SECONDARY PREVENTION OF STROKE - DOES DIPYRIDAMOLE ADD TO ASPIRIN

Background and Purpose : The purpose of this paper is to evaluate, in the light of all available evidence, the place of aspirin alone and of aspirin combined with dipyridamole in the secondary prevention of cer ebrovascular accidents. Methods : We performed a meta-analysis of all identified double blind, controlled, studies in secondary prevention o f cerebrovascular accidents for the following categories : studies com paring aspirin with placebo ; studies comparing aspirin plus dipyridam ole with placebo ; studies comparing aspirin plus dipyridamole with as pirin alone. An indirect comparison was carried out to compare the res ults obtained with aspirin alone and those obtained with aspirin combi ned with dipyridamole. Results : The meta-analysis of trials involving aspirin alone against placebo showed a risk reduction on strokes (17% reduction, p = 0.02), ''important vascular events'', i.e. a combinati on of vascular deaths, non-fatal strokes and non-fatal myocardial infa rction (18% reduction, p = 0.003). Fatal vascular events (vascular dea ths and fatal strokes) did not seem to be reduced at all. The overall mortality was reduced by 10%, but this reduction failed to reach stati stical significance (p = 0.23). The meta-analysis of trials involving aspirin combined with dipyridamole showed more important risk reductio ns on every outcome whether fatal or not. Strokes were reduced by 42% (p < 0.001), fatal strokes by 43% (p = 0.02) and vascular deaths by 24 % (p = 0.07, not significant). The overall mortality was reduced by 30 % (p = 0.004). Direct comparisons of aspirin with aspirin plus dipyrid amole did not indicate differences between the two treatment regimens. However the sample sizes involved in these comparisons were far too s mall to be informative. Indirect comparisons yielded statistically sig nificant results in favour of the combination in terms of ''important vascular events'' (p = 0.007), all strokes (p = 0.007) and fatal strok es (p = 0.03). The results were also in favour of the combination but not statistically significant in terms of all deaths (p = 0.10) and va scular deaths (p = 0.08). Conclusions : Aspirin used alone reduces sec ondary occurrence of vascular events in cerebrovascular patients. Ther e is no evidence, however, of a reduction of fatal events (vascular de aths and fatal strokes). In contrast, aspirin in combination with dipy ridamole reduces non-fatal as well as fatal events. These results as w ell as the indirect comparisons of the risk reductions suggest that th e combination of aspirin with dipyridamole may be superior to aspirin alone ; this hypothesis is presently tested in a large randomized tria l.