|Home | About | Journals | Submit | Contact Us | Français|
I disagree with Sudlow's recommendations.1 The cited ESPRIT study had lots of limitations. Firstly, during the study inclusion criteria changed from a three arm to a two arm design.2 3 Secondly, patients and physicians were not blinded to the treatment regimen. The resulting confounder therefore cannot be estimated. Thirdly, possible lifestyle changes, comorbidity, and co-treatment were not under examination.
Adherence in the dipyridamole-aspirin group was much less (2.6-fold) than in the aspirin group. The on-treatment analysis showed only a small benefit for bleeding complications (hazard ratio 0.58, 95% confidence interval 0.35 to 0.97). The analysis by intention to treat showed small benefits only in combined end points, which was driven by the single benefit in the occurrence of non-fatal strokes. The combined treatment showed no advantage in death. The new occurrence of disability was not reported. Furthermore, the real benefit from a number needed to treat of 104 is small.4
In my view the published data of ESPRIT do not alter the recommendations of the Antithrombotic Trialists' Collaboration.5
Competing interests: None declared.