OBJECTIVE Many patients with fecal incontinence demonstrate a functional deficit of the internal anal sphincter or the external sphincter muscles without any apparent structural defects. Few patients are amenable to repair or substitution of the sphincter. However, sacral nerve stimulation appears to offer a valid treatment option for fecal incontinence. The objectives of this study were: to evaluate the efficacy of temporary stimulation of the sacral nerve roots (percutaneous nerve evaluation) in patients with functional fecal incontinence; to determine the mechanisms of possible improvement; and to evaluate if temporary stimulation could be reproduced and maintained by implanting a permanent neurostimulation system. METHODS Twenty-three patients with fecal incontinence, 18 females and 5 males, median age of 54.9 years (range 28-71), underwent a percutaneous nerve evaluation test. Eleven patients (47.8 percent) also had urinary disorders: urge incontinence (4), stress incontinence (3), and retention (4). Associated disorders included perineal and rectal pain (1), spastic paraparesis (1), and syringomyelia (1). All patients underwent a preliminary evaluation using stationary anal manovolumetry, pudendal nerve terminal motor latency measurements, and anal ultrasound. A percutaneous electrode for the stimulation of the sacral nerve roots was positioned at the level of the third sacral foramen (S3) in 20 patients and S2 in 2 patients (1 patient missing). Stimulation parameters used were: pulse width 210 microsec, frequency 25 Hz, and average amplitude of 2.8 V (range 1-6). The electrode was left in place for a minimum of 7 days. Five patients were successively implanted with a permanent sacral electrode with a stimulation frequency of 16 to 18 Hz and amplitude of 1.1- 4.9 V. RESULTS Seventeen of the 19 patients (89.4 percent) who completed the minimum percutaneous nerve evaluation period of 7 days (median 10.7 (range 7-30)), had a reduction of liquid or solid stool incontinence by more than 50 percent, and fourteen (73.6 percent) were completely continent for stool. The most important changes revealed by manovolumetry were an increase in resting pressure (P < 0.001) and voluntary contraction (P = 0.041), reduction of initial pressure for first sensation (P = 0.049) and urge to defecate (P = 0.002), and a reduction of the rectal volume for urge sensation (P = 0.006). The percutaneous nerve evaluation results were reproduced at a median follow-up of 19.2 months (range 5 to 37) in the 5 patients who received a permanent implant. CONCLUSIONS Temporary stimulation of the sacral roots (percutaneous nerve evaluation) can be of help in those patients with fecal incontinence, and the results are reproduced with permanent implantation. The positive effect on continence seems to be derived from not only the direct efferent stimulation on the pelvic floor and the striated sphincter muscle, but also from modulating afferent stimulation of the autonomous neural system, inhibition of the rectal detrusor, activation of the internal anal sphincter, and modulation of sacral reflexes that regulate rectal sensitivity and motility.