Sample size requirements for evaluating heart valves with constant risk events

A method is described for computing sample size requirements for a clinical study of a new heart valve, according to the guidance document recently revised by the FDA. The FDA requirements specify a one-sample hypothesis test in which the complication rates for the study valve are compared to fixed values determined from previous experience with approved devices. These values, which can differ for different complications and for mechanical or tissue valve types, are called OPC's (Objective Performance Criteria). The method described, using the Poisson distribution, provides a simple formula for computing the valve-years required corresponding to any OPC. At the OPC level required by the PDA guidelines (1.2% year), the minimum follow-up necessary is about 800 valve-years.