INTRODUCTION: Artificial anal sphincter has been proposed in severe anal incontinence when local treatment is unsuitable or has failed. The outcome after implantation of this newly developed device has not yet been determined. METHODS: Twenty-four consecutive patients (7 men; median age, 44; standard deviation, 18; range, 14–80 years) implanted since May 1996 for a minimum of six months at three centers were prospectively evaluated. Continence (scoring system, 0 (normal) to 120 (complete incontinence)), rectal emptying, and general satisfaction were assessed clinically and at anal manometry preoperatively and postoperatively at a six-month interval. The causes of incontinence included anal trauma (9 cases), neuropathy (6 cases), neurologic disorders (4 cases), congenital malformations (3 cases), and prolapse (2 cases). Median duration of incontinence was 7.5 (standard deviation, 8) years. Stomas pre-existed in two cases and was created at implantation in one. RESULTS: Median follow-up was 20 (standard deviation, 8; range, 10–35) months. Seven patients had their devices explanted, and reimplantation was successfully performed in three of these cases. At the end of follow-up, 20 (83 percent) patients had an implanted activated device. Fecal incontinence score dropped significantly from a median 106 (standard deviation, 13) preoperatively to 19 (standard deviation, 32), 25 (standard deviation, 29), and 25 (standard deviation, 25) at six months, one year, and the end of follow-up, respectively (P<0.0001). Minor and major emptying difficulties occurred in seven and two patients, respectively. A high degree of satisfaction was achieved at the end of follow-up in 18 (75 percent) of the total series. Median anal pressures at rest on manometry increased significantly from 28 (standard deviation, 17; range, 5–76) mmHg preoperatively to 60 (standard deviation, 17; range, 38–96) mmHg with a closed cuff at the end of follow-up. These pressures dropped to 30 (standard deviation, 16; range, 9–65) mmHg with an open cuff, and reocclusion time lasted a median of 4.6 minutes (standard deviation, 3 minutes; range, 38 seconds to 10 minutes). CONCLUSION: Artificial anal sphincter provided prolonged and reasonably good functional results in severe incontinence, reproducing an efficient sphincteric mechanism and allowing satisfactory anal occlusion and rectal emptying in approximately 75 percent of cases in this study. The definitive explantation rate was kept low by careful patient selection and appropriate surgical and perioperative management.