Purpose: To review the effectiveness of medical treatments for stroke
prevention in patients at elevated risk for stroke. Data Sources: Engl
ish-language articles published after 1977 and indexed in MEDLINE unde
r the following Medical Subject Heading terms: anticoagulants, aspirin
, dipyridamole, ticlopidine, or sulfinpyrazone, combined with cerebrov
ascular disorders. Study Selection: Randomized controlled trials of an
ticoagulant or platelet antiaggregant treatment reporting subsequent s
troke and myocardial infarction, death, or complications in persons wi
th asymptomatic carotid stenosis or bruit, transient ischemic attack (
TIA), previous stroke, nonvalvular atrial fibrillation, or other vascu
lar diseases. Data Extraction: Of 900 articles identified, 33 were sel
ected by two independent reviewers and abstracted for outcome events a
nd person-years of follow-up. Results: In patients with nonvalvular at
rial fibrillation, warfarin is highly effective in reducing stroke and
death but may result in more complications. Aspirin appears to be les
s effective and less risky than anticoagulation. In patients with TIA
or minor stroke, both aspirin and ticlopidine reduce the risk for stro
ke. In patients who have had myocardial infarction, warfarin is effect
ive but had high complication rates in the reviewed studies. Aspirin s
lightly reduces the risk for stroke. Conclusions: Warfarin is strongly
recommended for persons with nonvalvular atrial fibrillation who are
older than 60 years or who have additional risk factors for stroke. As
pirin is recommended for persons at elevated risk for bleeding while r
eceiving anticoagulants. For persons with TIA or minor stroke, aspirin
should be used first. Patients who do not respond to or tolerate aspi
rin or who have had a major stroke are reasonable candidates for ticlo
pidine. For patients who have had myocardial infarction, aspirin is re
commended for the prevention of secondary myocardial infarction but no
t of stroke.