Anaphylaxis is a severe, life-threatening, generalized or systemic hypersensitivity reaction. While there is agreement as to this definition of anaphylaxis, the clinical presentation is often variable and it is not uncommon for there to be debate after the event as to whether anaphylaxis had actually occurred. The management of anaphylaxis falls into two distinct phases: (1) emergency treatment and resuscitation of a patient with acute anaphylaxis and (2) the search for a cause for the event and the formulation of a plan to prevent and treat possible further episodes of anaphylaxis. Both aspects are important in preventing death from anaphylaxis and are covered in this review.
Anaphylaxis is distinguished from other allergic reactions by the severity and systemic nature of the reaction. However, it is not clear why one person with specific immunoglobulin E (IgE) to an allergen will have an anaphylactic reaction on exposure, another individual only a local reaction and in a third individual no reaction . Some risk factors have been defined, such as low levels of platelet activating factor (PAF) acetylhydrolase and low levels of serum angiotensin-converting enzyme, both of which independently increase the risk of an allergic individual developing anaphylaxis on allergen exposure [2,3]. Local and systemic allergic reactions occur via similar mechanisms that differ in location and magnitude. It should be noted that fatal allergic reactions can occur without anaphylaxis being present. For example, angioedema affecting the upper airway may be a lethal local reaction and other reactions may kill by inhalation of vomit. Anaphylaxis results from the release of mediators by mast cell and basophils during degranulation, which entails more than just the release of toxic mediators such as histamine and heparin. Also released are proteases such as tryptase and chymase, cytokines and chemokines including interleukin (IL)-3, IL-4, IL-5, IL-13, granulocyte–macrophage colony-stimulating factor, tumour necrosis factor-α and chemokine ligand 3, lipid mediators such as leukotrienes and PAF and kinins such as bradykinin. Different combinations of the above mediators may be released in different circumstances . Many of these mediators are preformed and stored in the granules, whereas others are produced de novo on activation of mast cells and basophils. Degranulation can be mediated by cross-linking of IgE bound to membrane high-affinity IgE receptor (FcεRI), or by non-IgE-mediated mechanisms. The distinction between these mechanisms can be important diagnostically, but in practice their clinical presentation and management of the acute emergency they cause are indistinct.
The clinical presentation of anaphylaxis is variable and there continues to be debate about its clinical definition [5,6]. Many different organ systems may be affected. The skin may itch (pruritus) with or without weals (urticaria) and/or swelling (angioedema). There may be nausea, abdominal pain, vomiting and/or diarrhoea. Swelling may involve the lip, tongue, throat and/or upper airway impairing swallowing (dysphagia), speech (dysphonia) or breathing (with stridor and/or asphyxiation). There may be sneezing, runny nose (rhinorrhoea) and itching of the external ear canal. The lungs can be affected with cough, wheeze and bronchospasm with a corresponding fall in the peak expiratory flow rate. Cardiovascular events include hypotension, fainting (syncope), altered mental state and chest pain. In addition to marked anxiety, the patient may experience an ‘impending sense of doom’. Notwithstanding the debate around exactly what constitutes anaphylaxis, it is agreed that it represents a systemic rather than local reaction, and that it is severe and potentially life-threatening. There appears to be a consensus that for the term ‘anaphylaxis’ to be used there should have occurred in an appropriate clinical context a physiologically significant disturbance of one or more of the airway, breathing or circulation (‘ABC’). This pithy ABC definition is of great practical help in informing and advising patients so that they may recognize potentially life-threatening reactions in order to self-manage them appropriately (see below) and ensures that all agencies which patients may access issue uniform, clear, non-confused medical advice to patients.
Anaphylactic versus anaphylactoid – a dangerous distinction
The terms ‘anaphylactic’ and ‘anaphylactoid’ should be avoided. Both involve mast cell and basophil stimulation and result in identical clinical consequences. The belief held by some that anaphylactoid reactions are not as severe is not true, as both are potentially fatal and require (identical) emergency treatment. Delay in treating a reaction because it is labelled ‘anaphylactoid’ can be life-threatening. For this reason many advocate that the term anaphylactoid should be abandoned. The European consensus terms are allergic anaphylaxis (i.e. IgE-mediated anaphylaxis) and non-allergic anaphylaxis (i.e. non-IgE-mediated anaphylaxis).
Allergic (IgE-mediated) anaphylaxis results from the cross-linking of specific IgE bound to membrane FcεRI by the allergen, or in other words type 1 hypersensitivity by the Gell and Coombs classification . The breaking of immunological tolerance to otherwise harmless allergens with consequent production of allergen-specific IgE is not the subject of this review. Although this occurs more often in patients with co-existent eczema or asthma, it can occur in any individual.
Non-allergic (non-IgE)-mediated anaphylaxis occurs when mast cells and basophils are activated directly by processes that appear to bypass the need for membrane FcεRI cross-linking. The mechanisms by which such reactions occur are less well understood, but clearly imply cellular activation via other cell surface receptors or actions at intracellular target sites. Such anaphylactic reactions may occur, for example, to radiocontrast media, salicylates, IgA and opioid drugs [9,10].
Acute management of anaphylaxis
The evidence base for the management of acute anaphylaxis is limited, given the ethical and practical difficulties inherent in performing randomized clinical trials in medical emergencies. It is thus unsurprising that guidelines for the treatment of anaphylaxis vary . However, in all protocols and guidelines adrenaline is the mainstay of treatment (Fig. 1). This is true regardless of the cause of anaphylaxis, although there are separate guidelines for the management of anaphylaxis associated with administration of drugs during general anaesthesia , as such reactions can be managed in environments with immediate availability of intensive monitoring and life-support by highly skilled staff.
Ambiguity about the definition of anaphylaxis should not lead to a delay in its recognition with consequent delayed or inadequate treatment. A broad definition of anaphylaxis is most useful in the emergency setting, such as that from the Academy of Allergology and Clinical Immunology Nomenclature Committee: ‘Anaphylaxis is a severe, life-threatening, generalized or systemic hypersensitivity reaction’. This definition covers both IgE-mediated and non-IgE-mediated anaphylaxis.
Adrenaline should be administered promptly to patients having anaphylaxis with life-threatening disturbance of airway, breathing or circulation. It should be given by intramuscular injection, as this leads to more rapid absorption (with maximal plasma concentration in 8 min) and higher plasma concentrations than administration by subcutaneous injection [14,15]. Adrenaline should be administered at the mid-point of the anterolateral thigh (the bottom right of the right trouser pocket) with a needle long enough to reach the muscle . Adult autoinjectors contain 300 µg of adrenaline, and junior/paediatric versions contain 150 µg (Fig. 2). An adult autoinjector is recommended for children weighing > 30 kg. When injecting adrenaline, a needle long enough to reach the muscle should be used: for many people, this is longer than the needle in the autoinjectors and so health care professionals giving adrenaline are recommended to use vial/needle/syringe in most circumstances. Furthermore, in the United Kingdom the recommended dose of adrenaline is 0·5 mg as opposed to the 0·3 mg for self-treatment. Intravenous adrenaline should be given only by those experienced in its administration, when the patient is monitored intensively .
Although there are no randomized clinical trials for the use of adrenaline in anaphylaxis, current opinion favours its use when anaphylaxis is likely and there is a significant physiological disturbance of A, B or C . It is important that treatment with adrenaline is prompt, as delayed administration appears to be associated with poorer survival [19,20]. However, the administration of adrenaline carries risks, including cardiac arrhythmias, myocardial infarction and hypertensive intracerebral bleeds. Therefore, adrenaline should not be used in non-life-threatening allergic reactions .
A number of drugs can interact adversely with adrenaline (Table 1) Beta-blockers (used commonly to treat high blood pressure and angina) not only reduce the efficacy of adrenaline, but can lead to paradoxical bronchoconstriction, hypertension and bradycardia after adrenaline because of unopposed alpha-receptor stimulation . Tricyclic anti-depressant drugs and monoamine oxidase inhibitors increase the risk of cardiac arrhythmias from adrenaline administration. However, because of the potentially life-saving effect of adrenaline in anaphylaxis, administration of adrenaline is still recommended . The time to consider drug reactions with adrenaline is not during an acute episode of anaphylaxis when prompt treatment is life-saving. All patients considered to be at risk of anaphylaxis and certainly all patients who have been prescribed adrenaline autoinjectors should have their regular drug therapy reviewed. Drugs that interact with adrenaline should be substituted by others or be prescribed only with caution if the benefits of their continued use outweigh the risks of interaction with adrenaline should the patient have an episode of anaphylaxis.
Table 1. Drugs that may interact with adrenaline.
May antagonize hypertensive effects of adrenaline
Beta-blockers (including eye drops)
(1) Reduce efficacy of adrenaline (2) Unopposed alpha-receptor stimulation can lead to bronchoconstriction, bradycardia and severe hypertension
Possible risk of hypertension
Dopaminergics (entacapone and rasagiline)
Possibly enhance effects of adrenaline
Ergotamine and methysergide
Increased risk of ergot poisoning
Linezolid is a reversible, non-selective monoamine oxidase inhibitor, see below
Monoamine-oxidase inhibitors (both MAO-A and MAO-B)
Increased risk of arrhythmias and hypertension (may cause hypertensive crisis)
Increased risk of arrhythmias and hypertension
Patients with anaphylaxis become hypotensive because of a central effect of the reaction on the heart or a peripheral effect because of a combination of vasodilation and leakage of fluid from the intravascular compartment. Shocked patients must be kept lying down and may benefit from raising their legs. In patients with an insufficient response to adrenaline, a rapid intravenous fluid challenge with crystalloid should be given and the response monitored. Large volumes of fluid may be needed during the resuscitation of a patient with anaphylaxis. There is no evidence to favour the use of colloids or crystalloids over the other in the treatment of anaphylaxis. Indeed, some colloids (e.g. gelofusin) can cause anaphylaxis, and this should be considered in patients where there appears to be a link between the onset of anaphylaxis and a colloid infusion, or in those receiving colloids who fail to improve.
Anti-histamines may help in the treatment of anaphylaxis but the evidence for their use is weak . Theoretically, they may help with histamine-mediated pathology, but not with the effects resulting from other mediators, and may have limited efficacy in preventing ongoing mast cells and basophil activation . While most protocols recommend their use, it is important that they are viewed as a second-line treatment and that the administration of anti-histamines does not delay treatment with adrenaline and fluids [11,17].
Corticosteroids have a role in the management of anaphylaxis, but their use should not delay the administration of adrenaline or fluid resuscitation. Their actions require de-novo protein synthesis and are thus, by definition, delayed by at least some hours. They are therefore of no immediate benefit, but are likely to be therapeutically useful (in view of their multiple anti-inflammatory actions) in minimizing later pathology that might otherwise follow the acute reaction. The severity of the immediate reaction does not predict whether there will also be a biphasic (i.e. delayed, later) response and therefore it appears logical and wise to administer steroids to all patients with anaphylaxis [25,26]. There is little clinical trial evidence relating to the efficacy of the administration of steroids for anaphylaxis, but the early administration of steroids in patients suffering from acute asthma has been demonstrated to be beneficial .
The patient should be laid flat provided that breathing is not uncomfortable in that position. Placing a patient with low blood pressure in a seated or upright position may be dangerous and should be avoided . Light-headedness or loss of consciousness indicate low blood pressure, which may be helped by improving the return of blood to the heart by lying flat and elevating the legs if required. Other treatments that should be administered depend upon their availability and clinical need. High-flow oxygen should be used if available to maintain blood oxygen saturations. Wheeze or asthma should be treated with bronchodilators (nebulized β-blockers if available). If reacting to an infused drug or colloid, its administration should be stopped. Vasopressors and inotropes other than intramuscular adrenaline (e.g. metaraminol, noradrenaline, vasopressin, glucagon) should be used only under appropriate specialist supervision [17,29].
Investigating the cause of an episode of anaphylaxis
Did anaphylaxis occur?
After appropriate resuscitation from the acute episode, the first question that requires answering is: ‘Was this acute episode caused by anaphylaxis or by something else?’. The differential diagnosis of anaphylaxis includes asthma, other causes of shock (e.g. cardiogenic shock), other cardiovascular or respiratory events, urticaria without anaphylaxis, hereditary angioedema and psychiatric disorders such as panic attacks. Hereditary angioedema is caused by the autosomal dominant inheritance of C1 esterase inhibitor deficiency and leads to episodes of angioedema (usually without rash). The condition is investigated by the measurement of serum complement C3 and C4 levels and C1 esterase inhibitor levels and function [30–32].
It can often be difficult to determine whether anaphylaxis had occurred in a patient being seen subsequently in a specialized immunology or allergy clinic, especially if the notes from the time of the event are unavailable or inadequate, as there is only the patient's (and/or relatives) recollection of events upon which to base one's clinical judgement. It is thus most important for any attending doctor (at a hospital accident and emergency department or elsewhere) to document events carefully at the time, including the extent and duration of any physiological disturbances of breathing, oxygenation, pulse and blood pressure. It is also important for blood to be taken following resuscitation during the acute event for later testing to seek objective evidence of mast cell activation as the explanation for the acute event. Histamine has a very short half-life and following anaphylaxis serum histamine levels peak at 5 min and return to baseline by 15–30 min. Its lability and the unforgiving ex-vivo sample handling requirements for histamine assay make serial changes in serum histamine concentration levels impractical to monitor. In contrast, assay of tryptase released from mast cells during their activation is a very helpful assay to perform. Current assays measure native tryptase, the concentration of which peaks at around 2 h after anaphylaxis, with a serum half-life in vivo of around 2 h [33–35]. Thus 4 ml samples of clotted blood taken following immediate resuscitation (when convenient), and at 2 h after the onset of symptoms show elevated tryptase levels in most patients with anaphylaxis when compared with levels checked 24 h later (or baseline levels checked when the patient is referred to a specialized clinic for investigation) . Such a pattern is strongly suggestive of systemic mast cell degranulation, with the release of inflammatory mediators into the circulation. However, not all anaphylactic reactions are associated with a rise in tryptase to levels greater than the quoted reference ranges. The sensitivity of using tryptase to detect anaphylaxis can be increased if the percentage difference between the baseline and peak concentration are considered, regardless of whether or not the peak concentration is outside the reference range [34,36].
A persistently raised tryptase level occurs in mastocytosis, a disorder that involves an increase in mast cell numbers, and may be associated with myeloproliferation which can present with variable symptoms and signs including headache, diarrhoea, flushing, urticaria, angioedema, hypotension and shock. A number of triggers can lead to significant mediator release in mastocytosis, such as stings (even in the absence of IgE-mediated allergy), stress, drugs (aspirin, opioids, muscle relaxants, antibiotics), radiographic contrast media and alcohol [35,37].
Deaths from anaphylaxis may be underestimated, as there may be few signs of an allergic reaction having occurred on post-mortem examination . Death in anaphylaxis is caused either by shock (which may be misdiagnosed as myocardial infarction) or asphyxia (which may be misdiagnosed as asthma). The Royal College of Pathologists’ Guidelines on Autopsy Practice recommend a number of measures when anaphylaxis is suspected, including examination of the heart, coronary artery, lungs and vocal cords, the securing of pre-mortem blood samples where possible and the sampling of cadaveric blood for specific IgEs and tryptase . Post-mortem tryptase can be useful in the investigation of fatal anaphylaxis. Tryptase remains stable in serum for many days after death, so that there is no need to take a sample within hours of death for tryptase assay. Care should be taken in interpreting results as tryptase can be raised in other causes of death and discussion with a clinical immunologist is recommended [38,39].
What was the cause of anaphylaxis?
Most IgE-mediated reactions occur within 30 min of exposure to the allergen , so a careful history about exposure to possible causes of anaphylaxis should be obtained, especially if there have been several episodes of anaphylaxis or other allergic reactions. Assays for allergen-specific IgE antibodies have been reported to give false negative results when blood has been taken for testing soon after anaphylaxis, leading to the suggestion that such testing should be delayed by a few weeks . However, in another case, levels of specific IgE to suxamethonium did not drop following anaphylaxis . Skin prick and serum allergen-specific IgE tests to suspected allergens may provide very useful additional information, but no one test is perfect and false negative and positive results may occur. Challenge testing can provide further information, but again is not perfect. Attribution of cause requires consideration of all relevant clinical information and test results, a subject covered in more detail elsewhere within this series .
Anaphylaxis is caused most often by a reaction to an ingested or injected agent. The most common suspected causes of fatal anaphylaxis in the United Kingdom are antibiotics, anaesthetic agents, other drugs, nuts, seafoods, other foods, insect stings and contrast media . Inhaled allergens rarely cause anaphylaxis, although can result in severe asthma.
Anaphylaxis occurring at or around the time of general anaesthesia has a number of unique features. First, its identification is more difficult as the patient is unable to report any symptoms of anaphylaxis – often the first sign ofanaphylaxis is pulselessness or difficulty in inflating the lungs . Subsequent identification of the causative agent(s) is complicated by the reaction having occurred after concomitant exposure to a number of possible precipitants, including anaesthetic agents, latex and possibly colloids. Specific IgE testing is of limited use in the investigation of anaesthetic reactions as many of the reactions are non-IgE-mediated . Many of the drugs that the patient will have received can cause hypotension through their pharmacological properties and this needs to be distinguished from an anaphylactic reaction. Therefore perioperative anaphylaxis is investigated most appropriately in joint clinics carried out by anaesthetists and immunologists/allergists .
Anaphylaxis may occur after exercise in some individuals, and in a significant proportion of exercise-induced anaphylaxis, allergy to food – especially to wheat – is an important co-factor. In such individuals avoidance of the implicated food for 6–12 h before exercise is advisable . Despite full investigation, in a significant proportion of cases anaphylaxis may occur for no discernible reason (termed ‘idiopathic anaphylaxis’).
Prevention of anaphylaxis – problems of quantifying risk
Predicting who is at risk of anaphylaxis is difficult. The commonly held perception by the public that the severity of allergic reactions worsens with each exposure is not true. Fatal anaphylaxis occurs typically in those who have previously had only mild reactions or, for stings and most drugs, no previous allergic reactions at all. What factors lead to anaphylaxis in one patient and a milder reaction in another patient with similar clinical histories and test results is not well understood. There are some indicators of increased risk; for example, fatal reactions to foods occur more often in young people, whereas fatal reactions to insect stings occur more often in older people. Fatal drug reactions occur more often in the older population, but this is likely to be due to the fact that this population is prescribed more drugs. The presence of asthma, and in particular poorly controlled asthma, is a risk factor for fatal reactions [45,46]. The type of allergen, presence of co-existing atopic disease, low levels of platelet-activating factor acetylhydrolase and low levels of serum angiotensin-converting enzyme are also risk factors for severe reactions [2,3]. However, these factors alone do not allow the accurate identification of those individuals who will have anaphylaxis and those who will have only a milder reaction.
Prevention of anaphylaxis mandates the identification of any identifiable responsible allergens, and their successful avoidance by the individual. Different individuals have symptoms at different threshold levels of exposure, with peanut allergy being an example of an allergy where small amounts can be tolerated by some individuals, but trace amounts can be fatal in others who have similar skin prick and blood test results . For such allergens it is not possible to determine a safe level of exposure for an individual and therefore complete avoidance is recommended.
Adrenaline autoinjectors offer a partial safety net should allergen avoidance fail. However, they should not be considered as a means of allowing risky behaviour or deliberate allergen exposure. There are many reasons why an adrenaline autoinjector may not be used by a patient suffering an episode of anaphylaxis, including the patient not carrying the autoinjector on their person, not replacing autoinjectors when they have been used or are out of date, collapsing too quickly to use it or being unfamiliar with how to administer it. In one case, the deceased was found holding an unused adrenaline autoinjector. Even when adrenaline autoinjectors are administered correctly and promptly, and other appropriate treatment is given, anaphylaxis can still be fatal . Largely because of the difficulties in predicting who is at risk of anaphylaxis, current practice in prescribing adrenaline autoinjectors varies [18,48–51]. Some of the factors to be taken into consideration when deciding whether to prescribe an adrenaline autoinjector are listed in Table 2 and Table 3 summarizes the features, investigation and management options of different causes of anaphylaxis.
Table 2. Factors that may influence whether an adrenaline autoinjector is prescribed.
• Is this allergen one which causes fatal reactions?
• Is this allergen one to which trace amounts can cause reactions?
• Likelihood of re-exposure
• Severity (but lack of severity of previous reactions does not exclude a future severe reaction)
• Previous reaction to trace amounts
• Asthma (is it well controlled, ‘brittle’, etc.?)
• Cardiac disease
• Other medications
• Distance from hospital (but even those close to hospital would benefit from early adrenaline from an adrenaline autoinjector)
• Ability to learn to use and administer autoinjector correctly
• Likelihood of unintentional exposure to allergen
• Likelihood of using autoinjector only for life-threatening episode
• Awareness that carrying autoinjector does not allow risky behaviour, e.g. risking allergen exposure
• Patient choice
Table 3. Clinical features, diagnostic tests and management of different causes of anaphylaxis.
Cause of anaphylaxis
Range of possible clinical features
Laboratory and other tests to consider
Oral itching and swelling Urticaria Vomiting and/or diarrhoea Stridor Wheeze Hypotension Collapse Cardiac arrest and death
Skin prick tests Specific IgEs Challenge testing
Avoidance of responsible allergen Patient education Epipen provision and training
Stinging insect venom allergy
Local itching, urticaria and/or swelling Distal itching, urticaria and/or swelling Wheeze Hypotension Collapse Cardiac arrest and death
Skin prick tests Specific IgEs Western blotting if required
Patient education Behaviour modification to reduce risk of further stings Epipen provision and training Desensitization immunotherapy
Urticarial or other rash Hypotension Collapse Cardiac arrest Death
Mast cell tryptase Skin prick tests (require careful interpretation) Specific IgEs (only available for limited number of drugs and require careful interpretation) Intradermal testing Challenge testing
Avoidance of drug Marking of medical notes/records Inform all health professionals involved in patient's care Patient education Alert bracelet or equivalent
Local rash and swelling Wheeze Hypotension Collapse Cardiac arrest Death
Mast cell tryptase Skin prick tests Specific IgEs Latex challenge
Avoidance of latex Marking of medical notes/records Inform all health professionals involved in patient's care Patient education Alert bracelet or equivalent
Urticaria Angioedema Wheeze Hypotension Collapse
Mast cell tryptase Skin prick tests Specific IgEs Exercise challenge
Exclusion of food allergy (e.g. wheat) Anti-histamines prior to exercise Doxepin H2 antagonists Montelukast
Urticaria Angioedema Wheeze Hypotension Collapse
Mast cell tryptase Skin prick tests Specific IgEs
Exclusion of above Anti-histamines Doxepin H2 antagonists Montelukast
Comorbidities such as asthma, cardiovascular disease and thyroid disease should be managed optimally and drugs that can worsen anaphylaxis (e.g. beta-blockers, ACE inhibitors, angiotensin II receptor antagonists) avoided where possible. However, one should not lose sight of the most important preventative measure: patient education and avoidance of triggers. The publication of guides to help to improve understanding of allergies and the interpretation of different allergy tests , and the work of organizations such as the Anaphylaxis Campaign (http://www.anaphylaxis.org.uk), are of great importance in this regard.
Anaphylaxis is a medical emergency requiring prompt intervention. Although protocols vary in some of the details of management, adrenaline is the mainstay of treatment and early administration saves lives. However, even with prompt appropriate treatment, anaphylaxis can still kill. The prevention of anaphylaxis requires the identification of at risk patients, their education and behaviour modification. Therefore, all patients who suffer an episode of anaphylaxis should be referred for review by an immunologist or allergist. Determining who is at risk of an anaphylactic reaction rather than a milder allergic reaction can prove difficult. Information from a number of sources, including the detailed history, is important. New tests which may have the potential for improved risk assessment include the testing for specific IgE antibodies to recombinant proteins (rather than whole allergen extracts), mapping of the epitope-specific binding of IgE antibodies using microarray technology and basophil stimulation assays using the patient's own IgE-sensitized basophils [36,53–55].
We would like to thank Richard Pumphrey for his help and input in reviewing the article.