• stroke;
  • meta-analysis;
  • observational studies;
  • anti-inflammatory agents;
  • nonsteroidal;
  • epidemiology



To perform a quantitative systematic review of observational studies on the risk of stroke associated with the use of individual NSAIDs.

Methods and results

Searches were conducted using the Medline database within PubMed (1990–2008). Observational cohort or case–control studies were eligible if reported on the risk of cardiovascular events associated with individual NSAIDs versus the nonuse of NSAIDs. We found 3193 articles, in which 75 were eligible for review and abstraction. Of the 75 articles, 6 reported relative risk (RR) of stroke. Data were abstracted into a database using a standardized entry form. Two authors assessed study quality, and discrepancies were resolved by consensus.

The pooled RR of all subtypes of incident stroke was increased with the current use of rofecoxib (RR = 1.64, 95% CI = 1.15–2.33) and diclofenac (RR = 1.27, 95% CI = 1.08–1.48). The pooled estimates for naproxen, ibuprofen, and celecoxib were close to unity. The risk of ischemic stroke was also increased with rofecoxib (RR = 1.82, 95% CI = 1.09–3.04) and diclofenac (RR = 1.20, 95% CI = 0.99–1.45). Data were inadequate to estimate the pooled RR by dose and duration, for other individual NSAIDs or nonischemic stroke subtypes.


This meta-analysis supports an increased risk of ischemic stroke with the current use of rofecoxib and diclofenac. Additional studies are required to evaluate most individual NSAIDS, the effect of dose and duration, and the subtypes of stroke. Copyright © 2011 John Wiley & Sons, Ltd.