OBJECTIVE: The goal was to evaluate, in a clinical study, the predicted performance of the transverse tripolar system for spinal cord stimulation, particularly the steering of paresthesia, paresthesia coverage, and the therapeutic range of stimulation. METHODS: Six transverse tripolar electrodes were implanted in the lower thoracic region in four patients experiencing chronic neuropathic pain. Electrode positions, relative to the spinal cord, were estimated from computed tomographic scans. A dual-channel stimulator was used for initial percutaneous tests, and an implanted single-channel stimulator was used for follow-up test sessions. Nine "balance" settings and several cathode-anode combinations were used with the dual-channel and single-channel stimulator, respectively. In each test, the increase of paresthesia coverage from the perception threshold to the discomfort threshold was registered on a body map and the corresponding voltages were recorded. RESULTS: Paresthesia steering occurred in all but one patient. The normalized steering score, enabling quantitative comparisons of paresthesia steering among tests and patients, showed that maximum paresthesia steering occurred when the electrode was at least 3 mm dorsal to the spinal cord and centered <2 mm from its midline. Paresthesia coverage included 70 to 100% of the body up to the electrode level, unless the electrode migrated or had broken wires. The therapeutic range, defined as the discomfort/perception of paresthesia threshold ratio, varied from 1.6 to 4.0. CONCLUSION: The clinical performance of transverse tripolar stimulation is in accordance with the characteristics predicted by computer modeling. It enables finer control of paresthesia than that achieved by polarity changes in conventional spinal cord stimulation systems.