A lack of uniform methodology used by different authors in the assessment of different puncture techniques in spinal anesthesia formed the basis of the current study, which compared under randomized conditions the incidence of post spinal headache after a median or paramedian (lateral) approach. MATERIALS AND METHODS. Two hundred and fifty ASA physical status II and III patients, aged 50-85 years, scheduled for transurethral prostate surgery under spinal anesthesia were investigated. The patients were comparable with regard to weight and height (Table 1). No premedication was given and, 30 min prior to surgery, all patients received normal saline 400-500 ml i.v. The patients were randomly divided into two groups of 125 patients each to receive 4 ml 0.5% bupivacaine in 5% glucose (specific gravity 1.017 at 20 degrees C) using the median or paramedian (lateral) approach according to the following scheme (Table 2): I: 4 ml 0.5% bupivacaine/median approach; II: 4 ml 0.5% bupivacaine/paramedian approach. The study was carried out in a double-blind fashion. Neither the patient nor the investigator evaluating the post spinal headache was aware of which technique had been used. Lumbar puncture was performed by a midline approach at the L3-4 interspace using a 25-gauge (Whitacre) spinal needle with the patient in the sitting position group I. The bevel of the spinal needle was directly laterally, so that the dural fibers that run longitudinally were spread rather than transected. When using the paramedian approach (group II), patients were placed in the flexed lateral decubitus position and the spinal needle inserted 1 cm medial and 1 cm lateral and caudad to the lowest part of the posterior superior iliac spine and then directed medially and cephalad at an angle of 55 degrees into the subarachnoid space. Postoperatively, patients were allowed to move as soon as possible; no prophylactic bed rest was ordered. Starting from the 1st postoperative day, patients were evaluated by an independent observer and asked whether they were suffering from any problems concerning anesthesia. Typical post-puncture headache was defined as invariably bifrontal and occipital, frequently involving the neck and upper shoulders, and being aggravated by the upright position. Statistical analysis of the data was performed using the Mann-Whitney rank-sum test for unpaired samples. A P value of less than 0.05 was considered statistically significant. RESULTS. Twenty-six of 250 patients (10.4%) developed post spinal headaches. Comparing both groups, 11/125 (8.8%) patients in the median group (group I) versus 15/125 (12%) in the paramedian group (group II) had typical post-puncture headaches. Within the group of patients aged 50-60 years, the paramedian approach (group II) showed a significantly higher headache rate compared with group I (P less than 0.05). Neurologic sequelae were not observed; 6 patients received epidural injections of autologous blood while the rest of the patients suffering from post spinal headache were treated conservatively with bed rest, analgesics, and fluids. CONCLUSIONS. The results indicate that the incidence of post spinal headache is higher in younger patients when using the paramedian (lateral) approach. However, our findings suggest that the choice of lumbar puncture technique--median or paramedian--is of little importance in regard to post-puncture headache in elderly patients. The paramedian approach is especially useful when degenerative changes are encountered in the interspinous structures in elderly patients, when an ideal position is difficult to achieve.