Study Objectives: To determine the approximate incidence of transient neurologic symptoms (TNS) [formerly known as transient radicular irritation (TRI)] associated with procaine spinal anesthesia, and whether fentanyl prolongs the duration of procaine spinal anesthesia. Design: Unrandomized Pilot study. Setting: Community teaching hospital. Patients: 106 consecutive patients scheduled for spinal anesthesia for procedures anticipated to last less than 90 minutes. Interventions: All patients received 5% procaine for spinal anesthesia. Fentanyl 20 mu g was added for procedures anticipated to last longer than 45 minutes (but less than 90 min). Intraoperatively the adequacy of duration, Im,el, and intensity of anesthesia were observed. Time from injection of local anesthetic until knee-bending was recorded. Three days postoperatively, patients were questioned intensively in an effort to determine whether back pain and/or symptoms consistent with TNS had occurred.. Measurements and Main Results: Duration of anesthesia was adequate in ail but one instance. The intensity and the sensory level of anesthesia were satisfactory with one exception, a woman who had an unexpectedly low sensory level (LI) after 60 mg of procaine for cerclage, and who was also was the only patient to develop TNS. The incidence of TNS (0.9%) was markedly less than that reported after lidocaine and similar to the incidence observed after bupivacaine, Mild back pain without radiation occurred in II patients (10%) an incidence that is similar to that seen after bupivacaine and lidocaine, Compared with procaine alone, the addition of fentanyl significantly (p = 0.0001) prolonged the fine to bending knees from 72 minutes to 97 minutes. Conclusions: Procaine may be a useful alternative to lidocaine for short procedures, and it is less likely to produce TNS. Fentanyl prolongs motor block when added to procaine. (C) 1998 by Elsevier Science Inc.