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


Review: Risk for cardiovascular events is increased with rofecoxib, diclofenac, and indomethacin but not celecoxib


ACP J Club. 2007 Jan-Feb;146:24. doi:10.7326/ACPJC-2007-146-1-024

Clinical Impact Ratings

GIM/FP/GP: 7 stars

Cardiology: 5 stars

Geriatrics: 6 stars

Rheumatology: 7 stars

Source Citation

McGettigan P, Henry D. Cardiovascular risk and inhibition of cyclooxygenase: a systematic review of the observational studies of selective and nonselective inhibitors of cyclooxygenase 2. JAMA. 2006;296:1633-44. [PubMed ID: 16968831]



What are the risks for cardiovascular (CV) events in patients using cyclooxygenase 2 (COX-2) inhibitors and nonsteroidal antiinflammatory drugs (NSAIDs)?


Data sources: MEDLINE and EMBASE/Excerpta Medica (1985 to January 2006), Cochrane Library, Google Scholar, epidemiologic research Web sites, abstracts of scientific meetings, and bibliographies of relevant studies.

Study selection and assessment: Case–control or cohort studies that reported CV risks associated with use of selective COX-2 inhibitors and conventional NSAIDs with nonuse or remote exposure as the reference exposure. 17 case–control (8 nested) and 6 cohort studies (n = 1 659 288) met the selection criteria. Methodological quality of the individual studies was assessed with the Newcastle-Ottawa Scale. All studies scored from 7 to 8 out of 9 points.

Outcomes: CV events (included fatal or nonfatal acute myocardial infarction, CV death, coronary heart disease death, sudden cardiac death, unstable angina pectoris, thromboembolic CV event, and ischemic stroke).

Main results

COX-2 inhibitors studied were celecoxib (11 studies), rofecoxib (11 studies), and meloxicam (3 studies); NSAIDs studied were naproxen (15 studies), diclofenac (9 studies), ibuprofen (13 studies), indomethacin (6 studies), and piroxicam (4 studies). Point estimates were pooled using a random-effects model. Results are in the Table. Among COX-2 inhibitors, an increase in CV events was seen with rofecoxib, particularly at doses > 25 mg/d. Celecoxib, at the doses used in these studies, did not increase risk for CV events. Among NSAIDs, case–control studies showed increased risk with diclofenac and indomethacin.


A dose-related increased risk for cardiovascular events exists with rofecoxib. An increased risk could not be excluded with doses of celecoxib > 200 mg/d. Among nonsteroidal antiinflammatory drugs, diclofenac and indomethacin had risks similar to those of rofecoxib.

Sources of funding: National Health and Medical Research Council of Australia and National Heart Foundation Australia.

For correspondence: Dr. D. Henry, Newcastle Mater Hospital, Waratah, New South Wales, Australia. E-mail

Table. Risk for cardiovascular events with cyclooxygenase 2 inhibitors and nonsteroidal antiinflammatory drugs*

Drug Case–control studies Cohort studies Case–control and cohort combined RR (CI)
Number of studies RR (95% CI) Number of studies RR (CI)
Celecoxib 8 1.01 (0.90 to 1.13)† 3 1.22 (0.69 to 2.16)† 1.06 (0.91 to 1.23)†
Rofecoxib 9 1.31 (1.18 to 1.46) 2 1.53 (0.68 to 3.44)† 1.35 (1.15 to 1.59)
Rofecoxib > 25 mg/d 4 1.89 (1.43 to 2.51) 2 2.46 (1.29 to 4.71) 2.19 (1.64 to 2.91)
Meloxicam 3 1.25 (1.00 to 1.55) 0 - 1.25 (1.00 to 1.55)
Naproxen 12 0.96 (0.84 to 1.10)† 3 0.94 (0.85 to 1.04)† 0.97 (0.87 to 1.07)†
Diclofenac 7 1.36 (1.21 to 1.54) 2 1.36 (0.51 to 3.65)† 1.40 (1.16 to 1.70)
Ibuprofen 11 1.06 (0.95 to 1.18)† 5 1.12 (0.90 to 1.38)† 1.07 (0.97 to 1.18)†
Indomethacin 6 1.30 (1.07 to 1.60) 0 - 1.30 (1.07 to 1.60)
Piroxicam 4 1.06 (0.70 to 1.59)† 0 - 1.06 (0.70 to 1.59)†

*RR = relative risk; CI defined in Glossary. A random-effects model was used.
†Not significant.


Ever since the sudden withdrawal of rofecoxib in 2004, the CV safety of all COX-2 selective NSAIDs has been called into question, and even older nonselective NSAIDs are being drawn into the fray. We urgently need to know which, if any, of these drugs are safe to use.

The review by McGettigan and Henry seems to provide some reasonably clear answers, implying that rofecoxib is definitely out (at any dose) and that low-dose celecoxib is possibly safe, as are naproxen and ibuprofen. However, notwithstanding the use of a quality yardstick for observational studies, these conclusions are based entirely on observational studies that are intrinsically weaker than randomized controlled trials. Observational studies are relied on heavily in drug safety research, but this is mainly because sufficient numbers of large randomized trials are often not available. This is not particularly true for COX-2 inhibitors.

Kearney and colleagues (1) did a meta-analysis of COX-2 inhibitor trials looking specifically at CV risk. They concluded that there is an increased CV risk for all COX-2 inhibitors and, indeed, for many of the older NSAIDs, too. McGettigan and Henry point out that a dose-response relation was noted for celecoxib, with lower doses seeming to be safer in the Kearney and colleagues' review. The emerging picture, as Sánchez-Delgano highlights (2), is consistent with the risks and benefits of NSAIDs being more related to dose and potency than to COX selectivity. The debate about whether CV risk is or is not related to COX-2 selectivity may be academic. All NSAIDs are problematic, with CV risk probably increasing with potency, dose, and duration of treatment, with the possible exception of naproxen (1).

Colin P. Bradley, MD, MICGP
University College Cork
Cork, Ireland


1. 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]

2. Sánchez-Delgado EJ. Life without COX 2 inhibitors: risks and benefits are determined by dose and potency [Letter]. BMJ. 2006;332:1451-2. [PubMed ID: 16777893]