Drug Insight: Cyclo-Oxygenase-2 Inhibitors--A Critical Appraisal

Burkhard Hinz; Bertold Renner; Kay Brune


Nat Clin Pract Rheumatol. 2007;3(10):552-560. 

In This Article

Cardiovascular Side Effects of NSAIDs and COX2 Inhibitors

COX2 inhibitors have been associated with an increased incidence of cardiovascular adverse effects ( Table 2 ). In the VIGOR study the incidence of myocardial infarction was significantly increased in patients receiving rofecoxib (50 mg/d) in comparison with patients randomized to receive naproxen (1 g/d).[19] In the CLASS and TARGET trials, however, there was no statistically significant difference in the incidence of cardiovascular events between patients receiving celecoxib, lumiracoxib or NSAIDs.[21,33] Data from these trials have not been accepted as proof that COX2 inhibitors increase cardiovascular risk for several reasons: these trials were not sufficiently powered to detect cardiovascular events; different targeted diseases (rheumatoid arthritis in VIGOR, predominantly osteoarthritis in CLASS and TARGET) and entry criteria (VIGOR precluded aspirin consumption) mean that these studies cannot be directly compared; and finally, and from the current point of view most importantly, an at that time unexpected cardioprotective effect of naproxen could have influenced the outcomes of the VIGOR study. Continuous and regular administration of naproxen 500 mg twice daily has meanwhile been shown to affect platelet COX1 activity and subsequent platelet aggregation throughout the dosing interval in some, but not all, patients.[34]

Three well-conducted prospective randomized clinical trials demonstrated significant reduction in new adenoma formation associated with the use of COX2 inhibitors in patients with a previous history of colorectal carcinomas.[35,36,37] Two of these studies (adenomatous polyposis prevention study with rofecoxib [APPROVE] and adenoma prevention with celecoxib [APC]), however, showed that rofecoxib and celecoxib might increase the incidence of cardiac infarctions and other cardiovascular reactions.[35,36] This effect became fully evident after a prolonged period of treatment in otherwise healthy patients included in these trials. Moreover, in high-risk patients, short-term treatment with valdecoxib or parecoxib was associated with an increased number of severe thromboembolic events.[38] As a whole, six placebo-controlled trials have established that COX2 inhibitors confer a cardiovascular risk, which manifests as myocardial infarction, stroke, and congestive heart failure.

These observations had various pharmaco–political consequences: rofecoxib (Vioxx®) and valdecoxib (Bextra®; Pharmacia & Upjohn, North Peapack, NJ) were withdrawn from the market; and regulatory bodies were prompted to request changes in the labeling of both selective and nonselective COX inhibitors, including those available for over-the-counter use.[39] To minimize this risk, the respective substances should be taken at the lowest effective dose for the shortest possible duration of treatment.[40,41]

In contrast to COX2 inhibitors, no placebo-controlled randomized trial has been designed to define the cardiovascular risk of NSAIDs. Although the data from the prematurely and unconventionally terminated Alzheimer's Disease Anti-inflammatory Prevention (ADAPT) study of celecoxib, naproxen, and placebo in Alzheimer's disease is suggestive of an increased cardiovascular and cerebrovascular risk of naproxen,[42] these data are by no means definitive. The study was not designed to detect differences in cardiovascular and cerebrovascular risks, and the composite outcomes were specified post hoc. Moreover, these results are in contrast to a meta-analysis of 138 (published and unpublished) randomized trials[43] in which it was concluded that the incidence of serious vascular events is similar between a COX2 inhibitor and any non-naproxen NSAIDs, and that the risk associated with naproxen is in the placebo range. The summary rate ratio for vascular events, compared with placebo, was 0.92 for naproxen, 1.51 for ibuprofen, and 1.63 for diclofenac. Furthermore, population-based nested case–control studies have shown an increased risk of myocardial infarction associated with the current use of both COX2 inhibitors and NSAIDs.[44,45]

Finally, comparable rates of thrombotic cardiovascular events have been reported for the highly selective COX2 inhibitor etoricoxib and the NSAID diclofenac in the MEDAL study program,[46] which involved approximately 35,000 patients with osteoarthritis or rheumatoid arthritis. The participants were randomly assigned to either two daily doses of etoricoxib (60 or 90 mg), or a 150 mg/d dose of diclofenac, and then followed for an average of 18 months. The primary endpoint—the rate of any thrombotic cardiovascular event—was 1.24 events per 100 patient-years of use for those on etoricoxib and 1.30 events per 100 patient-years of use for diclofenac. Clearly, because of the absence of a nontreatment arm as a result of ethical constraints, this trial does not provide estimates of an absolute cardiovascular risk for patients taking diclofenac or etoricoxib as compared with those not receiving COX inhibitors; however, the absence of a significant difference in the relative cardiovascular risks suggests that there is presently no rationale for a further differentiation of COX-2 inhibitors and NSAIDs in terms of cardiovascular safety.

It has been proposed that COX2 inhibitors might elicit thrombotic reactions by inhibiting the COX2-dependent synthesis of vasoprotective prostacyclin in endothelial cells, while leaving the platelet COX1-derived formation of prothrombotic thromboxanes unaltered, thereby resulting in an imbalance between the eicosanoids.[47] Recent evidence, however, supports the view that disturbance of this balance cannot be the sole mechanism conferring the cardiovascular risk of COX2 inhibitors. Patients with coronary artery bypass grafts experienced severe cardiovascular events when administered valdecoxib or parecoxib, even though they were receiving apparent protection with low-dose aspirin and clopidogrel, both of which are well-established platelet aggregation inhibitors.[38] Likewise, a subgroup analysis from the APC trial showed that aspirin might not abrogate the potential cardiovascular harm of celecoxib.[36] In most long-term studies with COX2 inhibitors, an increased cardiovascular risk became evident after prolonged drug administration only. Finally, COX2-derived prostaglandins upregulate the expression of the thrombin inhibitor thrombomodulin in human smooth muscle cells, thus providing a molecular basis for a hitherto unknown platelet-independent mechanism underlying the prothrombotic effects of both NSAIDs and COX2 inhibitors.[48]

Claims that NSAIDs inhibit COX1, thereby conferring cardioprotection, proved wrong because platelet COX1 activity has to be suppressed by more than 95% to translate into inhibition of platelet aggregation.[49] Such a complete COX1 inhibition over the whole dosing interval is only achieved by low-dose aspirin and, in some individuals, by high-dose naproxen (500 mg twice daily).[34,49] Other NSAIDs such as ibuprofen and diclofenac suppress COX1 (>95%) only at peak plasma concentrations. This is consistent with an increased cardiovascular risk with high-dose diclofenac and ibuprofen, but not with naproxen.[43]

Consequently, it is plausible that a permanent blockade of COX2-dependent prostaglandins, including prostacyclin, is the (currently) most plausible explanation for the cardiovascular hazard conferred by selective and nonselective inhibitors. In fact, prostacyclin, which is suppressed by over 60% by both NSAIDs and COX2 inhibitors,[50] is not only a potent inhibitor of platelet aggregation, but also interferes with processes leading to hypertension, atherogenesis and cardiac dysfunction. In a randomized study, diclofenac, given in a delayed release form, elicits an even more pronounced inhibition of intravascular COX2-dependent prostaglandin formation than rofecoxib and celecoxib.[15] It has been suggested in this context that both degree and time-course of intravascular COX2 inhibition might determine the differential profile of cardiovascular side effects associated with NSAIDs and COX2 inhibitors.

In the cardiorenal hypothesis, changes in arterial blood pressure have been proposed to underlie the long-term cardiovascular side effects of both NSAIDs and COX2 inhibitors. The involvement of COX2 in human renal function is supported by numerous clinical studies that showed that COX2 inhibitors, similarly to NSAIDs, can cause peripheral edema, hypertension, and exacerbation of pre-existing hypertension by inhibiting water and salt excretion by the kidneys.[51,52,53] These observations are of major importance given that relatively small changes in blood pressure could have a significant impact on cardiovascular events. In patients with osteoarthritis, increases in systolic blood pressure of 1–5 mmHg have been associated with 7,100–35,700 additional ischemic heart disease and stroke events over 1 year.[54] Likewise, the hypertension optimal treatment (HOT) study showed that diastolic blood pressure differences of +4 mmHg can lead to a 28% increase in myocardial infarctions.[55]

The use of NSAIDs in patients at cardiovascular risk poses additional problems. According to published[56,57] and unpublished studies, ibuprofen and naproxen (the latter at higher than nonprescription doses) can interfere with the antiplatelet activity of low-dose aspirin when either is co-administered with aspirin. The underlying mechanism might be a competitive inhibition at the acetylation site of platelet COX1. Both ibuprofen (reversible inhibition) and aspirin (irreversible inhibition) occupy nearby sites on COX1, such that the presence of ibuprofen interferes with aspirin binding. Once ibuprofen leaves the binding site, COX1 will not be inhibited because aspirin will already have been metabolized or eliminated. The clinical implication of this interaction is unclear. It is, however, potentially important because the cardioprotective effect of aspirin, when used for secondary prevention of myocardial infarction, could be decreased or negated if NSAIDs are used too. In this context, a small epidemiological study of survivors of myocardial infarction suggested that concurrent ibuprofen but not diclofenac undermined the efficacy of aspirin in preventing a second myocardial infarction.[58] Current recommendations by the FDA,[59] therefore, advise patients who use immediate release aspirin and ibuprofen to take a single dose of 400 mg ibuprofen at least 30 minutes or longer after aspirin ingestion, or more than 8 hours before aspirin ingestion.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.