This study was designed to determine the role of the afferent peripheral neural input in the regression of sensory analgesia during epidural analgesia. Eighteen patients who underwent minor obstetric or gynecologic surgery under lumbar epidural analgesia with lidocaine were divided into three groups. Group A received electrical single twitch stimulation at the spinal nerve dermatome one segment rostrad of the proximal extent of sensory paralysis. Group B received electrical twitch stimulation at the anterior crural region where nerve block may have been most profound in this study. Group C was assigned as a control and did not receive any stimulation. The patients received the stimulation for 60 min (i.e., from 80 to 140 min after the epidural injection). The range of sensory analgesia at 140 min was compared with that at 80 min. The regression of sensory analgesia in group B or C was 1 +/- 0 (mean +/- SD) spinal nerve segments, but in group A it was 5 +/- 1, which was significantly (P less than 0.0025) larger than the others. We conclude that the afferent peripheral neural input to a spinal segment where the nerve block is incomplete may accelerate the regression of sensory analgesia from epidural analgesia.