Restless Legs Syndrome (RLS) is a sensory-motor disorder which produces sleep disturbance. Using data from a large clinical trial of gabapentin enacarbil (GEn) we sought to assess the ability of baseline, and changes from baseline, in clinical trial endpoints to predict treatment response.Data were derived from a randomized, double-blind, placebo-controlled, crossover polysomnography study of gabapentin enacarbil 1200 mg (n = 121) or placebo (n = 123). Efficacy evaluations included: sleep measures from polysomnography, subjective sleep measures, Suggested Immobilization Test (SIT) measures, and International Restless Legs Severity Scale (IRLS) and Clinical Global Impression-Improvement (CGI-I). Correlations were evaluated using Spearman's rank correlation coefficients. Predictors of treatment response were separately assessed for GEn and placebo using categorical IRLS and CGI-I outcomes. Stepwise logistic regression models ascertained which combination of baseline and change from baseline variables predicted response.Moderate to large correlations were observed between changes in the IRLS and changes in subjective sleep for both GEn and placebo, substantially larger for GEn than placebo. Small to moderate correlations were present between the change in IRLS and the change in SIT-discomfort for both GEn and placebo. In the stepwise regression, for both GEn and placebo, baseline and change from baseline SIT discomfort, as well as change in sleep quality, were strong predictors of response.Changes in sleep quality, and baseline and changes in SIT discomfort were prominent predictors of treatment response for GEn and placebo. Predictors of treatment response may allow for more targeted enrollment in future clinical trials and may provide insights into the efficacy of RLS treatments.