Long-Term Results of Artificial Anal Sphincter Implantation for Severe Anal Incontinence (2025)

Abstract

Treatment of severe anal incontinence by implantation of an artificial anal sphincter (AAS) was first reported by us in 1987, 1 and the first series were published in 1989 and 1992. 2,3 The first patients were implanted with an unmodified urinary sphincter, but subsequently this device was modified to meet the demands of a bowel sphincter (higher closing pressure and increased strength of the cuff).

Because production of the sphincter was suspended, only a few patients underwent implantation until 1996, when the device again became available to a few centers. Recently published results from these centers have confirmed our early positive evaluation of the method. 4–6 The follow-up varied from 4 to 76 months in the first two of these series 4,5 and from 4 to 12 months in the last. We herein present the first long-term results with a follow-up of ≥5 years with an AAS.

PATIENTS AND METHODS

Between 1987 and 1993, 17 patients (11 women and 6 men; median age 46 [range 32 to 65 years]) underwent implantation of the AAS. The first six patients received a urinary sphincter (AMS 800, American Medical Systems, Minneapolis, MN); the last 11 received a modified version of the urinary sphincter (the cuff-tab was strengthened, the cuffs were made wider, and the pressure-regulating balloon was enlarged so that a cuff pressure of up to 90 cm H₂O could be obtained). The indications for implantation are shown in Table 1. Previous surgery for anal incontinence included four patients with external sphincter repair from trauma (obstetric in three), one with a failed postanal repair, and one with a failed nonstimulated graciloplasty. All patients were incontinent for solid stool and had an incontinence score of 5 on a modified Williams scale. 7 This scale ranges from 1 to 5 (1: full continence to feces and flatus; 5, frequent episodes of incontinence to solid and liquid stool).

The technique has been described previously. 1–3 In short, the inflatable cuff is placed around the anal canal through two small incisions at 3 and 9 o’clock. The cuff is placed as high as possible, preferably at the anorectal junction. The pressure-regulating balloon is placed extraperitoneally to the left or right of the bladder, and the pump by which the patient can inflate and deflate the cuff is placed in the labia majora or in the scrotum.

Patients underwent a standard preoperative bowel preparation. Antibiotic prophylaxis in the first part of the study consisted of cefuroxime 750 mg, metronidazole 500 mg, and clindamycin 600 mg three times daily for 8 days, starting at the induction of anesthesia. In the last part of the study, prophylaxis was changed to gentamicin 240 mg once daily and metronidazole 500 mg three times daily for 5 days.

RESULTS

Of the 17 patients, 8 had a functioning AAS ≥5 years after the primary implantation (median 7 years [range 5 to 10 years]). Table 2 shows the reason for nonfunctioning sphincter at follow-up in nine patients. Two patients died within the first 3 years after the implantation from unrelated causes (cardiac disease, nephropathy). Explantation from infection was performed in two patients in the immediate postoperative period; one was receiving immunosuppressive treatment because of nephritis. In one patient, explantation was performed 7 months after implantation because of infection, which started as an erosion of the pump through the skin from forceful compression. Thus, the explantation rate from infection was 3/17 (18%; 95% CI 4% to 43%).

In two patients, the device was removed because of mechanical malfunction within the first year after the primary implantation. Further, in two patients the device was explanted followed by colostomy 2 and 3 years after the implantation: in one patient severe chronic diarrhea developed as a result of diabetic enteropathy, which rendered her incontinent, and in the other patient severe rectal emptying dysfunction developed, although no anatomic obstruction could be demonstrated.

Of the eight patients with a functioning sphincter after ≥5 years, two had a urinary sphincter and six a modified bowel sphincter. Four of the six patients who received implants after previous incontinence surgery failed were in the long-term follow-up group.

In five of the eight patients, revisional procedures were performed (63%; 95% CI 24 to 91) (Table 3). In one patient, refilling of fluid was performed percutaneously through an access port in the control pump. One further patient had incomplete rectal emptying, also without a demonstrable anatomic obstruction. This patient managed the condition by regular enemas and did not want the AAS removed.

Continence scores on a modified Williams scale 7 at follow-up are shown in Table 4. Also, the patient with occasional episodes of incontinence of solids found that the improvement justified the surgical procedure and did not want the device explanted. Resting anal pressure (deflated cuff) was 10 to 28 cm H₂O and squeeze pressure (cuff inflated) was 45 to 86 cm H₂O.

DISCUSSION

This study shows that long-term improvement of anal continence with an AAS can be expected in approximately 50% of patients with severe anal incontinence, depending on the etiology of the incontinence. This confirms our initial results. Recently published series of 5 to 13 patients report better results 4–6 ; this may be due in part to a different patient selection. In the beginning, we restricted the use of the device to patients with anal incontinence resulting from neurologic disease, because no other treatment options existed for these patients. Incontinence in this group may be particularly difficult to treat because sensory disturbances are common. Further, one patient with incontinence from diabetic neuropathy developed enteropathy with severe diarrhea, which the AAS could not retain. In the two recent studies mentioned, 4,5 only 3 of 12 and 2 of 13 patients, respectively, had incontinence caused by neurologic disorders.

In our series, six patients received an AAS after local sphincter repair failed. The result in this group of patients was clearly superior to that in patients with neurologic disorders. In the series of Lehur et al, 5 the AAS was used as primary treatment in some of the patients with traumatic anal incontinence, when local repair based on clinical and ultrasound examination was considered unlikely to succeed. This is undoubtedly a correct approach, because an attempt at local sphincter repair when clinical and instrumental examinations (intraanal ultrasound, magnetic resonance imaging) indicate little chance of success will make a second procedure, whether implantation of an AAS or a dynamic graciloplasty, more difficult and hence possibly less successful.

Implantation of foreign material in the anorectal region is likely to carry a higher risk of infection than implantation in other parts of the body. The explantation rate of the device from infection was in the present series 18%; this compares well with other published series, where the rate varied between 15% and 17%. 4–6 Preliminary results from an American multicenter trial showed explantation from infection in 3 of 13 patients (23%). 7 To reduce the infection rate, a meticulous perioperative regimen is mandatory; this includes not only the handling of the device and effective antibiotic prophylaxis but also strict attention to certain technical details. The cuff must be kept in place around the anal canal and must not slide down against the perineal skin, because this might result in erosion through the skin. This is accomplished by placing the cuff above the anococcygeal raphe posteriorly and above the corresponding raphe or muscle anteriorly. If the latter is not present, the tendinous part of the transversus perinei muscle may be substituted. In the patient with anal atresia, it was possible to place the cuff above a rudimentary puborectalis muscle.

Five of the eight patients followed for >5 years had revision procedures. This is a somewhat higher rate than in other published series. This difference may partly be due to shorter follow-up in these series, but also the fact that a higher proportion had implantation of the more recent version of the system, where rupture of the cuff and balloon is less likely to occur.

An alternative to an AAS, except in some patients with neurologic conditioned incontinence, is the dynamic graciloplasty. 8–10 The clinical results of this procedure are comparable to those of the AAS, 9,10 and the frequency of revision procedures is no less. 11 Rectal evacuation problems, which occurred in two of our patients without any sign of physical obstruction, also occur after dynamic graciloplasty. 10,11 The reason for this complication is obscure, but it may be due to preexisting evacuation problems that went unnoticed at the time of surgery because of longstanding incontinence. It is too early to determine whether these two surgical procedures can be used interchangeably. In our series, a patient with a failed graciloplasty was successfully treated with an AAS, the gracilis muscle being left in place, and there seems to be no reason why a failed AAS could not be treated with a dynamic graciloplasty. In patients with previous perianal infection or irradiation, the graciloplasty should probably be performed.

Sacral root stimulation, a new approach for the treatment of end-stage anal incontinence, probably requires an anatomically intact sphincter and is consequently an alternative to the AAS in only a limited group of patients. 12

In conclusion, the AAS seems to be a valid treatment for end-stage anal incontinence in selected patients. Infectious complications still present a problem, whereas mechanical problems are less frequent with the modification of the device now available. Long-term results in this series were adequate in approximately half of the patients. Further clinical trials are necessary to define more precisely the indications for this treatment compared with other modalities.

Footnotes

References