrent idiopathic anaphylaxis. Arch Intern Med. 1979; 139:1032–1034]. These authors studied 18 patients who had “recurrent, life-threatening episodes of anaphylaxis with no known cause.” Reference [Op. cit., p. 1032].
Each patient had experienced at least three episodes of anaphylaxis which included “urticaria [hives] and angioedema [large areas of deep skin edema] associated with bronchospasm [contraction of muscles that constrict the lung's air passages], laryngeal edema [swelling in the area of the layrnx or “voice box’], and/or hypotension [low arterial blood pressure].” And, the authors added, “Episodes were life-threatening in many instances.” Reference [Op. cit., p. 1032].
Patient ages ranged from 18 to 52, with 10 male and 8 female patients. Reference [Op. cit., p. 1032]. Periods over which attacks had occurred ranged between six months and 21 years, and frequency of attacks ranged between daily and every four months. Reference [Op. cit., p. 1033].
Right away I noticed a difference between these patients and those of the Bacal, et al., study. While the earlier paper suggested that IA resolved, with or without treatment, fairly rapidly, four of Lieberman and Taylor’s patients had had IA for periods ranging from six to 21 years and another ten had had symptoms for 2 to 4 years. Reference [Op. cit., p. 1033]. So two-thirds of Lieberman and Taylor's patients had had IA longer than Bacal, et al. had studied it.
Of the Lieberman and Taylor patients who were not lost to follow-up, five went into remission and five continued to have attacks. Reference [Op. cit., p. 1034]. Only half of these patients experienced the spontaneous remission that Bacal, et al., seemed to feel was inevitable.
It may be interesting to note that the five patients who went into remission had averaged between 4 and 5 attacks in all, whereas four of the five patients who had continued to have attacks had had many more attacks, ranging between 18 and “over thirty.” Reference [Op. cit., p. 1033].
Also three of the five patients who went into remission had not experienced syncope [fainting], shock [bodily cells receiving too little oxygen and blood flow], or hypotension as symptoms of their attacks, whereas the five who had continued having attacks had all experienced those symptoms. Reference [Op. cit., p. 1033].
The workup given each patient in the Lieberman and Taylor study was more extensive than that done by Bacal, et al., and included detailed histories, physical examinations, CBC, blood chemistry studies, chest x-ray, ECG [electrocardiogram] during attacks, IgE level, liver-spleen scan, bone marrow examination, biopsies of both normal and abnormal skin, CH50 [total hemolytic complement], C3 and C4 levels, histamine levels during and between attacks, skin tests to foods, airborne allergens, and penicillin, and RAST [radioallergosorbent test] to selected foods. Reference [Op. cit., p. 1033].
“In addition, each patient underwent stringent dietary manipulation using an elimination-provocation allergy diet.” Reference [Op. cit., p. 1032]. The results of all of these tests were within normal limits, with the exception of plasma histamine and ECG during attacks. Reference [Op. cit., p. 1032].
Illustrating the experience of these patients, a summary of one case was given:
The patient is a 26-year-old woman who has had a three-year history of recurrent attacks of severe urticaria and angioedema with syncope on two occasions and hoarseness, stridor [harsh or high-pitched noise when breathing, caused by obstruction], wheeze, and dysphagia [difficulty in swallowing] on several occasions. Episodes have occurred as often as daily. Although she has lost exact count, during the past three years there have been well over 30 episodes requiring epinephrine injections for urticaria and dyspnea [labored breathing or shortness of breath] associated with stridor or syncope. The patient has been seen by four different allergists. She has been hospitalized on two occasions, once for a diagnostic evaluation and once because of a severe anaphylactic hypotensive episode. All the studies previously noted were performed on this patient. Penicillin skin testing was performed because of the suspicion that the episodes could have been due to penicillin exposure while at work. All studies were normal except for an elevated plasma histamine level during an attack (40 ng/dL) that returned to normal between attacks and a skin biopsy specimen from an urticarial lesion that showed changes consistent with urticaria but no increase in mast cells.
Because of the life-threatening nature of the attacks and the fact that their frequency interfered with her work, the patient was hospitalized and placed on a diet consisting only of water. She had two episodes during the week of hospitalization. Each episode was observed by a physician and in neither instance, to the best of our knowledge, did the patient have access to food or drugs. The attacks have not been eliminated by 30 mg of prednisone orally daily and 50 mg of diphenhydramine hydrochloride (Benadryl) orally four times a day. At last follow-up, she continues to have approximately one attack per week requiring the self-injection of epinephrine. Reference [Op. cit., p. 1032–3].
This woman’s situation does not seem to fit the “benign prognosis” predicted by Bacal, et al., and Lieberman and Taylor comment, “…Unlike our series, in which five of our ten patients continue to have attacks even after having received medication to prevent them, Bacal et al were able to at least partially stabilize the majority of their cases….” Reference [Op. cit., p. 1033].
The authors also note that plasma histamine was elevated in two of their patients in whom it was measured during an attack, while all five of the patients whose histamine levels were measured when not in an attack exhibited normal results. They suggest that, “This implies, in the absence of any identifiable antigen or provocative agent, spontaneous mast cell and basophil degranulation.” Reference [Op. cit., p. 1033].
The authors noted that patients “responded well to acute therapy for anaphylaxis,” most requiring only epinephrine and Benadryl, although on five occasions, hospitalization was necessary. “Instruction of the patients in the self-injection of epinephrine has, in most instances, eliminated the need for emergency-room visits.” Reference [Op. cit., p. 1034].
However, Lieberman and Taylor note, “Prophylaxis of attacks has been more difficult.” In two severe cases, where 30 mg of prednisone was prescribed, “There was no detectable benefit in either case. This is in contrast to the series of Bacal and associates.” Reference [Op. cit., p. 1034]. This observation is very important because in the ensuing years the researcher-clinicians at Northwestern University Medical School would highlight responsiveness to corticosteroid therapy as the hallmark of IA, as opposed to anaphylaxis from other causes. Reference [For example: Greenberger PA. Idiopathic anaphylaxis. Immunol Allergy Clin N Am. 2007; 27:273–93].
While encountering differences in prophylaxis and severity in their patients, Lieberman and Taylor agree, “Fortunately, the problem subsides spontaneously in many instances. Five of the ten patients we are still following up have undergone a spontaneous remission.” Reference [Lieberman P, Taylor WW. Recurrent idiopathic anaphylaxis. Arch Intern Med. 1979; 139:1032–1034]. This is a much lower rate of remission than Bacal, et al., reported, but given the relatively small samples of patients in these two early studies, the differences may not be statistically significant.
1980: Response to the Lieberman-Taylor paper
In searching the medical literature, I found no public response to the Bacal, et al., paper, and only a single letter in response to that of Lieberman and Taylor.
In the June 1980 issue of the Archives of Internal Medicine, Dr. D.J. Salberg, of Allen Park, Michigan, wrote to the editor to make the following fairly minor points:
- He quibbled with the use of the term "anaphylaxis," given that there was no evidence of the reactions being IgE-mediated. He suggested the the use of the term "anaphylactoid" instead. Reference [Salberg DJ. Idiopathic anaphylaxis and histamine levels [letter]. Arch Intern Med. 1980; 140:864.].(Note: Current practice is to use the term anaphylaxis, regardless of whether the cause of a reaction is IgE-mediate, immune in nature but not IgE-mediated, or non-immune.)
- He disputed the authors’s assertion that the "normal plasma histmaine level is less than 10 ng/mL." He noted that other investigators had established that "normal plasma histamine content is actually less than 1 ng/mL," although he acknowledged the essential role of "meticulous handling of samples" in achieving accuracy in measurement of plasma histamine levels. Reference [Salberg DJ. Idiopathic anaphylaxis and histamine levels [letter]. Arch Intern Med. 1980; 140:864.].
1981: Urinary histamine in IA
A 1981 paper by Gwendolyn Myers, Mildred Donlon, and Michael Kaliner outlined a methodology for measuring the small amount of intact histamine excreted in urine. The authors note, “Two patients with systemic mastocytosis and two with idiopathic anaphylaxis had elevated urine histamine levels. Monitoring urine histamine may be useful in assessment of conditions in which histamine plays a role.” Reference [Myers G, Donlon M, Kaliner M. Measurement of urinary histamine: development of methodology and normal values. J Allergy Clin Immunol. 1981; 67:305–11. Abstract only.].
|Previous section||Next section|
1980s: A cautionary note and more publications
A 1984 editorial by Dr. Albert L. Sheffer in the New England Journal of Medicine — a prestigious journal read by doctors from a variety of medical specialties — discussed a case report in the same issue about a woman who had become sensitized to progesterone administered to her to prevent abortion and who later reacted allergically to progesterone produced within her own body.
He commented, “This case report clearly emphasizes the importance of a detailed historical assessment in the identification of etiologic factors in idiopathic anaphylaxis.” Reference [Sheffer AL. Unraveling the mystery of idiopathic anaphylaxis. NEJM. 1984; 311(19):1248–1249.].
While acknowledging that anaphylactic reactions could be related to hapten-induced IgG production, he seemed to be cautioning against thinking that anaphylaxis could be truly idiopathic. This is a sentiment that I have heard repeated by many doctors, including some who specialize in allergy and immunology. The idea of repeated non-allergenically mediated anaphylaxis seems to be one that a fair number of doctors find very hard to accept.
This is an important point: Resistance to the idea that anaphylaxis can be truly idiopathic constitutes one of the reasons why some doctors look askance at the diagnosis — and adds yet another reason to why being diagnosed with IA can make a patient feel as though he or she has grown a second head.
Progesterone sensitivity and recurrent anaphylaxis
The case report that prompted Sheffer’s remarks was authored by a group of researchers that included Michael Kaliner, Dean Metcalfe, and William J. Meggs. Reference [Meggs WJ, Pescovitz OH, Metcalfe D, et al. Progesterone sensitivity as a cause of recurrent anaphylaxis. N Engl J Med 1984; 311:1236–8].
Here is how the authors summarize the case:
Anaphylactic or anaphylactoid attacks usually occur in immediate response to a specific inciting agent or event. Patients with recurring anaphylaxis in whom there is no evidence of an external cause are classified as having recurrent idiopathic anaphylaxis.… One such patient had weekly life-threatening anaphylactic episodes that were partly controlled with intensive medical therapy combined with tracheal fenestration [incision into the trachea to open an airway, today often called a tracheotomy]; however, spontaneous remission of the attacks during lactation led to the suspicion of sensitivity to sex hormones. Three clinical observations confirmed this suspicion: provocative challenges with minute quantities of progesterone caused an anaphylactic event, inhibition of pituitary gonadotropin release by luteinizing hormone-releasing hormone analogue controlled the disease for nine months, and oophorectormy [removal of one or both ovaries] was curative. Reference [Op. cit., p. 1236].
The woman, who was 36 at the time of the case report, was described as having had “a lifelong history of seasonal allergic rhinitis and asthma, previous anaphylactic reactions to penicillin and streptomycin, and a strong family history of allergy.” Reference [Ibid.]. For two years she was diagnosed with “chronic ‘idiopathic’ urticaria [hives],” but then she began having acute attacks of anaphylaxis every five to ten days. The severity of her upper airway angioedema [swelling] required a permanent opening in her trachea. The authors report that “Medications that did not produce noticeable improvement included acetylsalicylic acid, indomethacin, vitamin E, prednisolone (up to 100 mg per day for three months), and oral disodium cromglycate.” Reference [Ibid.].
The patient’s gynecological history, “…Included intermittent use of oral contraceptives, seven pregnancies with four miscarriages, and parental progesterone administration during the completed pregnancies.” Reference [Ibid.]. Less than a year after the onset of her attacks of anaphylaxis, the patient once again became pregnant. She had only a few minor attacks in her first trimester but in the second and third trimesterds she had attacks every three to seven days. After the birth of a normal, healthy boy, while the patient was breastfeeding the child for seven months, she had no attacks of anaphylaxis. It was only after she spontaneously stopped lactating that the bouts of anaphylaxis returned.
The researchers performed tests in an attempt to simulate the hormonal conditions during lactation. Most interventions worsened her condition, until they tried “a long-acting analogue of luteinizing hormone-releasing hormone” (LHRHa). Reference [Op. cit., p. 1237]. The LHRHa therapy controlled her attacks for nine months, until she developed symptoms that led to a hysterectomy that included removal of her ovaries. Thereafter, without any medications, the patient remained attack-free for the nine months prior to the publication of the case report. Reference [Ibid.].
In discussing the case, the authors offer the idea that perhaps prior administration of injected or oral medications somehow contributed to her developing hypersensitivity to her own progesterone. A point not made by the authors, but perhaps inferred by Dr. Sheffer and readers of the case report, was that if one looked hard enough and long enough, a cause might be found for a purportedly idiopathic case of recurrent anaphylaxis.
Food skin testing in IA
In 1986 a research group from the Mayo Clinic contributed a paper on skin testing for food allergies to the literature on IA. Reference [Stricker WE, Anorve-Lopez E, Reed CE. Food skin testing in patients with idiopathic anaphylaxis. J Allergy Clin Immunol. 1986; 77:516-9.]. Stricker, Anorve-Lopez and Reed began with a group of 168 patients who were seen at the Mayo Clinic between 1980 and 1984. All received a set of 79 skin print tests “to help evaluate and identify an antigen source responsible for systemic anaphylactic reactions….”
The researchers chose food substances for the skin prick tests that have been reported to cause anaphylaxis, along with a selection of spices and condiments, including seeds. As an aside, the foods that caused skin prick reactions were (listed in decreasing order of the number of positive reactions):
- Chamomile (8 positive reactions)
- Aniseed, castor bean, millet (5 each)
- Celery and filbert (4 each)
- Horseradish, mushroom, mustard, poppy seed, shrimp, and sunflower (3 each)
- Almond, brewer’s yeast, cashew nut, chestnut, coconut, crab, ginger, hops, lobster, mango, pea, and sesame seed (2 each)
- Artichoke, baker’s yeast, beet, brazil nut, caraway seed, chicory, clam, cottonseed, fennel, flaxseed, garbanzo bean, halibut, honey, lentil, lima beans, milk, nutmeg, pistachio, thyme, tumeric, and walnut (1 each)
- Allspice, apple, banana, bay leaf, black pepper, buckwheat, cataloupe, chicken, chili pepper, chocolate, cinnamon, clove, cod, corn, cumin seed, dill seed, egg, garlic, juniper berry, orange, oyster, peach, peanut, potato, psyllium seed, raspberry, sage, salmon, soybean, strawberry, sweet potato, tangerine, tapioca, vanilla (zero positive reactions) Reference [Op. cit., p. 517].
Only 102 of the patients were judged to have “symptoms severe enough to be included in the present study.” Reference [Op. cit., p. 516]. Five other patientes were found to have “ sensitivities to various dyes and preserves, including one patient with sodium metabisulfite sensitivity,” and were eliminated from the study group. The remaining patients ranged in age from 10 to 73 years old, with roughly equal number of men and women.
After the patients were skin-prick tested, they were challenged with the foods to which they showed a positive reaction. If the patient developed a severe systemic reaction within four hours, that was considered to be a positive result. Two additional patients were “considered as having a positive clinical correlation because elimination of a skin test positive food was followed by complete elimination of anaphylaxis for longer than 24 months.” Reference [Op. cit., p. 518]. Of all the patients included in the study, 32 (31%) had one or more positive skin tests, while “seven (7% of the 102 patients) had a positive clinical correlation.” Reference [Ibid.].
The conclusions that Stricker, et al., draw from this study are interesting. First, they note that if skin testing had not been done, seven percent of the patients would have been diagnosed with IA when they were actually having reactions to food to which they were sensitive. They also note that the patients who had a positive clinical correlation to a positive skin test were significantly younger (averaging 27.7 years) than the group as a whole (average age of 40.8 years). They conclude: “This may indicate that one should be more aggressive in skin testing younger individuals in search of an obscure food antigen.” (It should also be noted that this study predated the commercial availability of a wide array of reagents for RAST testing.)
The researchers note that while patients whose positive skin-prick testing correlated with the clinical result of ingesting the food did very well if they eliminated those foods from their diets, patients who had positive skin-prick tests without clinical correlation “have not fared as well.” Even if they eliminated all of the foods that tested positive (via skin prick), they showed no improvement in the frequency or severity of their attacks. They add, “It is not clear why some of these skin test positive antigens provoked anaphylaxis and others did not.” Reference [Ibid.].
While extensive skin testing of food allergens revealed meaningful food reactions in only 7% of the 102 patients studied, the authors end their report with the recommendation that a battery of uncommon food antigen skin tests be used to ensure that someone is not diagnosed with IA who actually suffers from an uncommon food allergy.
Northwestern continues IA research
I have no way of knowing whether Sheffer’s comments gave voice to feelings that were widespread at that time, but between 1983 and 1989, the group at Northwestern churned out at least seven papers related to IA. That level of activity suggests either a kind of campaign to increase acceptance of the idea of IA or an attempt to establish a kind proprietary research ownership of the subject. I do not have access to any of these papers, but since the leader of that effort, the late Dr. Roy Patterson, later edited a volume summarizing his group’s research on IA that I do have, I hope that I will not omit any significant contribution to understanding made by these researchers during the mid- to late-1980s.
|Previous section||Next section|
To summarize what I've learned from the sources I've found:
Page last updated: May 30, 2011