Cardiovascular risk associated with celecoxib or etoricoxib: a meta-analysis of randomized controlled trials which adopted comparison with placebo or naproxen.

AIM The present meta-analysis attempted to assess whether an unfavourable cardiovascular risk profile could be identified in the case of two COX2 selective inhibitors (COXIBs), namely celecoxib and etoricoxib. Based on the data from the literature, our meta-analysis aimed to assess the probability of major cardiovascular events reported with the use of celecoxib or etoricoxib and compare this with the results seen in patients assigned to the placebo group. Furthermore, the risk of cardiovascular events found by using celecoxib or etoricoxib was also compared with that associated with the use of naproxen, a nonselective non-steroidal anti-inflammatory drug (NSAID) chosen as our reference drug. METHODS The studies had to be randomized controlled trials with at least 4-week duration. Studies were included if they compared celecoxib or etoricoxib against placebo or naproxen. Moreover, the selected studies had to have determined the risk, odds or incidence of myocardial infarction, stroke or cardiovascular death. For the comparisons versus placebo, the endpoints of interest were "serious vascular events", "non-fatal myocardial infarction", "non-fatal stroke" and "death from cardiovascular causes", whereas "myocardial infarction" and "stroke" were the endpoints of interest concerning the comparison versus naproxen. RESULTS From the evaluation of 41 studies comparing celecoxib with placebo, we found a significantly higher incidence of serious vascular events in the celecoxib group compared to controls treated with placebo (rate ratio 1.598, 95% CI: 1.048 to 2.438; P=0.029). Furthermore, in patients allocated to treatment with celecoxib, we found an incidence rate of non-fatal acute myocardial infarction that was three times higher compared with the placebo group (rate ratio 3.074, 95% CI: 1.375-6.873, P=0.006). In contrast, we did not find any significant difference with regard to the incidence of nonfatal stroke and that of death from cardiovascular causes by comparing celecoxib and placebo. In addition, by examining cardiovascular outcomes that emerged from the 17 trials which compared etoricoxib with placebo, it was not possible to demonstrate statistically significant differences in incidence for each of the explored endpoints. With regard to the comparison of each coxib with the non-selective COX2 inhibitor naproxen, we did not find any significant difference for either the odds of myocardial infarction or that of stroke. CONCLUSION On the basis of our meta-analysis, we can state that symptomatic benefits induced by the prolonged administration of celecoxib may be partially invalidated by a concomitant increase in vascular risk, particularly the increased risk of myocardial infarction found in celecoxib-treated patients, compared to controls taking placebo. In contrast, treatment with etoricoxib proved not to result in an increased risk of serious vascular events when compared with both the placebo and naproxen. Our meta-analysis also denotes that the alternative to COXIBs, represented by naproxen, does not show significant benefit in terms of reduced cardiovascular risk. Therefore, considering that the increase in incidence rate of cardiovascular events associated with treatment with celecoxib is small in absolute terms, it is reasonable to state that celecoxib is still a drug whose benefits outweigh the potential adverse effects on the cardiovascular system.