Artificial anal sphincter: complications and functional results of a large personal series.

PURPOSE: This study was undertaken to evaluate the technique of artificial sphincter for fecal incontinence, with its complications and risk factors, the functional results, and which variables derived from demographic data, preoperative studies, device characteristics, technical details, perioperative findings, and complications could influence the outcome.
METHODS: The Acticon Neosphincter was implanted in 53 patients (35 females), median age 46 years, with total anal incontinence not amenable to sphincter repair or after failed sphincteroplasty. In females with associated rectocele, this was synchronously corrected. Six (11 percent) patients already had a colostomy, but no proximal stoma was constructed at the time of implantation. Causes of incontinence were congenital, 13; iatrogenic, 13; obstetric, 10; neurogenic, 9; trauma, 4; idiopathic, 2; and perineal colostomy, 2. Physiologic testing before and after the operation and preoperative endosonography were done when they were available. Quality of life was assessed in 25 patients. Mean follow-up was 26.5 (range, 7-55) months.
RESULTS: Perioperative events occurred in 14 (26 percent) patients: abnormal bleeding, 7; vaginal perforation, 4; rectal perforation without apparent contamination, 2; and unobserved urethral perforation, 1. Early complications were mainly related to sepsis in 8 (15 percent) patients and wound complication in 8 (15 percent). Sepsis could not be statistically associated with any of the variables studied here. Wound separation was associated with fibrosis (p = 0.003) and tension of the wound (p = 0.001). Late complications were: cuff and/or pump erosion, 9 (18 percent) patients; infection, 3 (6 percent); impaction, 11 (22 percent); pain, 4 (8 percent); and mechanical failures, 2 (4 percent). None of those complications showed a statistical association with any of the variable studied here. There were 10 (19 percent) definitive explants caused by septic or skin complications. Only 26 (60 percent) of 43 patients with the device in action use the pump (patients' decision). Normal continence was achieved in 65 percent of patients and continence to solid stool in 98 percent. The Cleveland Clinic score of incontinence (0-20, maximal incontinence) changed from 17 +/- 3 preoperatively to 4 +/- 3 postoperatively (p = 0.000). An early complication of the perianal wound influenced the functional results: postimplant score > 4 < or = 4 (p = 0.009). Resting and squeeze pressures changed significantly after activation (p = 0.000). Quality of life measured in four subscales changed significantly in all the subscales (p = 0.000).
CONCLUSIONS: The artificial anal sphincter restores continence to solid stool in almost all severely incontinent patients, two-thirds of whom achieve practically normal continence. Quality of life improves significantly. Infection and skin erosion are the cause of the majority of explants. No predictable factors of functional success could be found in this study.