Terms and conditions: semantic complexity and aspirin resistance.

The term aspirin resistance has been used to describe the occurrence of cardiovascular events despite regular aspirin intake at recommended doses.1 In this context, such treatment failures resemble those with any drugs, including statins, -blockers, and ACE inhibitors.2,3 In secondary prevention, the clinical effectiveness of aspirin has been clearly demonstrated and is comparable to that of these other agents in that they all reduce nonfatal cardiovascular disease events by 25% to 30% and fatal events by 15% to 20% in randomized trials.3 These results represent a small to moderate but clinically worthwhile reduction in risk. Conversely, they suggest that 70% to 75% of nonfatal and 80% to 85% of fatal events are not prevented by these drugs. Mechanistic approaches to investigating aspirin resistance have relied heavily on ex vivo evaluations of platelet function. Although thrombosis is the proximate cause of virtually all occlusive vascular events,4 other factors, such as vascular function,5 and perhaps interactions with other blood cells, such as monocytes,6 are also probably relevant. It is unknown precisely how the impact of aspirin on the ex vivo response to selected concentrations of single aggregating agonists might model its efficacy in preventing clinical events in vivo. Multiple factors may confound platelet aggregometry, including posture, time of day, smoking, exercise, and blood cholesterol.7–9 Indeed, platelet aggregability may recover despite sustained inhibition of thromboxane during chronic dosing with aspirin.10 The term aspirin resistance is insufficiently precise to offer a credible basis for clinical decision making. More usefully, the multiple potential causes of treatment failure on aspirin might be pursued and named accordingly.11 Aspirin irreversibly acetylates a serine residue at position 530 on the cyclooxygenase (COX) enzyme, thus inhibiting the first step in the transformation of arachidonic acid to the platelet agonist thromboxane A2. The irreversible nature of COX inhibition underlies the ability of low doses of aspirin administered chronically to inhibit platelet aggregation in vivo.13 There is a nonlinear relationship of inhibition of platelet thromboxane A2 generation with inhibition of thromboxane-mediated platelet aggregation, requiring in excess of 95% inhibition to influence function.14 Although inhibition of platelet COX-1 at low aspirin doses is sufficient to explain the benefits on cardiovascular disease consistently observed in randomized trials, direct comparisons of clinical efficacy with higher aspirin doses, in which other mechanisms may be operative,15 have not been performed. Aspirin resistance has relied primarily on quantitative interpretations of the impact of aspirin on platelet aggregation ex vivo,16 occasionally on serum thromboxane B2, and in one instance on urinary 11-dehydrothromboxane B2 excretion.18 Although all of these approaches have been useful in exploring the clinical pharmacology of aspirin, none of them have been related quantitatively to clinical outcomes in individuals. Thus, the inference that a quantitative response in one of these variables to aspirin administration might predict the efficacy of aspirin in preventing a heart attack or stroke in that individual is presently unsubstantiated. Furthermore, chance, bias, and/or confounding are plausible alternative explanations for the findings from all of these studies. These limitations include technical reproducibility, adequate blinding of the investigators, assurance of compliance, and controls for recognized modifiers of the aggregation response mentioned above. Thus, it would seem premature and unwarranted to suggest that measurement of aggregation or thromboxane generation after dosing with aspirin could be used to categorize patients as “resistant” or “responsive” to aspirin in a way that reliably predicted clinical outcome and guided therapeutic decision making. Aside from aspects of trial design and technical limitations, it is quite possible that distinct molecular mechanisms may indeed contribute to treatment failure with aspirin. These include genetic variability in the target cyclooxygenases19 or indeed in proteins relevant to aspirin disposition. Although cyclooxygenase polymorphisms have been described,20 they have yet to be related to clinical outcome. One plausible explanation for the occurrence of an event despite aspirin intake may relate to drug–drug interactions. For example, nonsteroidal antiinflammatory drugs may interact pharmacodynamically with aspirin, compromising its ability to sustain inhibition of platelet thromboxane formation.21