The effect of added phenylephrine on the duration of sensory analgesia during lidocaine spinal anesthesia was determined in 65 ASA class I-III patients randomly divided into three groups. Group 1 (n = 25) received 62.5 mg lidocaine in 7.5% glucose; group 2 (n = 21) received lidocaine with 2 mg phenylephrine; and group 3 (n = 19) received lidocaine with 5 mg phenylephrine. The level of analgesia to pin prick was assessed by an anesthesiologist unaware of the drug combination used. The mean +/- SD cephalad level of analgesia did not differ among the groups. In group 1, the times for two- and for four-segment regression of the level of analgesia, and the time for regression of analgesia to the T-12 dermatome, were 77 +/- 19 (1 SD), 99 +/- 24, and 109 +/- 26 min, respectively. The corresponding values were 98 +/- 25, 118 +/- 27, and 130 +/- 36 min in group 2 and 124 +/- 32, 142 +/- 31, and 162 +/- 35 min in group 3. All the regression times in group 2 were significantly longer than those in group 1 (P less than 0.05). All the regression times in group 3 were significantly longer than those in group 2 (P less than 0.02). It is concluded that clinically useful prolongation of sensory analgesia may be obtained by addition of phenylephrine to lidocaine during spinal anesthesia.