A metaregression analysis of the dose-response effect of aspirin on stroke

Arch Intern Med. 1999 Jun 14;159(11):1248-53. doi: 10.1001/archinte.159.11.1248.


Background: We evaluated whether the risk of stroke depends on aspirin dose in patients with a previous transient ischemic attack or stroke.

Methods: We conducted a metaregression analysis of stroke by using published randomized, placebo-controlled trials. We analyzed studies of patients who had recently had a transient ischemic attack or stroke (ie, secondary prevention). We abstracted data on the treatment regimen and stroke. To evaluate the dose-response relationship, we conducted a metaregression analysis of study-specific risk ratios by means of weighted linear regression.

Results: Eleven randomized, placebo-controlled trials contributed a total of 5228 patients randomized to aspirin only and 4401 patients randomized to placebo only. The slope of the dose-response curve was virtually flat across a wide range of aspirin doses from 50 to 1500 mg/d (P = .49 for test of slope not =0). Summarizing across studies, aspirin decreases the risk of stroke by about 15% (risk ratio, 0.85;95% confidence interval, 0.77-0.94).

Conclusions: Aspirin reduces the risk of stroke by approximately 15%, and this effect is uniform across aspirin doses from 50 to 1500 mg/d. The lowest effective aspirin dose has not yet been identified, but it could be lower than 50 mg/d.

Publication types

  • Meta-Analysis
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Aged
  • Aspirin / administration & dosage*
  • Cerebrovascular Disorders / prevention & control*
  • Dose-Response Relationship, Drug
  • Female
  • Fibrinolytic Agents / administration & dosage*
  • Humans
  • Linear Models
  • Male
  • Middle Aged
  • Odds Ratio
  • Platelet Aggregation Inhibitors / administration & dosage*
  • Randomized Controlled Trials as Topic
  • Risk


  • Fibrinolytic Agents
  • Platelet Aggregation Inhibitors
  • Aspirin