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.