We retrospectively evaluated the effect of exercise on the degree of decrement in ulnar, spinal accessory, and facial repetitive nerve stimulation (RNS) in 179 patients with myasthenia gravis (MG) to assess whether exercise increases the diagnostic yield of identifying significant decrements. The mean worsening of decrement following exercise was 1.9% (ulnar nerve), 1.9% (spinal accessory nerve), and 1.3% (facial nerve). Abnormal (> or =10%) decrement solely following exercise occurred in the ulnar nerve in 7% of patients, accessory nerve in 5%, and facial nerve in 7%. When analyzed according to Myasthenia Gravis Foundation of America class of disease, the likelihood of producing > or =10% decrement only after exercise was greatest in class I MG with facial RNS and in class II and III generalized MG with ulnar and spinal accessory RNS. In all other disease stages, the likelihood of producing > or =10% decrement only after exercise was < or =7%. This study suggests that exercise increases the yield of diagnosis of MG by RNS in only a small percent of patients. Therefore, in most patients with suspected MG, RNS at rest is sufficient and the additional time required for prolonged postexercise RNS may be better spent in examining other muscle-nerve combinations.