Current issues of ACP Journal Club are published in Annals of Internal Medicine


Review: Selective COX-2 inhibitors increase vascular events more than placebo and naproxen, but not more than other NSAIDs


ACP J Club. 2006 Nov-Dec;145:66. doi:10.7326/ACPJC-2006-145-3-066

Clinical Impact Ratings

GIM/FP/GP: 5 stars

Hematol/Thrombo: 4 stars

Rheumatology: 6 stars

Source Citation

Kearney PM, Baigent C, Godwin J, et al. Do selective cyclo-oxygenase-2 inhibitors and traditional non-steroidal anti-inflammatory drugs increase the risk of atherothrombosis? Meta-analysis of randomised trials. BMJ. 2006;332:1302-8. [PubMed ID: 16740558]



Do selective cyclooxygenase-2 (COX-2) inhibitors increase risk for serious vascular events more than placebo or traditional nonsteroidal antiinflammatory drugs (NSAIDs)?


Data sources: MEDLINE and EMBASE/Excerpta Medica (1966 to April 2005), U.S. Food and Drug Administration Web site, and drug manufacturers.

Study selection and assessment: Randomized controlled trials (RCTs) ≥ 4 weeks in duration that compared a selective COX-2 inhibitor with placebo or a traditional NSAID. 138 RCTs (n= 145 373) met the selection criteria. Investigators and manufacturers provided details on the number of vascular events and person-time at risk.

Outcomes: Myocardial infarction (MI), stroke, and vascular death, and a composite endpoint of all vascular events.

Main results

Selective COX-2 inhibitors increased risk for all vascular events and MI, but not stroke or vascular death, more than placebo and naproxen (Table). COX-2 inhibitors and nonnaproxen NSAIDs did not differ for all vascular events, MI, or vascular death; risk for stroke was lower with COX-2 inhibitors (Table). When data from indirect and direct comparisons were used, the rate ratios for all vascular events of traditional NSAIDs compared with placebo were 0.9 (95% CI 0.7 to 1.3) for naproxen, 1.5 (CI 1.0 to 2.4) for ibuprofen, and 1.6 (CI 1.1 to 2.4) for diclofenac.


Selective cyclooxygenase-2 inhibitors increase risk for serious vascular events more than placebo and naproxen, but not more than nonnaproxen traditional nonsteroidal antiinflammatory drugs.

Sources of funding: UK Medical Research Council; British Heart Foundation; Cancer Research UK.

For correspondence: Dr. C. Baigent, University of Oxford, Oxford, England, UK. E-mail

Table. Risk for vascular events with selective cyclooxygenase-2 (COX-2) inhibitors vs placebo, naproxen, or other traditional nonsteroidal antiinflammatory drugs (NSAIDs) at 4 to 208 weeks*

Comparisons Number of trials (person-y) Outcomes Weighted event rates per person-y RRI (95% CI) NNH (CI)
COX-2 vs placebo 121 (31 129) Myocardial infarction 0.6% vs 0.3% 86% (33 to 159) 351 (190 to 913)
All vascular events 1.3% vs 0.9% 42% (13 to 78) 269 (145 to 869)
Stroke 0.4% vs 0.4% 2% (−29 to 47) Not significant
Vascular death 0.4% vs 0.2% 49% (−3 to 129) Not significant
COX-2 vs naproxen 42 (27 338) Myocardial infarction 0.6% vs 0.3% 104% (41 to 196) 353 (187 to 894)
All vascular events 1.2% vs 0.7% 57% (21 to 103) 238 (132 to 646)
Stroke 0.4% vs 0.4% 10% (−27 to 65) Not significant
Vascular death 0.3% vs 0.2% 47% (−10 to 140) Not significant
COX-2 vs other NSAIDs 51 (29 247) Myocardial infarction 0.5% vs 0.4% 20% (−15 to 68) Not significant
All vascular events 0.9% vs 1.1% 12% (−12 to 31) Not significant
Stroke 0.2% vs 0.4% 38% (5 to 59) 663 (427 to 5038)
Vascular death 0.2% vs 0.3% 33% (−6 to 57) Not significant

*Abbreviations defined in Glossary; weighted event rates, RRI, RRR, NNH, NNT, and CI calculated from control rates and rate ratios in article using the Peto 1-step approximation method.


The review by Kearney and colleagues is the most extensive and methodologically sound systematic review of the cardiovascular effects of selective COX-2 inhibitors yet published. Not surprisingly, it found that COX-2 inhibitors increase risk for “all vascular events” (regardless of whether patients were taking aspirin at the same time), mainly because of an increase in the incidence of MI, but it did not find evidence of any difference among the individual COX-2 inhibitors. An even more important contribution of this review is that it shows that traditional NSAIDs, with the exception of aspirin and possibly naproxen, cause an increase in vascular events that seems similar to that of COX-2 inhibitors.

What are the clinical implications of these findings? They reinforce that when both traditional and selective NSAIDs are prescribed, 3 factors must be considered—the likelihood that the patient will benefit symptomatically compared with more conservative measures, the patient's chance of a serious gastrointestinal complication from NSAID treatment (not evaluated in this review), and the patient's risk factors for cardiovascular disease. All patients should be fully informed about these factors and encouraged to take the drug for the shortest possible period.

A number of clinically important questions about the cardiovascular effects of selective COX-2 inhibitors could not be answered by the review by Kearney and colleagues because of the (thankfully!) low prevalence of vascular events. These questions include the risk for vascular events with low versus high drug doses and how soon after the drugs are started that cardiovascular risk increases. Studies of intravenous administration of selective COX-2 inhibitors in patients after coronary artery bypass surgery have shown increased risk for vascular events within days (1), suggesting that the risk with oral medications may also increase rapidly. An individual-patient meta-analysis planned by the same authors, which may address some of these issues, is anticipated with interest.

Andreas Laupacis, MD
Institute for Clinical Evaluative Sciences
Toronto, Ontario, Canada


1. Nussmeier NA, Whelton AA, Brown MT, et al. Complications of the COX-2 inhibitors parecoxib and valdecoxib after cardiac surgery. N Engl J Med. 2005;352:1081-91. [PubMed ID: 15713945]