The concept of equivalence and its application to the assessment of thrombolytic effects.

Abstract

Very large clinical trials have become the norm in the evaluation of thrombolytic agents, and these 'megatrials' are administratively complex and expensive. It remains to be seen whether new thrombolytics will lead to further large reductions in fatality from an acute myocardial infarction, but new agents may well have advantages in areas such as safety and ease of administration, in addition to other clinical benefits (i.e. fewer cases of cardiac shock, heart failure and atrial fibrillation). The problem is how to introduce such new agents without a megatrial for each one. Endpoints other than fatality have some advantages and, in thrombolysis, angiographic studies are a necessary step in the development of new agents. However, such studies may not always correlate precisely with the results of mortality endpoint studies. Measurements of the resolution of ST segment elevation in myocardial infarction seem to provide a very useful method of assessing thrombolysis, but although such a technique can be applied to large numbers of patients, it cannot totally replace mortality endpoint trials. The 'equivalence' of two treatments is a clinical, not a statistical, concept, although statistical principles that allow equivalence to be investigated with medium-sized trials should be applied. Demonstrating equivalence in outcome between the new thrombolytic reteplase and streptokinase was the aim of the INJECT study.