Anaphylaxis

Introduction
Background

Portier and Richet first coined the term anaphylaxis in 1902 when a second vaccinating dose of sea anemone toxin caused a dog's death. They named this response anaphylaxis, which is derived from the Greek words a - (against) and – phylaxis (immunity, protection).

Anaphylaxis is an acute multiorgan system reaction, potentially fatal, caused by the release of chemical mediators from mast cells and basophils.1,2 The most common organ systems involved include the cutaneous, respiratory, cardiovascular, and gastrointestinal systems.

Anaphylaxis has no universally accepted clinical definition. The traditional nomenclature for anaphylaxis reserves the term anaphylactic for IgE-dependent reactions and the term anaphylactoid for IgE-independent events, which are clinically indistinguishable. The World Allergy Organization, which is an international umbrella organization representing more than 70 national and regional professional societies dedicated to allergy and clinical immunology, has recommended replacing this terminology with immunologic (IgE-mediated and non-IgE-mediated [eg, IgG and immune complex complement-mediated]) and non-immunologic anaphylaxis.3

Clinically, anaphylaxis is considered likely to be present if any 1 of the 3 following criteria is satisfied within minutes to hours:
Acute symptoms involving skin, mucosal surface, or both, and at least one of the following: respiratory compromise, hypotension, or end-organ dysfunction
Two or more of the following occur rapidly after exposure to a likely allergen: hypotension, respiratory compromise, persistent gastrointestinal symptoms, or involvement of skin or mucosal surface
Hypotension develops after exposure to an allergen known to cause symptoms for that patient: age-specific low blood pressure or decline of systolic blood pressure of more than 30% compared to baseline

In clinical practice, however, delaying treatment until the development of symptoms affecting multiple organs is risky, since the ultimate severity of anaphylaxis is difficult to predict from its outset.

Click here to read the updated practice parameter on the diagnosis and management of anaphylaxis from the Joint Task Force on Practice Parameters; American Academy of Allergy, Asthma and Immunology; American College of Allergy, Asthma and Immunology; and Joint Council of Allergy, Asthma and Immunology. Note that guidelines for the emergency medical treatment of anaphylaxis vary internationally.4

Pathophysiology

When mast cells and basophils degranulate, whether by IgE- or non–IgE-mediated mechanisms, preformed histamine and newly generated leukotrienes, prostaglandins, and platelet activating factor (PAF) are released. The physiologic responses to these mediators include smooth muscle spasm in the respiratory and gastrointestinal tract, vasodilation, increased vascular permeability, and stimulation of sensory nerve endings.

These physiologic events lead to some or all of the classic symptoms of anaphylaxis: flushing; urticaria/angioedema; pruritus; bronchospasm; laryngeal edema; abdominal cramping with nausea, vomiting, and diarrhea; and feeling of impending doom. Concomitant signs and symptoms can include rhinorrhea, dysphonia, metallic taste, uterine cramps, light-headedness, and headache. Hypotension, cardiac arrhythmias, syncope, and shock can result from intravascular volume loss, vasodilation, and myocardial dysfunction. Increased vascular permeability can produce a shift of 35% of vascular volume to the extravascular space within 10 minutes.

Additional mediators activate other pathways of inflammation: the neutral proteases, tryptase and chymase; proteoglycans such as heparin and chondroitin sulfate; and chemokines and cytokines. These mediators can activate the kallikrein-kinin contact system, the complement cascade, and coagulation pathways. The development and severity of anaphylaxis also depend on the responsiveness of cells targeted by these mediators. IL-4 and IL-13 are cytokines important in the initial generation of antibody and inflammatory cell responses to anaphylaxis. No comparable studies have been conducted in humans, but anaphylactic effects in mice depend on IL-4Rα-dependent IL-4/IL-13 activation of the transcription factor, STAT-6 (signal transducer and activator of transcription 6).5 Eosinophils may be inflammatory (release cytotoxic granule-associated proteins, for example) or anti-inflammatory (metabolize vasoactive mediators, for example).

Under rigid experimental conditions, histamine infusion alone is sufficient to produce most of the symptoms of anaphylaxis. Histamine mediates its effects through activation of histamine 1 (H1) and histamine 2 (H2) receptors. Vasodilation, hypotension, and flushing are mediated by both H1 receptors and H2 receptors. H1 receptors alone mediate coronary artery vasoconstriction, tachycardia, vascular permeability, pruritus, bronchospasm, and rhinorrhea. H2 receptors increase atrial and ventricular contractility, atrial chronotropy, and coronary artery vasodilation. H3 receptors in experimental models of canine anaphylaxis appear to influence cardiovascular responses to norepinephrine. The importance of H3 receptors in humans is unknown.
Effects on the Cardiovascular System

Anaphylaxis has been associated clinically with myocardial ischemia, atrial and ventricular arrhythmias, conduction defects, and T-wave abnormalities. Whether such changes are related to direct mediator effects on the myocardium, to exacerbation of preexisting myocardial insufficiency by the adverse hemodynamic effects of anaphylaxis, to epinephrine released endogenously by the adrenals in response to stress, or to therapeutically injected epinephrine is unclear.

Since mast cells accumulate at sites of coronary atherosclerotic plaques, and immunoglobulins bound to mast cells can trigger mast cell degranulation, some investigators have suggested that anaphylaxis may promote plaque rupture, thus risking myocardial ischemia. Stimulation of the H1 histamine receptor may also produce coronary artery vasospasm. PAF also delays atrioventricular conduction, decreases coronary artery blood flow, and has negative inotropic effects.

Calcitonin gene-related peptide (CGRP), a sensory neurotransmitter that is widely distributed in cardiovascular tissues, may help to counteract coronary artery vasoconstriction during anaphylaxis. CGRP relaxes vascular smooth muscle and has cardioprotective effects in animal models of anaphylaxis.

While tachycardia is the rule, bradycardia can occur during anaphylaxis. Thus, bradycardia may not be as useful to separate anaphylaxis from a vasodepressor reaction as previously thought. Relative bradycardia (initial tachycardia followed by decreased heart rate despite worsening hypotension) has been reported in the setting of experimentally induced insect sting anaphylaxis.

Two distinct physiologic responses occur in mammals experiencing hypovolemia.6 The initial response to hypovolemia is a baroreceptor-mediated increase in overall cardiac sympathetic drive and a concomitant withdrawal of resting vagal drive, which together produce peripheral vasoconstriction and tachycardia. When effective blood volume decreases by 20-30 percent, a second phase follows, which is characterized by withdrawal of vasoconstrictor drive, relative or absolute bradycardia, increased vasopressin, further catecholamine release as the adrenals become more active, and hypotension. Hypotension in this hypovolemic setting is independent of the bradycardia, since it persists when the bradycardia reverses with atropine administration.

The implications of relative or absolute bradycardia in human anaphylaxis and hypovolemic shock have not been studied. However, one retrospective review of approximately 11,000 trauma patients found that mortality was lower with the 29 percent of hypotensive patients who were bradycardic when they were compared to the group of hypotensive individuals who were tachycardic, after adjustment for other mortality factors.7 Thus, bradycardia may have a specific compensatory role in these settings.

Conduction defects and sympatholytic medications may also produce bradycardia. Excessive venous pooling with decreased venous return (also seen in vasodepressor reactions) may activate tension-sensitive sensory receptors in the inferoposterior portions of the left ventricle, thus resulting in a cardio-inhibitory (Bezold-Jarisch) reflex that stimulates the vagus nerve and causes bradycardia.
Frequency
United States

The true incidence is unknown. Moneret-Vautrin et al reviewed the published literature and stated that severe anaphylaxis affects at least 1-3 persons per 10,000 population.8 Neugut et al estimated that 1-15% of the US population is at risk of experiencing an anaphylactic or anaphylactoid reaction.9 They estimated that the rate of actual anaphylaxis to food was 0.0004%, 0.7-10% for penicillin, 0.22-1% for radiocontrast media (RCM), and 0.5-5% after insect stings.

A population-based study from Olmsted County, Minnesota, found an average annual incidence of anaphylaxis of 21 cases per 100,000 person-years.10 Ingestion of a suspect food was the cause in 36% of cases; a medication, subcutaneous immunotherapy (SCIT), or a diagnostic agent was the cause in 17% of cases; and an insect sting was the cause in 15% of cases. Thirty-two percent of cases were considered idiopathic. Episodes of anaphylaxis occurred more frequently from July through September, which is attributable to insect stings.

In a study of patients referred to a university-affiliated allergy-immunology practice in Memphis, Tennessee, food was the cause of anaphylaxis in 34% of patients, medications in 20%, and exercise in 7% (anaphylaxis due to insect stings or SCIT were excluded from the study).11 The cause of anaphylaxis could not be determined in 59% of patients (ie, they were diagnosed with idiopathic anaphylaxis). A separate study estimated the number of cases of idiopathic anaphylaxis in the United States to be 20,000-47,000 cases per year (approximately 8-19 episodes per 100,000 person-years).

International

The incidence of anaphylaxis does not appear to vary significantly between countries. Two European studies detected a lower average annual incidence than found in the Olmsted County study (3.2 cases of anaphylactic shock per 100,000 person-years in Denmark; 9.8 cases of out-of-hospital anaphylaxis per 100,000 person-years in Munich, Germany12 ). Rates in Europe range from 1-3 cases per 10,000.13,12 Simons and colleagues examined the rate of epinephrine prescriptions for a population of 1.15 million patients in Manitoba, Canada, and found that 0.95% of this population was prescribed epinephrine, an indicator of perceived risk that future anaphylaxis may occur.14
Mortality/Morbidity
Fatalities from anaphylaxis are infrequent but not rare. Estimates range from 0.65-2% of patients with anaphylaxis.15,16 The case-fatality rate from the Olmsted County study was 0.65%.10 Reactions to foods are thought to be the most common cause of anaphylaxis when it occurs outside of the hospital and are estimated to cause 150 deaths per year in the United States. Severe reactions to penicillin occur with a frequency of 1-5 cases per 10,000 patient courses, with fatalities in 1 case per 50,000-100,000 courses. Insect sting anaphylaxis causes approximately 25-50 deaths per year. Anaphylaxis to radiocontrast media (RCM) was estimated to have caused 500 deaths in 1982, although this number has likely decreased because of increased awareness and the use of pretreatment regimens and/or lower osmolar agents for patients with a history of RCM reaction.
In the United Kingdom, half of fatal anaphylaxis episodes have an iatrogenic cause (eg, anesthesia, antibiotics, RCM), while foods and insect stings each account for a quarter of the fatal episodes.
The most common causes of death are cardiovascular collapse and respiratory compromise. One report examined 214 anaphylactic fatalities for which the mode of death could be surmised in 196, 98 of which were due to asphyxia (49 lower airways [bronchospasm], 26 both upper and lower airways, and 23 upper airways [angioedema]). The fatalities from acute bronchospasm occurred almost exclusively in those with preexisting asthma. Another analysis of 23 unselected cases of fatal anaphylaxis determined that 16 of 20 “immediate” deaths (death occurring within one hour of symptom onset) and 16 of the 23 cases that underwent autopsy were due to upper airway edema.
Death can occur rapidly. An analysis of anaphylaxis fatalities occurring in the United Kingdom from 1992 to 2001 revealed the interval between initial onset of food anaphylaxis symptoms and fatal cardiopulmonary arrest averaged 25-35 minutes, which was longer than for drugs (mean, 10-20 minutes pre-hospital; 5 minutes in-hospital) or for insect stings (10-15 minutes).
Race
Race has no known effect on the risk of anaphylaxis.
Sex
In the Olmsted County study, men and women were equally affected.10
The Memphis study showed a slight female predominance.11
Earlier studies have suggested that episodes of anaphylaxis to intravenous muscle relaxants, aspirin, and latex are more common in women, while insect sting anaphylaxis is more common in men. These sex discrepancies are likely a function of exposure frequency.
Age
Anaphylaxis can occur at any age. In the Olmsted County study, the age range was 6 months to 89 years.10 The mean age was 29 ±19 years. The Memphis study had an age range of 1-79 years, with a mean of 37 years.11
Simons and colleagues noted the highest frequency of epinephrine prescriptions for boys aged 12-17 months (5.3%).14 The rate was 1.4% for those younger than 17 years, 0.9% for those aged 17-64 years, and 0.3% for those aged 65 years or older.
Severe food allergy is more common in children than in adults. However, the frequency in adults may be increasing, since severe food allergy often persists into adulthood.
Anaphylaxis to radiocontrast media (RCM), insect stings, and anesthetics has been reported to be more common in adults than in children. Whether this is a function of exposure frequency or increased sensitivity is unclear.

Other risk factors
Atopy is a risk factor. In the Olmsted County study, 53% of the patients with anaphylaxis had a history of atopic diseases (eg, allergic rhinitis, asthma, atopic dermatitis).10 The Memphis study detected atopy in 37% of the patients.11 Other studies have shown atopy to be a risk factor for anaphylaxis from foods, exercise-induced anaphylaxis, idiopathic anaphylaxis, radiocontrast reactions, and latex reactions. Underlying atopy does not appear to be a risk factor for reactions to penicillin or insect stings.
Route and timing of administration affect anaphylactic potential. The oral route of administration is less likely to cause a reaction, and such reactions are usually less severe, although fatal reactions occur following ingestions of foods by someone who is allergic. The longer the interval between exposures, the less likely an anaphylactic (IgE-mediated) reaction will recur. This is thought to be due to catabolism and decreased synthesis of allergen-specific IgE over time. This does not appear to be the case for IgE-independent reactions.
Geographic location previously was thought to be irrelevant, but a British report has challenged that assumption. Analysis of hospital discharge summaries after anaphylaxis suggests that individuals living in rural areas and in the southern and western parts of England have an increased incidence of anaphylaxis.
Asthma is a risk factor for fatal anaphylaxis.
Delayed administration of epinephrine is also a risk factor for fatal outcomes.17
Posture also influences anaphylaxis outcomes. In a retrospective review of pre-hospital anaphylactic fatalities in the UK, the postural history was known for ten individuals.18 Four of the ten fatalities were associated with the assumption of an upright or sitting posture during anaphylaxis. Postmortem findings were consistent with pulseless electrical activity and an "empty heart" attributed to reduced venous return from vasodilation and redistribution of intravascular volume from the central to the peripheral compartment.
Clinical
History

In most studies, the frequency of signs and symptoms of anaphylaxis is grouped by organ system. For example, 100% of patients with anaphylaxis in the Olmsted County study had cutaneous manifestations, consistent with the study's definition of anaphylaxis, which required one symptom of generalized mediator release (mostly skin manifestations).10 Nevertheless, other studies have reported that 90% of patients have cutaneous involvement. The Olmsted study observed that 69% had respiratory manifestations, 41% had cardiovascular involvement, and 24% had oral or gastrointestinal manifestations.10 Other studies have reported similar findings.

Children, however, may be different. An Australian study evaluated 57 children under age 16 years who presented to a pediatric emergency department over a three-year period. Cutaneous features were noted in 82.5%, whereas 95% had respiratory symptoms.

Patients often initially describe a sense of impending doom, accompanied by pruritus and flushing. This can evolve rapidly into the following symptoms, broken down by organ system:
Cutaneous/ocular - Flushing, urticaria, angioedema, cutaneous and/or conjunctival pruritus, warmth, and swelling
Respiratory - Nasal congestion, rhinorrhea, throat tightness, wheezing, shortness of breath, cough, hoarseness
Cardiovascular - Dizziness, weakness, syncope, chest pain, palpitations
Gastrointestinal – Dysphagia, nausea, vomiting, diarrhea, bloating, cramps
Neurologic - Headache, dizziness, blurred vision, and seizure (very rare and often associated with hypotension)
Other – Metallic taste, feeling of impending doom
Symptoms usually begin within 5-30 minutes from the time the culprit antigen is injected but can occur within seconds. If the antigen is ingested, symptoms usually occur within minutes to 2 hours. In rare cases, symptoms can be delayed in onset for several hours. Anaphylaxis, however, occurs as part of a clinical continuum. It can begin with relatively minor cutaneous symptoms and rapidly progress to life-threatening respiratory or cardiovascular manifestations. In general, the more rapidly anaphylaxis develops after exposure to an offending stimulus, the more likely the reaction is to be severe.

Physical

The first priority should be to assess the patient's airway, breathing, circulation, and adequacy of mentation (eg, alertness, orientation, coherence of thought).

Respiratory
Severe angioedema of the tongue and lips may obstruct airflow.
Laryngeal edema may manifest as stridor or severe air hunger.
Loss of voice, hoarseness, and/or dysphonia may occur.
Bronchospasm, airway edema, and mucus hypersecretion may manifest as wheezing. In the surgical setting, increased pressure of ventilation can be the only manifestation of bronchospasm.
Hypoxia can cause altered mentation.
Cardiovascular
Tachycardia is present in one fourth of patients, usually as a compensatory response to reduced intravascular volume or to stress from compensatory catecholamine release.
Bradycardia is more suggestive of a vasodepressor (vasovagal) reaction, although it has been observed in anaphylaxis (see Pathophysiology).
Hypotension (and resultant loss of consciousness) may be observed secondary to capillary leak, vasodilation, and hypoxic myocardial depression.
Cardiovascular collapse and shock can occur immediately, without any other findings. This is an especially important consideration in the surgical setting.
Cutaneous
Cutaneous findings may be delayed or absent in rapidly progressive anaphylaxis. Urticaria (hives) can occur anywhere on the body, often localizing to the superficial dermal layers of the palms, soles, and inner thighs. The lesions vary in size and are erythematous, raised, and highly pruritic.
Angioedema (soft tissue swelling) is also commonly observed. These lesions involve the deeper dermal layers of skin. It is usually nonpruritic and nonpitting. Common areas of involvement are the larynx, lips, eyelids, hands, feet, and genitalia.
Generalized (whole-body) erythema (or flushing) without urticaria or angioedema is also occasionally observed.
Gastrointestinal: Vomiting, diarrhea, and abdominal distention are frequently observed.
Causes

Immunologic

IgE-mediated anaphylaxis: This is the classic form of anaphylaxis, whereby a sensitizing antigen elicits an IgE antibody response in a susceptible individual. The antigen-specific IgE antibodies then bind to mast cells and basophils. Subsequent exposure to the sensitizing antigen causes cross-linking of cell-bound IgE, resulting in mast cell (and/or basophil) degranulation. Typical examples of IgE-mediated anaphylaxis include the reactions to many drugs, insect stings, and foods.
Certain drugs cause IgE-mediated anaphylaxis. Most cases of IgE-mediated drug anaphylaxis in the United States are due to penicillin and other beta-lactam antibiotics.
Penicillin is metabolized to a major determinant, benzylpenicilloyl, and multiple minor determinants. Penicillin and its metabolites are haptens, small molecules that only elicit an immune response when conjugated with carrier proteins.
Other beta-lactam antibiotics may cross-react with penicillins or may have unique structures that also act as haptens. The incidence rate of anaphylaxis to cephalosporins in penicillin-anaphylactic patients appears to be much less than the 10% frequently quoted. Pichichero reviewed this complicated literature and offers specific guidance for the use of cephalosporins in patients who have a history of IgE-mediated reactions to penicillin.19
Patients with less well-defined reactions to penicillin have a very low risk (1-2%) of developing anaphylaxis to cephalosporins. The rate of skin-test reactivity to imipenem in patients with a known penicillin allergy is almost 50%. In contrast, no known in vitro or clinical crossreactivity exists between penicillins and aztreonam.
Many other drugs have been implicated less frequently in IgE-mediated anaphylaxis.
In the surgical setting, anaphylactic reactions are most often due to muscle relaxants and latex but can also be due to hypnotics, antibiotics, opioids, colloids, and other agents. Volatile anesthetic agents can cause immune-mediated hepatic toxicity but have not been implicated in anaphylactic reactions.
Insect stings, that is, venoms from Hymenoptera insects (eg, bees, yellow jackets, hornets, wasps, fire ants), can elicit an allergen-specific IgE antibody response. From 0.5%-3% of the US population experiences a systemic reaction after being stung.20
Hypersensitivity to foods is a problem encountered throughout the industrialized world.17 In the United States, an estimated 4 million Americans have well-substantiated food allergies. In Montreal, 1.5% of early elementary school students were found to be sensitized to peanuts. Reactions to foods are thought to be the most common prehospital (outpatient) cause of anaphylaxis and are estimated to cause 125 deaths per year in the United States.
Certain foods are more likely than others to elicit an IgE antibody response and lead to anaphylaxis. Foods likely to elicit an IgE antibody response in all age groups include peanuts, tree nuts, fish, and shellfish. Foods likely to elicit an IgE antibody response in children also include eggs, soy, and milk.
An analysis of 32 fatalities thought to be due to food-induced anaphylaxis revealed that peanuts likely were the responsible food in 62% of the cases. In placebo-controlled food challenges, peanut-sensitive patients can react to as little as 100 µg of peanut protein.21 The Olmsted County study, in agreement with earlier studies, found that food ingestion was the leading cause of anaphylaxis, accounting for as many as one third of all cases.10 Scombroid fish poisoning can occasionally mimic food-induced anaphylaxis. Bacteria in spoiled fish produce enzymes capable of decarboxylating histidine to produce biogenic amines, including histamine and cis-urocanic acid, which is also capable of mast cell degranulation.
Latex hypersensitivity is a phenomenon that has been recognized in the last 20 years, corresponding with the increased use of latex gloves because of the HIV and hepatitis B and C epidemics and the institution of universal precautions. In 1995, an estimated 8-17% of healthcare professionals were at risk for latex reactions. The incidence rate is decreasing, at least in part, because of increased awareness, improved manufacturing practices, and a change to unpowdered latex and nonlatex gloves.
Allergen-specific subcutaneous immunotherapy (SCIT) can cause IgE-mediated anaphylaxis. Allergy injections are a common trigger for anaphylaxis. This is not unexpected because the treatment is based on injecting an allergen to which the patient is sensitive. However, life-threatening reactions are rare. Three studies suggest that fatalities from SCIT occur at a rate of approximately 1 per 2,500,000 injections.22,23,24 A total of 104 fatalities due to SCIT and skin testing were reported from 1945-2001.

Risk factors for severe anaphylaxis due to immunotherapy include poorly controlled asthma, concurrent use of beta-blockers, high allergen dose, errors in administration, and lack of a sufficient observation period following the injection. Near-fatal reactions (NFR) to subcutaneous immunotherapy also have been examined retrospectively. Of 646 allergist-immunologists who responded to a survey on reactions, 273 reported NFR. The investigators defined an NFR as respiratory compromise, hypotension, or both, requiring emergency epinephrine. Hypotension was reported in 80% and respiratory failure occurred in 10% of NFRs, exclusively in subjects with asthma. Epinephrine was delayed or not administered in 6% of these cases.

Aspirin and nonsteroidal anti-inflammatory drugs
Aspirin and nonsteroidal anti-inflammatory drugs (NSAIDs) in the past have been classified as IgE-independent because reactions were thought to occur from aberrant metabolism of arachidonic acid. Isolated cutaneous reactions to aspirin/NSAIDs and bronchospasm in aspirin-sensitive asthmatics (often in association with nasal polyposis) are indeed mediated through IgE-independent mechanisms. Blockade of cyclooxygenase by these drugs causes the prostanoid pathway to shut down, resulting in an overproduction of leukotrienes via the 5-lipoxygenase pathway. These patients have marked cross-reactivity between aspirin and most NSAIDs.
Anaphylaxis after taking these drugs apparently occurs via a different mechanism that is more consistent with IgE-mediated anaphylaxis. With true anaphylaxis, the different cyclooxygenase inhibitors do not appear to cross-react. Anaphylaxis only occurs after 2 or more exposures to the implicated drug, suggesting a need for prior sensitization. Finally, patients with true anaphylaxis do not usually have underlying asthma, nasal polyposis, or urticaria. In one study of nearly 52,000 people taking NSAIDs, 35 developed anaphylactic shock.

Immunologic, IgE-independent reactions
Complement-mediated reactions
Anaphylaxis resulting from administration of blood products, including intravenous immunoglobulin, or animal antiserum is due, at least in part, to activation of complement. Immune complexes formed either in vivo or in vitro can activate the complement cascade.
Certain byproducts of the cascade, namely plasma-activated complement 3 (C3a), plasma-activated complement 4 (C4a), and plasma-activated complement 5 (C5a), are called anaphylatoxins and are capable of causing mast cell/basophil degranulation.
Exercise-induced anaphylaxis
This is a rare syndrome that can take one of two forms. The first form is ingestant dependent, requiring both exercise and the recent ingestion of particular foods (eg, wheat, celery) or medications (eg, NSAIDs) to cause an episode of anaphylaxis. In these patients, exercise alone does not produce an episode, and ingesting the culprit food or medication alone does not cause an episode.
The second form is characterized by intermittent episodes of anaphylaxis during exercise, independent of any food ingestion. Anaphylaxis does not necessarily occur during every episode of physical exertion.
Anaphylaxis associated with systemic mastocytosis
Anaphylaxis can be a manifestation of systemic mastocytosis, a disease characterized by excessive mast cell numbers in multiple organs.
Such patients appear to be at increased risk for food and venom reactions. Alcohol, vancomycin, opioids, radiocontrast media (RCM), and other biologic agents that can degranulate mast cells directly are discouraged.
Idiopathic anaphylaxis25
This is a syndrome of recurrent anaphylaxis for which no consistent triggers can be determined despite an exhaustive search. This recurrent syndrome should be distinguished from a single episode of anaphylaxis for which the etiology may be unclear.
Most patients treated with antihistamines and steroids have complete remission following tapering of steroids. Others require long-term prophylaxis with high doses of H1 antihistamines.
Idiopathic anaphylaxis can be categorized as infrequent (<6 episodes per year) or frequent (³6 episodes per year or 2 or more episodes within the last 2 months).25 One approach is expectant treatment with epinephrine, antihistamines, and prednisone for individuals who have infrequent episodes and a prolonged taper of prednisone for those with frequent episodes.
Most of these patients are female, and atopy appears to be an underlying risk factor.
Two thirds of patients have 5 or fewer episodes per year, while one third have more than 5 episodes per year.
A subpopulation of women develops anaphylaxis in relationship to their menstrual cycle; this phenomenon is known as catamenial anaphylaxis.26,27 In severe cases, these patients require manipulation of their hormonal levels by medical pituitary suppression and even oophorectomy. Most of these patients react to shifts in progesterone levels, and the diagnosis can be confirmed by provoking an anaphylactic event through administration of low doses of progesterone.
Nonimmunologic reactions
Certain agents (ie, direct mast cell activators) are thought to cause direct, nonimmunologic release of mediators from mast cells. These include opioids, RCM, dextrans, protamine, and vancomycin. Mechanisms underlying these reactions are poorly understood but may involve specific receptors (eg, opioids) or non–receptor-mediated mast cell activation (eg, hyperosmolarity). Evidence also exists that RCM, dextrans, and protamine can activate several inflammatory pathways, including complement, coagulation, and vasoactive (kallikrein-kinin) systems.

http://emedicine.medscape.com/article/135065-overview