Aspirin Intolerant Asthma

Aspirin-intolerant asthma (AIA) is a phenotype experienced by 10–20% of persistent asthmatics, in whom acute bronchoconstriction is induced by ingestion of aspirin and other nonsteroidal anti-inflammatory drugs (NSAIDs). These drugs share the ability to inhibit synthesis of prostanoids by blockade of cyclooxygenase (COX). Acute reactions to NSAIDs can be life threatening and may be associated with rhinoconjunctival and dermal symptoms. Drugs that selectively inhibit COX-2 appear to be better tolerated than nonselective inhibitors of COX-1 and COX-2.

Patients with aspirin-intolerant asthma usually have persistent underlying asthma, often associated with nasal polyposis. Pathologically, the bronchial and nasal airways of AIA subjects show chronic eosinophilia, with evidence of activation of eosinophils and mast cells during acute reactions. The etiology of AIA is unclear, but the proposed mechanism focuses on the inhibition by NSAIDs of the synthesis of a prostanoid, putatively prostaglandin E2, that would normally suppress local inflammatory reactions. The consequent synthesis of cysteinyl-leukotrienes and other leukocyte-derived mediators contributes to bronchoconstriction and other acute features.

Treatment of AIA involves avoidance of NSAIDs combined with conventional management of underlying asthma, with 75% of AIA patients requiring corticosteroids. Controlled desensitization with regular doses of an NSAID can provide protection against acute reactions.


The classical aspirin-intolerant asthma (AIA) syndrome was described by Samter and Beers in 1968 as a triad of rhinosinusitis (often with nasal polyps), asthma, and aspirin sensitivity. Patients with fullblown AIA are about twice as likely to be female as male, but no more likely to be atopic than the general population. Patients typically present in early middle age with nasal congestion, anosmia, and rhinorrhea, and many develop nasal polyps, often with secondary infection of the paranasal sinuses. Bronchoconstriction and airway inflammation usually emerge later, and this becomes severe, perennial asthma, with about 75% of patients needing oral or inhaled corticosteroids to maintain control of their symptoms. Acute respiratory reactions to nonsteroidal anti-inflammatory drugs (NSAIDs) may be accompanied by rhinoconjunctival symptoms and by dermal symptoms such as facial flushing and exacerbation of preexisting urticaria.

Aspirin challenges reveal that the tendency to bronchoconstrict in response to therapeutic doses of NSAIDs is more prevalent than the full-blown ‘aspirin triad’ recognized clinically. Based on patient history alone, the prevalence of NSAID intolerance in adult asthmatics is 3–5%, but it rises to 19% when consecutive asthmatic patients are challenged with oral aspirin. Prevalence in asthmatic children is less than 2% based on history alone, but 13–16% of postpubertal asthmatic children respond adversely when aspirin challenged. NSAID intolerance is overrepresented in the severe asthmatic population. Among patients who have experienced a near-fatal asthma exacerbation requiring treatment in the intensive care unit, 24% are aspirin-sensitive based on history alone. Most life-threatening acute exacerbations in AIA patients are directly due to inadvertent use of NSAIDs, and in a few cases the precipitating NSAID was prescribed by the patient’s own physician. However, over 40% of life-threatening asthma exacerbations in AIA patients cannot be attributed to NSAID ingestion, illustrating the severity of the underlying chronic asthma in these patients.


Aspirin (acetylsalicylic acid) and other members of the family of NSAIDs inhibit formation of the prostanoid family of lipid mediators by blocking isozymes of cyclooxygenase (COX). They are used therapeutically for their mild analgesic, antipyretic, and anti-inflammatory actions. NSAIDs are variably associated with adverse reactions including prolonged bleeding, nephritis, and gastric ulceration. In 1919, Cooke recognized that in some patients with asthma, aspirin may also precipitate life-threatening acute exacerbations. Such reactions have now been recognized in response to a wide variety of NSAIDs of diverse chemical classes (Table 1).

A family history of AIA is reported by only 6% of AIA patients, suggesting that any genetic influences are subtle and expressed phenotypically only in the presence of relevant environmental factors. Leukotriene C4 synthase is the terminal enzyme for the synthesis of the bronchoconstrictor cysteinylleukotrienes postulated to contribute to airway narrowing in AIA. A biallelic polymorphism in the promoter region of the LTC4 synthase gene has been reported, involving an A to C transversion at a position 444 bp upstream of the transcription start site. The variant 444C allele may lead to enhanced transcription of the LTC4 synthase gene in relevant cells. The prevalence of the variant allele has been reported to be significantly elevated in AIA patients from a Polish population, but this has not been replicated in US Caucasian or Japanese populations.


A classification system for allergic and pseudoallergic reactions to NSAIDs was proposed by Stevenson, Sanchez-Borges, and Szczeklik in 2001. AIA is classified as a type 1 pseudoallergic reaction in which asthmatic patients with a high frequency of sinusitis and nasal polyps experience lower respiratory tract reactions and/or rhinoconjunctival symptoms after exposure to therapeutic doses of aspirin and other NSAIDs. Reactions are dose-dependent and may be life threatening. The type 2 category describes urticarial reactions or angioedema induced by NSAIDs in patients with pre-existing chronic urticaria, and types 3 and 4 describe urticarial or sporadic reactions in patients who are otherwise normal. In contrast to types 1–4 above, in which patients show extensive cross-reactivity to many NSAIDs, patients classified in types 5–8 react adversely only to a single drug. These reactions include urticaria/angioedema, anaphylaxis, aseptic meningitis, and hypersensitivity pneumonitis.

A consistent, although not diagnostic, pathological finding in AIA patients is chronic eosinophilia in the blood, nasal polyps, and bronchoalveolar lavage (BAL) fluid. Immunohistochemical studies in bronchial biopsies have confirmed a marked bronchial mucosal eosinophilia in AIA patients, with eosinophil counts three- to fourfold higher than in aspirintolerant asthmatics and 10- to 15-fold higher than in normal subjects. After NSAID challenge, further increases in eosinophils, their activation marker ECP (eosinophil cationic protein), and histamine have been described in the nasal airways, BAL fluid, and plasma of AIA patients, suggesting activation of eosinophils and mast cells. In the nasal airway, aspirin challenge of AIA patients is associated with increments in nasal tryptase and histamine, strongly suggesting mast cell activation.

Tryptase and histamine levels also rise in the serum of patients experiencing systemic reactions to oral aspirin, but not in those with localized respiratory reactions. Release of tryptase is a recognized marker of mast cell activation in the pulmonary airway, and occurs within 5min of allergen challenge in allergic asthmatics. However, tryptase did not rise in the BAL fluid of AIA patients challenged with inhaled or endobronchial lysine-aspirin. In contrast, a rise in urinary levels of the PGD2 metabolite 9a, 11b-PGF2 following endobronchial lysine-aspirin challenge has been interpreted as evidence for mast cell activation in AIA. Together, the evidence suggests that inflammatory mediators released both from mast cells and from eosinophils contribute to the pulmonary and rhinoconjunctival symptoms following NSAID exposure in AIA patients.

Clinical Features

As there are no acceptable in vitro tests for AIA, confirmation of aspirin intolerance can only be obtained by NSAID challenge under controlled conditions. Lung function is monitored while patients ingest incremental oral doses of aspirin or inhaled doses of a lysine-aspirin conjugate or sulpyrine. Concomitant use of b2-adrenergic agonists, cromones, and inhaled corticosteroids may mask responses to NSAID challenge, leading to a high rate of false-negative results.

In a 3-day oral challenge protocol, incremental doses of aspirin are given at 3-hourly intervals up to a maximum of 650 mg. The challenge is terminated when forced expiratory volume in 1 s (FEV1) falls by at least 20%. Reactions begin around 50 min after oral aspirin ingestion, ranging from 20 to 120 min. A shorter protocol involves the inhalation of incremental aerosolized doses of lysine-aspirin, a soluble and nonirritant form (this is not available in the US). Respiratory reactions to inhaled lysineaspirin often occur within 1 min, so shorter dosing intervals of 30–60 min allow the entire challenge to be completed within 1 day.

Inhaled lysine-aspirin challenges are safer than oral challenges as reactions are localized to the airways and are easily reversible with inhaled b2-agonists. The sensitivity of inhaled lysine-aspirin challenge is similar to oral aspirin challenge.


The Cyclooxygenase Theory

The most successful model to explain acute respiratory reactions to NSAIDs is the cyclooxygenase theory promulgated by Szczeklik in 1975. This postulates that reactions are related directly to the pharmacological activity of NSAIDs in inhibiting isozymes of COX, the key enzymes in the synthesis of the prostanoid family of lipid-inflammatory mediators. Intolerance to an individual NSAID in vivo was shown to be predictable by its potency in inhibiting COX in vitro, with strong inhibitors (including aspirin, indomethacin, mefenamic acid, ibuprofen, and piroxicam) being common precipitants of adverse reactions, while weak inhibitors (such as sodium salicylate) precipitated reactions rarely or only at high doses. The concept that inhibition of prostanoid synthesis is the trigger for NSAID-induced reactions is recognized as a key advance.

The Role of Cysteinyl-Leukotrienes: The Shunting Hypothesis and the PGE2 Brake Hypothesis

The second key advance was the elucidation in 1979 of the structure of slow-reacting substance of anaphylaxis (SRS-A) as a mixture of potent bronchoconstrictor products of a related family of lipid mediators, the cysteinyl-leukotrienes (cysteinyl-LTs). The cysteinyl-LTs – LTC4, LTD4, and LTE4 – are now recognized to have important bronchoconstrictor and proinflammatory roles in many phenotypes of asthma. Their particular relevance to AIA emerged from the recognition that their biosynthetic pathway, the 5- lipoxygenase (5-LO) pathway, shares arachidonic acid as a common precursor with the prostanoid pathway (Figure 1). The leukotriene pathway is not inhibited by NSAIDs. Specific blockade of the prostanoid pathway by NSAIDs was therefore proposed to shunt arachidonate away from conversion into prostanoids, which have relatively little bronchoconstrictor activity, towards the formation of cysteinyl- LTs, which are highly potent bronchoconstrictors. This ‘shunting’ hypothesis is superficially attractive but measurements of lipid mediators in isolated leukocytes treated with NSAIDs argues against such a simple mechanism. It has therefore been superseded by an alternative notion, the ‘PGE2 brake’ hypothesis. This proposes that NSAIDs block the formation of an anti-inflammatory prostanoid, PGE2, which is otherwise known to suppress leukotriene synthesis by leukocytes. Exposure to NSAIDs may thus liberate the 5-LO pathway from suppression by endogenous PGE2 in vivo, at least in susceptible individuals.

There are three main lines of experimental evidence supporting this model:

  1. The triggering effect of NSAIDs on LT synthesis can be mimicked in vitro in a number of inflammatory leukocyte subtypes. Endogenous PGE2 suppresses, and NSAIDs consequently enhance, leukotriene synthesis in eosinophils, neutrophils, basophils, and macrophages, but apparently not in human lung mast cells. Eosinophils themselves generate sufficient PGE2 to suppress their own LTC4 synthesis by about 90%. Treatment of eosinophils with indomethacin inhibits endogenous PGE2 synthesis and stimulates LTC4 release. Replacement experiments showed that exogenous PGE2 restores the braking effect, returning LTC4 synthesis to the levels seen before indomethacin treatment. PGE2 may act in an autocrine or paracrine manner at EP2 receptors on the cell surface, followed by an increase in intracellular cAMP and activation of protein kinase A, but the subsequent steps by which 5-LO activity is inhibited are unknown.
  2. Eicosanoids can be measured in biological fluids, including bronchoalveolar lavage (BAL) fluid, and urine. Endoscopic challenge with lysine-aspirin reduces PGE2 and thromboxane A2 levels in the BAL fluid of AIA patients and aspirin-tolerant asthmatics, but a dramatic rise in BAL fluid cysteinyl- LTs is seen only in the AIA group. Following challenge with oral aspirin or inhaled lysine-aspirin, urinary LTE4 levels, used as a marker of whole-body cysteinyl-LT production, rise three to sevenfold in AIA patients, but not in aspirin-tolerant asthmatics; this response is not seen after methacholine-induced bronchoconstriction or placebo challenge. At the same time, there is a fall in urinary markers of prostanoid synthesis, such as 11-dehydro-thromboxane A2. In AIA patients, preinhalation of PGE2 before challenge with inhaled lysine-aspirin completely ablates the rise in urinary LTE4 and prevents the consequent bronchoconstriction, providing strong evidence that cysteinyl-LT synthesis has a functional role in the NSAID-induced airway narrowing. The protective effect of inhaled PGE2 does not correlate with its relatively weak bronchodilator activity, confirming that PGE2 preinhalation protects by restoring the suppression of cysteinyl-LT synthesis, not by dilating airways directly.
  3. The effector role of cysteinyl-LTs in adverse respiratory and rhinitic reactions to NSAIDs in most AIA patients has been confirmed with placebocontrolled clinical trials of specific leukotriene modifier drugs. The 5-LO inhibitors zileuton and ZD-2138 markedly blocked the rise in urinary LTE4 and the fall in FEV1 following oral aspirin challenge of AIA. Rhinoconjunctival and dermal reactions to oral aspirin are also blocked by zileuton. Antagonists of CysLT1 receptors block oral NSAID-induced respiratory reactions in AIA patients.

The PGE2 Brake is Derived from COX-1

The cyclooxygenase theory and the PGE2 brake hypothesis are thus crucial in understanding how NSAIDs trigger LT synthesis and bronchoconstriction, but they do not explain why only some asthmatics are susceptible to these responses. AIA patients tolerate selective COX-2 inhibitors including nimesulide, meloxicam, and rofecoxib, and respond adversely most often to NSAIDs with a greater selectivity for COX-1, such as aspirin and indomethacin. This suggests that cytoprotective/anti-inflammatory PGE2 in the lung is produced by constitutive COX-1. COX-1 is expressed in a large number of cell types, including mast cells, eosinophils, macrophages, vascular endothelial cells, bronchial epithelium, and bronchial smooth muscle. The exact cellular sources of the putative PGE2 brake remain unclear. AIA patients may overproduce cysteinyl-LTs chronically and acutely after NSAID exposure because of a defect in endogenous PGE2 synthesis. However, inhaled lysineaspirin equieffectively inhibits airway PGE2 synthesis both in AIA patients and in aspirin-tolerant asthmatics. Baseline levels of PGE2 and other prostanoids are not consistently different in the BAL fluid and urine of AIA patients and control groups. Immunohistochemical studies of AIA bronchial biopsies have found little evidence for a meaningful anomaly in the cellular expression of COX isozymes. One possibility is that a defective PGE2 brake may derive not from a failure of PGE2 synthesis, but of leukocytes to be suppressed by PGE2. An anomaly in EP2 receptor structure, expression, or signaling might be one explanation, and this would also be consistent with the lack of clinical benefit shown in AIA patients treated with the stable PGE1 analog misoprostol.

Suggestive evidence of an anomaly within the cysteinyl- LT biosynthetic pathway itself is that baseline cysteinyl-LT synthesis appears to be two- to sevenfold higher in AIA patients than in control groups, even in the absence of exposure to NSAIDs, as demonstrated by measurements of cysteinyl-LTs in BAL fluid, induced sputum, and urine. The increases seen after NSAID exposure are superimposed upon this chronically elevated baseline. Immunohistochemical analysis of bronchial biopsies from AIA and aspirin-tolerant asthmatic subjects shows that counts of cells immunostaining for LTC4 synthase, the terminal enzyme for cysteinyl-LT synthesis, were fivefold higher in AIA biopsies than in aspirintolerant asthma biopsies and 18-fold higher than in normal biopsies. The numbers of LTC4 synthasepositive cells in the bronchial mucosa correlated with elevated levels of cysteinyl-LTs in the BAL fluid and with bronchial responsiveness to inhaled lysine-aspirin. Persistent overproduction of cysteinyl-LTs in steady-state AIA may therefore be related to overexpression of LTC4 synthase in the bronchial wall, much of it within eosinophils and mast cells. This may also contribute to the further surge in cysteinyl- LT synthesis when PGE2 suppression is removed by NSAIDs.

Animal Models

Although the anti-inflammatory actions of NSAIDs and their effects of prolonged bleeding, nephritis, and gastric ulceration can be replicated in animal models, there has been relatively little interest in developing animal models of AIA. This may reflect, at least in part, the difficulty of replicating the diverse immunopathological features of asthma, especially chronic remodeling of structural airway tissues, and also of replicating the typically high degree of disease severity and glucocorticoid dependency observed in AIA patients.

Management and Current Therapy

National and international management guidelines for asthma (e.g., GINA guidelines) are based on disease severity and make no therapeutic distinctions between AIA and other asthma phenotypes. Asthmatics should avoid using NSAIDs either prescribed inadvertently or purchased over the counter. NSAID-induced acute reactions in susceptible asthmatics can be treated with nebulized b-2 adrenergic agonists, repeated frequently over several hours where necessary. Decongestants and antihistamines (topical or oral) can be used for associated rhinoconjunctival symptoms. In the most severe cases, intubation and mechanical ventilation in the intensive treatment unit may be required.

Treatment of chronic AIA focuses on anti-inflammatory therapy of the upper and lower airways using topical and inhaled/insufflated corticosteroids. Antibiotics may be required when purulent nasal secretions indicate infection, and many AIA patients require repeated polypectomies. Some AIA patients may require NSAIDs for concomitant inflammatory disease such as arthritis, and in such cases, aspirin desensitization may be an option. Acute respiratory reactions to NSAIDs in AIA patients are always followed by a refractory period lasting 2–5 days during which further reactions to NSAIDs cannot be induced. Sensitivity to NSAIDs re-emerges within a week, but regular low-dose NSAIDs can maintain the refractory state indefinitely.

Daily or alternate-day dosing of NSAIDs is therefore used clinically to desensitize AIA patients to inadvertent ingestion of NSAIDs and to allow NSAID therapy of concomitant diseases such as arthritis. Cross-desensitization occurs such that repeated dosing with one NSAID provides protection against adverse reactions to other NSAIDs. Chronic desensitization reduced the numbers of acute exacerbations and hospital admissions in AIA patients compared to a control group of AIA patients who avoided NSAIDs, associated with reductions in corticosteroid use. The mechanism of desensitization is unknown, but is tentatively linked to reductions both in the quantity of cysteinyl-LTs synthesized following NSAID exposure, and with reduced bronchial responsiveness to cysteinyl-LTs. In the nasal airway, desensitization has been linked to a reduction in expression of the CysLT1 receptor on infiltrating leukocytes.

In small studies in AIA patients, the antiallergic cromone sodium cromoglycate, the antiviral agent acyclovir, the antibiotic roxithromycin, and the longacting b2-agonist salmeterol have all been reported to block not only the bronchial responses to inhaled NSAIDs, but also the associated rise in urinary LTE4 excretion. The mechanism of these drugs is difficult to understand in this context, but may involve prevention of mast cell degranulation. On the basis of the pathophysiological evidence of a central role of cysteinyl-LTs in AIA, clinical trials of 5-LO inhibitors and CysLT1 receptor antagonists have been reported.

Zileuton improved lung function and rescue beta-2 agonist use and restored the sense of smell in a 6- week study of 40 AIA patients. In an 8-week crossover trial of montelukast in 80 AIA patients, there were significant improvements in lung function, use of rescue therapy, symptom scores, night-time awakenings, and asthma quality-of-life (QOL) scores compared with placebo. In the clinic, treatment response to leukotriene modifiers is variable, but a treatment trial is advised in patients with AIA uncontrolled by topical corticosteroids.

Further Reading

Drazen JM, Israel E, and O’Byrne PM (1999) Treatment of asthma with drugs modifying the leukotriene pathway. New England Journal of Medicine 340: 197–206.

Pavord ID and Tattersfield AE (1994) Bronchoprotective role for endogenous prostaglandin E2. Lancet 345: 436–438.

Samter M and Beers RF (1968) Intolerance to aspirin: clinical studies and consideration of its pathogenesis. Annals of Internal Medicine 68: 975–983.

Sanak M and Szczeklik A (2000) Genetics of aspirin-induced asthma. Thorax 55(supplement 2): S45–S47.

Stevenson D, Sanchez-Borges M, and Szczeklik A (2001) A classification of allergic and pseudoallergic reactions to drugs that inhibit cyclooxygenase enzymes. Annals of Allergy, Asthma and Immunology 87: 1–4. Stevenson DD, Simon RA, and Zuraw BL (2003) Sensitivity to aspirin and nonsteroidal antiinflammatory drugs. In: Adkinson NF, Bochner BS, Yunginger JW, et al. (eds.) Middleton’s Allergy: Principle and Practice, 6th edn., pp. 1695–1710. St Louis, MO: Mosby.

Szczeklik A, Gryglewski RJ, and Czerniawska-Mysik G (1975) Relationship of inhibition of prostaglandin biosynthesis of analogues to asthma attacks in aspirin sensitive patients. British Medical Journal 1: 66–69.

Szczeklik A and Stevenson DD (2003) Aspirin-induced asthma: advances in pathogenesis, diagnosis, and management. Journal of Allergy and Clinical Immunology 111: 913–921.