Evidence-based medicine and reteplase: Inductive arguments over deductive reasoning

Two trials of thrombolysis for acute myocardial infarction are contrasted a nd compared using a systematic approach developed for the interpretation of studies where equivalence is claimed. Reteplase has been compared with str eptokinase in a true equivalence trial (International Joint Efficacy Compar ison of Thrombolytics [INJECT] trial) and with tissue plasminogen activator in a failed superiority trial (Global Use of Strategies to Open Occluded C oronary Arteries [GUSTO] III). Important methodological distinctions and ec onomic insights are highlighted. Using this approach one can conclude chat the INJECT investigators have shown that reteplase is at lease as effective as streptokinase and that the test for equivalence was satisfied. The poin t estimate of the treatment effect for the primary outcome event favours re teplase over streptokinase, and the confidence intervals essentially elimin ate the possibility of saeptokinase superiority. The GUSTO III primary outc ome event, 30 day mortality, provides no convincing evidence that reteplase is equivalent to tissue plasminogen activator. Selective emphasis on one o r two post hoc outcome event clusters, such as death and nonfatal stroke, o r death and disabling stroke, provides some suggestive, but not conclusive, evidence for equivalence. Moreover, for some of the outcome events, and in particular the primary outcome event, the point estimate suggests that ret eplase is, in fact, slightly less effective than tissue plasminogen activat or.