In spite of several clinical trials, the risk of inducing intracerebral hemorrhage by using aspirin for prevention of cardiovascular disease remains a matter of considerable controversy. Thrift and colleagues used data from a large Australian study of stroke to address this issue.
They studied all cases of first stroke admitted to 13 major hospitals during a two-year period. Cases secondary to blood dyscrasias, anticoagulation therapy, tumors and vascular abnormalities were excluded. The 331 study patients were matched by age and sex with control subjects who lived in the same area. Among the extensive data gathered on participants was information concerning use of aspirin and nonsteroidal anti-inflammatory drugs (NSAIDs). This information was checked against physician records.
In the two weeks before the stroke or interview, 17 percent of cases and 18 percent of controls reported taking aspirin. Five percent of cases and six percent of controls reported taking regular low-dose aspirin. The use of moderate to high dosages of aspirin (more than 1,225 mg per week) was reported by 6 percent of case subjects but only 3 percent of control subjects. Multiple statistical analyses were conducted to identify subgroups of patients who might be at increased risk of intracranial hemorrhage associated with aspirin use. Although borderline statistical significance was found for elevated blood pressure, increasing age and increased serum cholesterol level, the authors conclude that there was no evidence of substantially increased risk in any patient subgroup. No increased risk of stroke was found in patients who reported use of NSAIDs.
The authors conclude that no meaningful increase in risk of intracerebral hemorrhage could be identified in users of aspirin or NSAIDs. Specifically, the dosages recommended for prevention of cardiovascular disease were not associated with stroke risk. Although heavy users of aspirin did show an increased risk of stroke in this study, the small numbers of patients involved make this finding tentative.