Application of futility analysis to refine jitter recordings in myasthenia gravis

Abstract

Introduction The current practice of single fiber electromyography (SFEMG) requires that 20 fiber pairs with normal jitter be collected to exclude myasthenia gravis (MG). We applied principles of futility analysis from clinical trials in an attempt to reduce that requirement. Methods We utilized conditional power futility analysis to assess the probability of an abnormal 20-pair SFEMG based on ongoing analysis of jitter as each pair is collected. Rules for early test termination in the presence of 0, 1 or 2 abnormal pairs were identified. These rules were then applied to previously collected SFEMG data. Results SFEMG could be stopped at just 12 pairs if all are normal and at 17 pairs if 1 is abnormal. The rules successfully determined when SFEMG could be stopped in 104/106 (98%) studies originally reported to be normal. Discussion If the first 12 SFEMG pairs have normal jitter, the study can be terminated and interpreted as normal.