To Ridker and colleagues, who commented on the GUSTO trial in the 15 May issue [1] and in an earlier issue of Annals [2], we reemphasize that the trial did not have the power to show significant treatment differences within all subgroups of interest, even prespecified subgroups. However, there was remarkable consistency of the treatment effect across different subgroups [3]. We agree with Farkouh and colleagues [4] that only qualitative differences should "excite the reader to apply the overall trial conclusions differently to subsets of patients." We also agree that early treatment is important.

Regarding the concern that the benefit of accelerated tissue plasminogen activator (tPA) was due to a differential number of tPA-treated patients receiving coronary bypass surgery, we emphasize that contrary to Ridker's claim, the difference in rates of bypass surgery (9.0% for tPA, 8.3% for streptokinase) was not "larger than any reported mortality difference" [1]. Furthermore, surgery rates were not different at all in the first day after enrollment (0.51% for tPA, 0.56% for streptokinase), by which time half the mortality benefit of tPA was already achieved [5]. With careful analysis, the reduction in mortality with tPA is not explained by the difference in surgery rates.

The analysis plan for comparing streptokinase arms and combining them if there was no mortality difference was based on sound rationale and was documented in the protocol before the study began. It was used neither to make differences more significant nor to hide data. Moreover, end point data cited by Ridker and colleagues [1], including death or nonfatal stroke, were fully presented for all four treatment arms in the initial GUSTO report [5].

We agree with Farkouh and colleagues on the importance of a biological mechanism to explain treatment differences. This was precisely the motivation for the large angiographic substudy, a feature of GUSTO that is unparalleled in previous megatrials. Treatment differences in early and complete reperfusion (the important biological mechanism) fully parallel and explain the observed differences in mortality. This point has unfortunately been overlooked by GUSTO critics.

In the conduct of the study and the ensuing debate, we have endeavored to disengage science from advocacy. We believe that we have unequivocally shown that the survival benefit of myocardial reperfusion therapy is critically linked to early and complete restoration of blood flow.