The North American study of plasmapheresis in Guillain-Barré syndrome (GBS) included early, standardized electrodiagnostic testing in 210 of the 245 patients. To determine the types of abnormalities and the relation to outcome, we analyzed the prospectively collected motor conduction data obtained during the first 30 days of illness. For each parameter of motor conduction data, means were computed based on institutional normal data. Mean distal motor and F wave latencies commonly were abnormal in these early studies. In univariate analysis, mean compound muscle action potential (CMAP) amplitude from distal and proximal stimulation and mean motor conduction velocity were found to be related to the four predetermined outcome measures. In multivariate analysis of the motor conduction parameters, mean distal CMAP amplitude was the single best predictor of prognosis; other measures did not add to its predictive value. In further multivariate analysis of all factors relating to outcome, mean distal CMAP amplitude was determined to be the most powerful predictor of outcome, followed by plasmapheresis. These two factors were always statistically significant over and above all other variables. Our results indicate that prognostic information can be obtained from motor conduction studies even when performed early in the course of GBS; a mean distal CMAP amplitude of 0 to 20% of the lower limit of normal is associated with a markedly increased probability of a poor outcome. Nevertheless, even in patients with a low mean distal CMAP amplitude, the predicted outcome is improved with plasmapheresis therapy.