Variability in response to aspirin: do we understand the clinical relevance?



    1. Division of Cardiology, Wilford Hall Medical Center and San Antonio Uniform Health Sciences Consortium, San Antonio, TX, USA; and *University of Kentucky, Lexington, KY, USA
    Search for more papers by this author

    1. Division of Cardiology, Wilford Hall Medical Center and San Antonio Uniform Health Sciences Consortium, San Antonio, TX, USA; and *University of Kentucky, Lexington, KY, USA
    Search for more papers by this author

Charles L. Campbell, Division of Cardiology, Wilford Hall Medical Center, 2001 Bergqust Drive, Suite 1, Lackland AFB, Texas, USA. Tel.: +1 210 292 6835; fax: +1 210 292 7737; e-mail:


Summary.  Aspirin, an irreversible inhibitor of platelet prostaglandin synthase activity, is the cornerstone of therapy for acute coronary syndromes. In recent years, laboratory and clinical data have accumulated that suggest there may be significant individual variability in the response to aspirin and that the effects of aspirin therapy vary significantly over time. There is, as of yet, no cohesive explanation for this variability. The term ‘aspirin resistance’ has been loosely applied to situations in which the clinical or ex vivo effects of aspirin are less than expected. In this review we discuss the clinical data regarding this phenomenon and the need for prospective evaluation of aspirin non-responders.


Platelet activation and aggregation are central to the pathogenesis of unstable angina (UA) and acute myocardial infarctions (MI), collectively referred to as acute coronary syndromes (ACS). Aspirin is an important therapy in ACS and is widely used in the prevention of ischemic events as well. The clinical benefits of aspirin are very probably related to its ability to inhibit platelet-derived prostaglandin G/H synthase, or cyclo-oxygenase 1 (COX-1), by covalently acetylating the enzyme preventing substrate binding. COX-1 is central to the formation of ecosanoids, the first enzyme in the pathway that converts arachidonic acid into prostaglandin G and H and later thromboxane A2 (TXA2). TXA2 is a vasoconstrictor that is known to cause platelet aggregation. Aspirin is rapidly absorbed from the stomach and has a half-life of 5–15 min in the circulation. The interaction between platelets and aspirin occurs quickly within the portal circulation and a single dose of aspirin at 325 mg will almost completely suppress thromboxane production within 15–30 min [1,2]. Given that platelets are anucleic, the effects of aspirin last for the approximately 10-day lifetime of the platelet. After a single dose of aspirin, COX activity is restored at the rate of normal platelet turnover, approximately 10% per day [3]. Some studies have suggested that normal hemostasis can return when approximately 20% of the platelets have uninhibited COX activity [4,5].

Additional COX enzymes have been characterized. The gene for the COX-2 enzyme is typically not expressed constitutively but is very sensitive to induction by inflammatory stimuli and growth factors. Aspirin is an approximately 170 times less potent inhibitor of the COX-2 enzyme [6]. Until recently COX-2 activity had not been noted in platelets. Weber et al. have, however, discovered small amounts of COX-2 in human platelets, perhaps originating in megakaryocytes where COX-2 activity functions as a regulator of hematopoiesis [7,8]. In situations with high platelet turnover this expression of COX-2 activity may be clinically relevant and there has been debate about the clinical importance of COX-2 inhibition in atherosclerosis [9]. An additional enzyme with COX activity (COX-3) has also been characterized. The clinical significance of this enzyme remains to be determined, but it appears to be sensitive to inhibition by acetaminophen and may be central to the antipyretic effects of this medication [10].

Aspirin is a cornerstone of therapy…or is it

Aspirin therapy is undoubtedly an important therapy in the setting of ACS [11,12]. Though the evidence is less compelling, aspirin is also recommended for secondary prevention of ischemic vascular events and for primary prevention in high-risk individuals. A recent meta-analysis of placebo-controlled aspirin trials, including more than 125 000 patients in a variety of clinical settings, found a 3% absolute reduction (RR  =22%) in ischemic events (cardiovascular death, MI, or stroke) associated with aspirin therapy [13]. In addition, in the community setting, retrospective analyses suggest that the widespread use of aspirin has been associated with a reduction in mortality among patients hospitalized with ACS [14,15]. In other analyses, it also appears that aspirin therapy is related to a therapeutic ‘frame-shift’ in which aspirin not only reduces the likelihood of experiencing an ischemic event, but also reduces the severity of breakthrough events [16,17].

Despite these results there remains some controversy with respect to chronic aspirin use. In the largest trial evaluating aspirin in the setting of secondary prevention, the Aspirin Myocardial Infarction Research Group (AMIS) trial where 4524 patients were randomized to receive 1 g of aspirin daily vs. placebo, there was a trend toward an increase in mortality (9.6% vs. 8.8%; P-value was not significant) [18]. In fact, in none of the large trials of aspirin was a significant reduction in mortality obtained [19]. In addition, the United States Food and Drug Administration recently refused to approve aspirin for use in the setting of primary prevention.

Also of interest are several retrospective analyses of large multicenter trials in which chronic aspirin therapy was found to be a marker for increased risk in the setting of ACS [17,20–22]. Most notably in the Thrombosis and Myocardial Infarction 11B trial (TIMI 11B), designed to evaluate the effectiveness of long-term treatment with low molecular weight heparin in the setting of UA and non-ST-elevation MI (UA/NSTEMI), patients who reported taking aspirin for more than a week prior to presentation were more likely to experience adverse events than those who were aspirin naïve. Data from this trial were used to develop a risk prediction model for patients presenting with UA/NSTEMI that was validated retrospectively by analyzing data from the Efficacy and Safety of Subcutaneous Enoxaparin in Unstable Angina and Non-Q-Wave Myocardial Infarction (ESSENCE) Trial and that has become widely accepted [23].

Analysis of some recent ACS trials also indicates that aspirin users have the most to gain from adjuvant antiplatelet and antithrombotic therapies. For example, in the Platelet IIb/IIIa in Unstable Angina: Receptor Suppression Using Integrilin therapy (PURSUIT) trial, treatment with eptifibatide resulted in a significant reduction in ischemic events only among prior aspirin users [20]. Similar results were found in analysis of TIMI 11B and of the Platelet Receptor Inhibition for Ischemic Syndrome Management in Patients Limited by Unstable Signs and Symptoms (PRISM-PLUS) trial in which the benefits of tirofiban and heparin were confined to those patients on chronic aspirin therapy at the time of entry into the trial [17,24].

The response to aspirin may not be uniform and may vary with time

Emerging as an area of research and further debate with respect to chronic aspirin therapy is the concept of aspirin resistance. Aspirin resistance has been variably defined; in some instances, including all patients who have experienced an ischemic event while taking aspirin (perhaps better referred to as treatment failures) and more often, as those patients in which aspirin therapy fails to achieve an arbitrarily defined reduction in the measured level of platelet activation or aggregation. The response to aspirin has been characterized in a variety of ways including measurement of urinary thromboxane B2 (TXB2) or P-selectin or by employing one of many in vitro platelet aggregation assays. It has been difficult to convincingly demonstrate the clinical relevance of the various in vitro tests of platelet function and no test has emerged as a clear gold standard for identifying patients resistant to aspirin. It is not surprising therefore that the reported incidence of an inadequate response to aspirin varies widely from 5 to 75%[25].

Ex vivo evidence of a variable response to aspirin

There are numerous trials showing variability in the baseline response to aspirin in different settings using a variety of ex vivo, laboratory based, assays of platelet function including platelet aggregability [26], classical platelet aggregometry [27–29], flow cytometric analyses of markers of platelet activation [30] and bleeding time (Table 1) [31]. Similar variability has also been noted in several studies using newly developed point-of-care assays [32–35]. Adding to these concerns are studies by Helgason et al. and more recently Pulcinelli et al. which demonstrate significant variability in the response to aspirin over time. Helgason et al. used platelet aggregometry to monitor patients with a history of ischemic stroke treated with between 325 and 1300 mg of aspirin. Of 303 participants, complete inhibition was achieved in only 151, and this effect was durable in only 104 (70%) on repeat analysis at 6 months [36]. In the more recent study, platelet aggregometry was used to periodically assess the effects of aspirin therapy in a group of patients with stable coronary disease. Aspirin-treated patients exhibited a gradual decline in platelet inhibition over the 24-month study period when compared to matched controls taking ticlopidine. By the end of the study period the effects of aspirin appeared to be negligible, while patients in the control group taking ticlopidine exhibited no decline in platelet inhibition [37].

Table 1.  Studies evaluating interindividual variations in response to aspirin
StudynTest% of patients considered partial or non-responders to aspirin
  • *

    Only the results from platelet-rich aggregometry from the study by Gum et al. were utilized to calculate the corrected mean.

Hurlen et al. [26]  93platelet aggregate ratio15%
Grotemeyer et al. [41] 180platelet aggregate ratio33%
Mueller et al. [42] 100whole blood aggregometry60%
Buchanan & Brister [31]  40bleeding time42%
Pappas et al. [28]  31whole blood adherencechange after aspirin normally distributed
Helgason et al. [36] 306platelet-rich plasma aggregometry26%
Valles et al. [52]  82platelet recruitment61%
Valettas et al. [30]  30flow cytometry57%
Gum et al. [33] 325platelet-rich platelet aggregometry5.5%
Andersen et al. [35]  71PFA-100®35%
Chen et al. [46] 151Ultegra rapid platelet function assay19.2%
Total1409 corrected mean 26%*

Perhaps nowhere is the controversial relationship between the clinical effectiveness of aspirin and aspirin resistance more evident than among patients undergoing coronary artery bypass grafting (CABG). There is ample evidence that the use of aspirin among these high-risk patients is effective in reducing the risk of early ischemic events [38]. In contrast with these clinical data, however, are two studies suggesting that these patients are almost entirely non-responsive to aspirin in the early postoperative period. Zimmerman et al. found that among 24 patients who had undergone CABG, 100 mg of aspirin was, on average, insufficient to inhibit ex vivo platelet aggregation or thromboxane synthesis in the first 10 days following surgery [39]. The same group later confirmed these results in a larger trial in which they found that among 93 patients undergoing CABG, in whom aspirin therapy had been terminated 7–10 days preoperatively, treatment with 100 mg of aspirin did not significantly inhibit ex vivo platelet COX activity. They also found that platelet aggregation was not altered by aspirin therapy in the first 10 postoperative days. Interestingly, the amount of COX-2 activity in platelets rose 16-fold during this period of time, perhaps secondary to increased platelet turnover. While there is conflicting evidence regarding the amount of platelet-derived COX-2 activity from other reports, the increase in platelet COX-2 production noted by Zimmerman et al. has prompted these and other authors to speculate that platelet COX-2-mediated production of TXA2 might be responsible for aspirin non-responsiveness in settings of high platelet turnover [8,40].

Clinical evidence of aspirin resistance

There is a growing body of literature that establishes a link between the variability in response to aspirin and adverse clinical events. Grotemeyer et al. in a study of 180 patients admitted with stroke, used an in vitro measurement of platelet reactivity performed 12 h after the administration of 500 mg of aspirin and found that 33% of the patients had platelets which were reactive despite aspirin therapy. Patients were discharged on 500 mg of aspirin three times daily. At 2 years of follow-up, patients who were initially non-responsive were more likely to suffer an adverse event including death, recurrent stroke, or MI (4.4% vs. 44%; P = 0.0001) [41]. Mueller et al. evaluated the response to aspirin using whole blood platelet aggregometry in 100 patients with intermittent claudication who underwent elective percutaneous balloon angioplasty and were treated with 100 mg day−1 aspirin. Again these authors found that there was considerable variation in the response to aspirin over the course of the year during which platelet aggregometry was performed four times. More importantly, patients found to be not fully responsive to aspirin therapy had an increase in adverse events as manifested by re-occlusion of vascular grafts [42].

Eikelboom et al. evaluated a subset of patients enrolled in the Heart Outcomes Prevention Evaluation (HOPE) trial for aspirin responsiveness by measuring levels of urinary 11-dehydro-TXB2. This is a stable urinary metabolite of TXA2 and a high urinary concentration among chronic aspirin users is thought to denote a lack of response to aspirin arising from either incomplete inhibition of platelet TXA2 production or from TXA2 production from sources other than platelets. In this case–control study, baseline 11-dehydro-TXB2 levels in 488 patients who suffered an adverse event during the HOPE trial (defined as MI, stroke, or death) was compared with that of age and gender-matched event-free controls. They found that when compared to controls, patients suffering an adverse event had significantly elevated urinary levels of 11-dehydro-TXB2 and that increasing quartiles of urinary thromboxane metabolite were associated with increasing risk [43].

A more recent report supports these earlier studies but also points out the pitfalls associated with ex vivo measurements of platelet function. Gum et al. evaluated 325 patients with stable coronary disease treated chronically with 325 mg of aspirin and found incomplete inhibition of platelet function in 5.5% of the patients using classical platelet aggregometry. An additional 23.8% were found to be semi-responders. These patients were also evaluated using a point of care platelet function analyzer (PFA-100, Dade Behring, Deerfield, IL, USA), which simulates primary hemostasis in whole blood samples. With this device, 9.5% of patients were found to be resistant to the effects of aspirin. There was no correlation between the methods, as only 1.2% of the patients were found to be resistant to aspirin by both methods [33,44]. Subsequently, the results of over 2 years of follow-up indicated that patients initially found to be resistant to aspirin by platelet aggregometry were at an increased risk of death, MI, or stroke [34]. However, there was no relationship between aspirin resistance as measured by the PFA-100 and subsequent adverse events (12.9% aspirin-sensitive, 15.1% aspirin-resistant, P = 0.4) [44]. This apparent disconnection may be explained by subsequent work which has demonstrated that platelet aggregation in the PFA-100 cartridge is sensitive to serum von Willebrand factor levels which were not measured in the study by Gum et al. [45].

In perhaps the most convincing study to date employing a point-of-care assay, Chen et al. used the Ultegra Rapid Platelet Function Analyzer (Accumetrics Inc., San Diego, CA, USA) to determine aspirin responsiveness among 151 patients scheduled for non-urgent percutaneous coronary intervention. Patients who were resistant to aspirin were much more likely to suffer post-procedure elevations in serum creatine kinase myocardial band levels, than were patients responsive to aspirin (51.7% vs. 24.6%; P = 0.006). This study is particularly interesting in light of the fact that these events occurred in the setting of dual antiplatelet therapy as all the patients were pretreated with clopidogrel > 12 h prior to their procedure [46].


Numerous mechanisms have been proposed to explain why some patients do not respond to aspirin, and the subject has been intensively reviewed elsewhere [25,40,47–51]. It is likely that more potential explanations will be discovered and that more than one mechanism will be operable in a given patient. As of yet no large clinical trial has been designed specifically to correlate clinical events and laboratory findings with respect to aspirin response. Given the lack of a clear ex vivo‘gold standard’ assay for platelet activity, this trial would probably need to link clinical events with multiple assessments of platelet biochemistry and function. Until such a study is completed it will remain difficult to determine if the breakthrough events experienced by patients treated with aspirin represent something as sinister as ‘resistance’ to aspirin or are related to more mundane issues such as aspirin dose, medication interactions, or medical non-compliance.


The authors wish to thank Lieutenant Colonel John Wightman MD, First Lieutenant Joel Elliott, Senior Airman Kenneth James and the members of the 791st Expeditionary Aeronautical Evacuation Squadron, deployed in support of Operation Iraqi Freedom and Operation Enduring Freedom, for support and assistance in the preparation of this manuscript.