Evidence has shown that electrical stimulation at the dorsal columns attenuated the "wind-up" phenomenon in dorsal horn neurons in nerve-injured rats. This study was aimed to test the effect of spinal cord stimulation (SCS) on temporal summation (TS), the clinical correlate of the wind-up phenomenon in patients with radicular leg pain. Eighteen patients with SCS implants were tested both 30 minutes after SCS activation ("ON") and 2 hours after turning it off ("OFF"), in a random order. Temporal summation was evaluated in the most painful site in the affected leg and in the corresponding area in the contralateral leg by applying a tonic painful heat stimulus (46.5°C; 120 seconds) and simultaneous recording of the perceived heat pain intensity. Patients were also requested to report their clinical pain intensity (0-100 numerical pain scale) during SCS "ON" and "OFF". The Wilcoxon signed rank test was used in the comparisons between SCS "ON" and "OFF". Spinal cord stimulation activation significantly attenuated clinical pain intensity (from 66 ± 18 to 27 ± 31, P < 0.001). In the nonpainful leg, SCS activation failed to produce an effect on TS (24 ± 20 vs 21 ± 24 in SCS "OFF" and "ON", respectively; P = 0.277). In contrast, a significant decrease in the magnitude of TS in the affected leg was observed in response to SCS activation (from 32 ± 33 to 19 ± 24; P = 0.017). These results suggest that attenuation of TS, which likely represents suppression of hyperexcitability in spinal cord neurons, is a possible mechanism underlying SCS analgesia in patients with neuropathic pain.