BACKGROUND: The development of postdural puncture headache is related to a patient's age, size, pregnancy, sex, type of dural puncture needle and direction of the needle bevel. We studied the effect of the needle size and type of dural puncture on postdural puncture headache after spinal anesthesia in the 40 patients. METHODS: Forty patients, belonging to ASA classes 1 and 2, were divided into 2 groups, one using the lumbar paramedian approach (n = 20, Group 1) on L4-5 interspinous space with a 24 G spinal needle and the the other using Taylor's approach (n = 20, Group 2) through S2 foramen with a 22G spinal needle. All patients received an infusion of lactated Ringer's solution (1,000 ml). After spinal anesthesia, we measured the incidence, onset, duration and severity of postdural puncture headache in the 40 patients. RESULTS: The incidence of postdural puncture headache is lower in Group 2 (0%) than in Group 1 (15%). The onset of postdural puncture headache occured within 24 hours in 100% of the patients in Group 1 and the duration of postdural puncture headache occured within 48 hours in 66.7% of the patients in Group 1. The severity of postdural puncture headache was mild and moderate in 3 cases of all the 3 cases with postdural puncture headache of Group 1. The location of postdural puncture headache was frontal in 2 cases and occipital in 1 case. CONCLUSIONS: The incidence of postdural puncture headache is zero in Taylor's approach group in 20 patients. Therefore we think that spinal anesthesia using Taylor's approach is safer than spinal anesthesia with the lumar approach for postdural puncture headache.