February 2005

Section 5 Gastrointestinal Tract and Abdomen

Clinical Evaluation

History

Figure 1. Workup and management of constipation

Constipation is the most common digestive complaint, with as much as 20% of the population reporting this symptom.¹ The meaning of the term constipation, however, is variable: when patients describe themselves as constipated, they may be referring to decreased stool frequency, reduced stool volume, altered stool consistency, or difficulty with defecation.² Accordingly, when a patient presents with a complaint of constipation, a thorough history of the presenting illness is essential [see Figure 1].

The patient should be asked about the frequency of bowel movements, the volume of stool per movement, the caliber of the stool, and, in particular, any changes in bowel habits over time. Patients with idiopathic constipation tend to have long-standing problems, with no abrupt change in bowel habits. Thus, if the history reveals constipation of sudden onset, an underlying cause (e.g., cancer) is more likely and should be sought. Other important symptoms that should lead to a search for a secondary cause are weight loss, anorexia, nausea and vomiting, rectal bleeding, changes in stool caliber, and fever. The patient should always be asked about previous colon cancer screening or other GI investigations. Although chronic constipation is common, severe constipation that has been present since early childhood should alert the clinician to the possibility of undiagnosed short-segment Hirschsprung disease; this rare diagnosis is easily missed if it is not given appropriate consideration. Other symptoms may be indicative of an outlet problem (e.g., rectocele or nonrelaxing puborectalis syndrome); such symptoms include requiring a prolonged period to evacuate stool from the rectum, a feeling of incomplete rectal emptying, and the need to support the perineum (through digitation of the vagina or rectum) to achieve complete evacuation.

Diet can contribute significantly to constipation. Because high-fiber foods tend to increase stool bulk and frequency, detailed information on dietary fiber intake should be obtained. Because dehydration increases fluid resorption from stool and thereby results in the formation of hard stools, total daily fluid intake should be determined as well. A specific effort should be made to assess intake of fluids that contain caffeine, which exerts a diuretic effect. Most patients with long-standing constipation will already have tried some form of self-medication. Such attempts should be documented, both to help assess the severity of the symptom and to determine the likelihood of response to simple measures.

Various other diseases and certain common medications [see Table 1] also can cause or contribute to constipation. When such factors are present, treating the underlying condition or changing medications can result in substantial improvement. Therefore, a thorough past medical history and an accurate medication history are essential. A family history of colonic neoplasia or inflammatory bowel disease is potentially suggestive and may lead to a more intensive search for secondary causes. Victims of physical or sexual abuse may present with constipation; however, they are unlikely to mention the abuse if not directly questioned about the possibility.

Physical Examination

During physical examination, it is important to make a quick assessment of the patient's nutritional status. In general, patients with idiopathic constipation should not appear malnourished; the appearance of malnutrition should prompt a more extensive search for a secondary cause. An abdominal examination should be conducted to look for any significant abdominal distention, tenderness, or masses. Distention is a common and expected finding with idiopathic constipation, but significant tenderness or masses should prompt a full investigation.

All patients presenting with constipation should undergo a rectal examination. The anus should be examined for evidence of scarring or stricture. A digital rectal examination should be done to assess anal tone; high anal tone and inability to increase pressure when asked to squeeze are common findings in patients with obstructed defecation resulting from a nonrelaxing puborectalis. An effort should be made to look for any anterior defect in the rectovaginal septum, which would indicate the presence of a rectocele; such a defect, if present, may be made more prominent by having the patient strain. The finding of a rectal mass warrants further investigation.

Patients with other secondary causes of constipation (e.g., hypothyroidism and hypercalcemia) often respond to treatment of the underlying disease or manipulation of medications. If such measures are ineffective, the constipation should be treated symptomatically, in much the same fashion as idiopathic constipation is. Patients requiring long-term opioid administration for pain control generally experience constipation as a side effect, and this effect does not dissipate with time. Thus, many of these patients will require laxative therapy for the duration of their opioid use.³

Mild to Moderate Idiopathic Constipation

In patients who have mild to moderate symptoms and no findings from the history or the physical examination that would indicate a secondary cause, extensive investigations are not necessary. Routine colonoscopy is not mandatory for patients younger than 50 years. For patients older than 50 years, the baseline risk of colorectal cancer is sufficiently high that screening colonoscopy is recommended even in the absence of symptoms. These older patients should therefore undergo routine colonoscopy, and many authors recommend that patients younger than 50 years undergo routine flexible sigmoidoscopy. Random endoscopic biopsies are unnecessary, because idiopathic constipation is not associated with abnormalities on routine processing of mucosal biopsies.

Severe, Long-standing Constipation or Refractory Constipation

In patients who have very severe constipation or in whom medical management fails, further investigative tests are warranted. These tests are conducted to classify patients into three categories, each of which calls for a different treatment approach: (1) slow-transit constipation, (2) nonrelaxing puborectalis, and (3) normal-transit constipation.⁴ The initial investigations should include assessment of colonic transit time to determine if slow-transit constipation is present, as well as evaluation of pelvic floor function to determine if a nonrelaxing puborectalis is present.

Figure 2. Colonic transit study

There are two main methods for evaluating colonic transit: the radiopaque marker study and colonic scintigraphy. Both tests have advantages and disadvantages. In general, the choice between them depends on local expertise; the radiopaque marker study is more widely available. For the radiopaque marker study, 20 radiopaque markers (prepackaged in gelatin capsules) are ingested, and an abdominal x-ray (which includes the pelvis) is taken on day 5. The patient abstains from laxatives for the duration of the study. At 3 days, most patients with normal transit have excreted more than 80% of the markers; however, because there is substantial variation among asymptomatic persons, only patients who retain more than 20% of the markers for at least 5 days are considered to have abnormal transit. Abnormal transit may be demonstrated either throughout the colon or within a limited portion thereof (most commonly, the sigmoid and the rectum) [see Figure 2].

Colonic scintigraphy shares certain principles with the radiopaque marker study. Patients ingest a meal containing a radioactive isotope, and abdominal images are obtained with a gamma camera at 12, 24, and 48 hours. The results provide a quantitative assessment of colonic transit. In addition, unlimited numbers of images may be taken with the single isotope dose, and this feature of the test may be especially useful in children. For optimal accuracy, this technique requires standardization, and its availability is generally limited to centers with specific interest and expertise in it.

Pelvic floor studies are valuable for ruling out obstructed defecation as a cause of constipation. The balloon expulsion test can be performed in the office as an initial screening measure.⁵ A balloon filled with 50 ml of water is attached to tubing and placed in the rectum; patients with a nonrelaxing puborectalis generally cannot expel the balloon from the rectum in 1 minute while sitting on a commode. It should be kept in mind, however, that as many as 12% of patients with normal pelvic floor function will have difficulty with balloon expulsion in this setting.⁶

A thorough pelvic floor evaluation is best conducted in a pelvic floor laboratory with a specific interest in anorectal function. In addition to the balloon expulsion test, the evaluation generally involves manometry, including assessment of the reflexive relaxation of the internal sphincter after rectal distention. The presence of this reflexive relaxation rules out Hirschsprung disease as a cause of constipation. In patients with a nonrelaxing puborectalis, manometry during straining effort demonstrates abnormal function of the external sphincter—either failure to relax to enable expulsion or, on occasion, paradoxical contraction. Similar findings during straining can be documented by means of electromyography (EMG) with a sponge electrode in the anal canal.

Figure 3. Schematic representation of normal defecation

Defecography is commonly performed as well. Barium paste is formulated so as to simulate a fecal bolus and placed in the rectum. The patient is asked to defecate on a radiolucent commode, and the event is recorded with fluoroscopy. During normal defecation, the puborectalis and the anal sphincter muscles relax, and the rectum assumes a more vertical position with respect to the anal canal, facilitating evacuation of stool [see Figure 3]. In a patient with a nonrelaxing puborectalis, defecography typically demonstrates failure to open the anorectal angle and persistence of the puborectalis impression during defecation, as well as failure to empty completely.⁷ Other important findings that may be noted include rectocele, internal intussusception, and rectal prolapse. When appropriately selected, patients with obstructed defecation resulting from such abnormalities may benefit from surgical correction; however, even when these anatomic abnormalities are present, they may not be the underlying cause of constipation. Interpretation of defecography is subjective, and there is wide normal variation. Therefore, the diagnosis of a nonrelaxing puborectalis should be based not on a single test result but, rather, on the totality of the diagnostic findings.

The diagnosis of nonrelaxing puborectalis syndrome is made in persons with constipation in whom there is evidence of failure of the pelvic floor to relax appropriately (or paradoxical contraction of the pelvic floor) with defecation. The cause of this condition is not known; however, the syndrome is thought to be acquired over time. Patients with an underlying neurologic disorder (e.g., multiple sclerosis or Parkinson disease) are prone to spasticity of the puborectalis and may experience severe constipation as a result.

Many cases of constipation can be managed with dietary and lifestyle changes, such as modifying the diet to include foods high in fiber and drinking adequate amounts of water. Physical inactivity is associated with constipation, and encouraging moderate exercise may lead to significant symptomatic relief. Fiber supplementation is a key component of therapy for mild constipation.⁸ Fiber products (e.g., psyllium, methlycellulose, and polycarbophil) increase stool bulk and stimulate colonic motility. Such products must be taken with sufficient amounts of fluid, or they may lead to stool hardening. Often, patients have already tried fiber products but did not achieve satisfactory results because the quantities were insufficient; daily doses as high as 20 g may be necessary for a therapeutic effect. Patients taking fiber products may experience an increase in flatulence, particularly with fermentable fiber products. To improve tolerance, the amount of fiber should be increased gradually, and patients should be informed that the effect of fiber may not be seen immediately.

Nonlaxative therapy should be stressed; however, if dietary changes and fiber supplementation fail, judicious use of laxatives can bring about significant symptomatic relief. It should be kept in mind that tachyphylaxis to laxatives is common and may lead to chronic dependence. Stool softeners, or emollient laxatives (e.g., ducosate sodium and mineral oil), enhance penetration of water and fat into the stool, thereby making it less hard. These agents may be of use on a relatively short-term basis. Ducosate sodium is less effective than fiber supplementation⁹; stool softeners should not be used as a substitute for fiber.

Stimulant laxatives, including cascara, anthraquinones (senna and rhubarb), castor oil, and bisacodyl, are common components of popular over-the-counter medications. These agents have direct neuromuscular or mucosal effects, resulting in enhanced GI motility and altered mucosal transport (and thus increased intestinal secretion).⁸ Long-term use or abuse of anthraquinones can lead to melanosis (discoloration of the colonic mucosa caused by pigment deposition in colonic macrophages).

Osmotic laxatives contain compounds that either are not absorbed or are poorly absorbed. If the solutions are hypertonic, they cause water to move into the bowel lumen to maintain tonicity.¹⁰ Common preparations include magnesium and phosphate salts. Ingestion of large amounts of such preparations can lead to hypermagnesemia or hyperphosphatemia, mainly in patients with renal failure. The large fluid shifts that result when these compounds are used for bowel preparation may be dangerous in patients with underlying heart disease. Polyethylene glycol (PEG) is a high-molecular-weight compound that is not absorbed and thus functions as an osmotic laxative. PEG preparations are commonly administered as isotonic solutions and therefore cause only minimal fluid or electrolyte shifts when consumed rapidly (as in bowel preparation). PEG compounds are available as laxatives that can be taken either intermittently or regularly.

Tegaserod, a 5-HT₄ partial agonist, has been shown to alleviate bloating and increase stool frequency by improving gut motility and decreasing visceral sensitivity.¹¹ It may be prescribed for women with constipation-predominant irritable bowel syndrome (IBS) (see below) and for either male or female patients younger than 65 years who have idiopathic constipation. Tegaserod has been associated with the development of diarrhea; typically, the diarrhea resolves when the drug is discontinued, but occasionally, it is severe. In addition, several cases of ischemic colitis have been reported in patients receiving tegaserod. Although no causal relation has been established, patients should be warned to cease taking tegaserod and immediately contact their physician if abdominal pain worsens.

Enemas and suppositories act via a number of mechanisms, including softening of the stool, stimulation of rectal contraction by rectal distention, and direct alteration of mucosal secretion. They may be useful for occasional administration.

Constipation-Predominant Irritable Bowel Syndrome

In patients with constipation, significant abdominal pain, and no identifiable secondary cause of constipation, the diagnosis of constipation-predominant IBS may be appropriate [see Table 2].^12,13 Often, patients with constipation-predominant IBS respond to reassurance and fiber supplementation. Tegaserod may be employed in female patients who do not respond to conservative measures.

Severe, Long-standing Constipation or Refractory Constipation

Nonrelaxing puborectalis Patients with constipation arising from a nonrelaxing puborectalis often benefit from biofeedback.¹⁴ In this modality, a device (e.g., an anorectal manometer) is used to monitor pelvic floor activity; electrodes may also be used for EMG biofeedback. Patients observe pressure changes (or EMG activity) during attempts to evacuate. Through trial and error, they are taught to modify their responses until appropriate relaxation is achieved, the aim being to retrain the pelvic floor to relax during defecation. Training may have to be reinforced at intervals. Accurate determination of the success rate of biofeedback is difficult, in that the published literature consists primarily of case series and most of the trials that have been conducted have not included a placebo arm. It has been estimated that the success rate may be as high as 70%; however, this estimate is probably overoptimistic.¹⁵

If biofeedback fails, injection of botulinum toxin into the puborectalis under ultrasonographic guidance may be attempted. To date, published reports have evaluated this approach only in relatively small study groups; the results, though not decisive, are promising, in that the use of botulinum toxin clearly brought about noticeable improvements in manometric and defecographic findings¹⁶ and symptomatic improvements in the majority of patients.¹⁷ Other experimental techniques available for treatment of nonrelaxing puborectalis syndrome are electrogalvanic stimulation¹⁸ and sacral nerve stimulation (SNS).¹⁹ Currently, surgical approaches do not play a role in the treatment of constipation secondary to a nonrelaxing puborectalis.

Slow-transit constipation Slow-transit constipation, also known as colonic inertia, is most common in young women and often starts at puberty. It is characterized by abnormally slow forward propulsion of colonic contents. The cause of slow-transit constipation is unknown, though abnormalities in a number of cellular and neuromuscular modulators of GI motility have been found in patients with this condition.^20,21 Although patients with idiopathic slow-transit constipation are frequently resistant to laxative therapy, many respond to osmotic PEG laxatives. Surgery should be considered as an option only in the most severely affected patients, in whom aggressive laxative therapy has repeatedly failed over a prolonged period. Even in specialized centers, only about 5% of patients presenting with constipation are considered appropriate candidates for surgical treatment.²²

The operation most commonly performed to treat slow-transit constipation is subtotal colectomy with ileorectal anastomosis, performed via either an open or a laparoscopic approach. The colon is removed to the level of the sacral promontory in a standard fashion; the ileorectal anastomosis may be either stapled or handsewn. Constipation is less likely to recur with this anastomosis than with an ileosigmoid anastomosis.²³

Surgical therapy is generally successful in improving bowel function: in most patients, stool frequency rises to one to three bowel movements a day. Unfortunately, surgery may not satisfactorily alleviate other symptoms (e.g., abdominal discomfort or bloating),²⁴ and patients should be made aware of this possibility before operation. The key to successful surgical treatment is patient selection. Overall, the majority of well-selected patients are satisfied with the results of surgical treatment^25,26; however, long-term postoperative complications, particularly small bowel obstruction, are common. In addition, patients may manifest symptoms of a more global GI dysmotility disorder in the long term.

Other surgical approaches sometimes employed in this setting are ileostomy [see 5:30 Intestinal Stomas], and colectomy with cecorectal anastomosis; however, data on the long-term effectiveness of these approaches in large numbers of patients are lacking. Completion proctectomy with IPAA and ileostomy are options for patients who remain severely symptomatic after ileorectal anastomosis but who manifest no evidence of proximal dysmotility.

Not infrequently, patients have both slow-transit constipation and a nonrelaxing puborectalis. In such cases, it is essential that the obstructed defecation be addressed before any surgical treatment is carried out. Even after biofeedback, if surgical therapy is attempted in this setting, as many as 50% of patients will be dissatisfied with the results.²⁷

Severe idiopathic constipation Patients who have severe constipation but show no signs of slow-transit constipation, pelvic floor dysfunction, or IBS should be treated with reassurance and symptomatic management. Osmotic laxatives—in particular, PEG products—may be very useful in this group. Operative treatment plays no role in management; however, experimental approaches (e.g., SNS) are being evaluated for possible use in this setting.

Fecal incontinence may be defined as the involuntary loss of rectal contents through the anal canal. It is a relatively common condition, occurring in an estimated 2.2% of persons in the United States.²⁸ Its exact prevalence is unknown, however, and appears to vary with the population being studied. For example, nearly 50% of nursing home patients are incontinent to stool.²⁸ Fecal incontinence is often treated inadequately, either because of underreporting of symptoms to the physician²⁹ or because of ignorance or disinterest on the physician's part.

Fecal incontinence makes a significant contribution to medical morbidity (e.g., urinary tract infections and decubitus ulcers), but its main impact is on quality of life. Affected patients experience embarrassment and shame, and many dramatically alter their lifestyle in an effort to avoid accidents.

Normal continence depends on a chain of interdependent processes, and disruption of any of the links in the chain can lead to incontinence. Frequently, a combination of factors is responsible for the incontinence.

To care about continence, persons must have adequate mental function, and to maintain normal continence, they must have an intact neurologic arc from the brain to the anal sphincter. A wide array of neurologic disorders can lead to incontinence, including dementia, strokes, spinal cord injury, multiple sclerosis, and diabetic autonomic neuropathy. So-called idiopathic fecal incontinence is caused by pelvic floor denervation resulting from traction injury to the pudendal nerves.³⁰ The injury is usually caused by straining and consequent pelvic floor descent during obstetrical delivery or by chronic straining at stool.

Conditions characterized by abnormal GI function, especially diarrheal states, can cause or exacerbate incontinence. Common causative conditions include infectious diarrhea and inflammatory bowel disease. Diarrhea-predominant IBS can contribute to incontinence in patients with other associated disorders. Fecal impaction is an important cause of incontinence, particularly in older and institutionalized populations.³¹

Figure 4. Obstetric sphincter injury

Abnormalities of the pelvic floor are frequent causes of incontinence. Some such abnormalities are congenital malformations (e.g., imperforate anus, rectal agenesis, and cloacal defect). More often, abnormalities are attributable to acquired sphincter injuries. Common causes of sphincter injury include obstetric injury, pelvic fracture, and traumatic impalement [see Figure 4]. One of the most frequent causes is an anorectal procedure, such as fistulotomy,³² sphincterotomy,³³ or anal dilatation.³⁴ Sphincter-sparing rectal resections can also lead to incontinence as a consequence of both the loss of the normal rectal reservoir and the sphincter injury caused by transanal introduction of intraluminal staplers.

Clinical Evaluation

History

Figure 5. Workup and management of fecal incontinence

A careful patient history and a directed physical examination are the most important elements of clinical evaluation for a patient with fecal incontinence [see Figure 5]. The patient should be asked about the onset and nature of the incontinence (e.g., whether the stool is liquid or solid and whether flatus is present), any associated changes in stool consistency or bowel habits, and the frequency of incontinence. A pertinent but thorough medical, surgical, and obstetric history should be obtained, and any underlying contributory conditions (e.g., colitis) should be treated. The impact of the incontinence on the patient's quality of life should be assessed, at least qualitatively.

Physical Examination

Physical examination should focus primarily on the perineum. Seepage and secondary perineal skin breakdown should be noted, as should scars from previous surgical treatment or trauma. Perineal body deformity is an important sign of obstetric injury, and gaping of the anus with traction on the buttocks is suggestive of rectal prolapse. When prolapse is suspected but not evident, the patient should be asked to strain while seated on a commode. Digital rectal examination is useful for detecting low rectal tumors and fecal impaction; it also provides a qualitative assessment of both resting sphincter tone and voluntary squeeze pressure.

Investigative Studies

Endoscopy should be performed on all incontinent patients to exclude a neoplastic or inflammatory condition. In most cases, flexible sigmoidoscopy is adequate, but if the patient has unexplained diarrhea, bleeding, or changed bowel habits, complete colonoscopy should be performed.

Figure 6a. Endoanal ultrasonogram: normal sphincter

Figure 6b. Endoanal ultrasonogram: sphincter defect

Anorectal testing is indicated for most patients with significant incontinence, particularly if operative treatment is being considered. The most important test is endoanal ultrasonography (EAUS), which yields a highly accurate assessment of sphincter integrity [see Figures 6a and 6b].³⁵ At some centers, magnetic resonance imaging has become the test of choice for evaluating the pelvic floor. Anal manometry provides a quantitative assessment of resting and squeeze anal pressures, which serve as indicators of internal anal sphincter function and external anal sphincter function, respectively. EMG may be used to diagnose neuropathic injury of the pelvic floor. Although concentric-needle EMG is the most accurate technique, most centers employ a glove-mounted intra-anal electrode to measure pudendal nerve conduction time (i.e., pudendal nerve terminal motor latency [PNTML]). The practical utility of PNTML testing is debatable, however, and opinions vary regarding the test's ability to predict successful outcomes after anal sphincter repair.^36,37 When the cause of incontinence is uncertain, dynamic imaging of the pelvic floor with defecography or MRI may reveal an occult pathologic state (e.g., occult rectal prolapse).

Biofeedback appears to be an effective therapy for fecal incontinence in a high percentage of patients.^38,39 It is an inherently attractive approach because it is simple, painless, and risk-free. However, the biofeedback literature consists mostly of small, uncontrolled, retrospective studies; a randomized, controlled trial from 2003 found that biofeedback had no advantages over standardized medical and nursing care (i.e., advice) or advice plus sphincter exercises.⁴⁰

Sphincteroplasty

Figure 7. Sphincteroplasty

Anal sphincter repair is the most widely accepted operation for fecal incontinence [see Figure 7]. In acute situations (e.g., when an obstetric sphincter injury is recognized), immediate direct repair is generally recommended. Unfortunately, as many as 75% of women have persistent external anal sphincter defects after primary repair, and about 60% have some degree of incontinence.⁴¹ If immediate repair is not attempted, surgical treatment should be delayed at least 3 to 6 months to permit resolution of local tissue inflammation and edema.

For incontinent patients with established sphincter defects, overlapping sphincteroplasty is the procedure of choice. Complete bowel preparation is carried out before the procedure, and prophylactic antibiotics are administered.

Operative technique Step 1: initial dissection. The patient is placed in the prone jackknife position, with the buttocks taped apart and a large roll beneath the hips. A curvilinear incision is made over the perineal body, and the anoderm and the anal canal mucosa are raised as an endodermal flap [see Figure 7, part a]. The vaginal wall is mobilized anteriorly.

Step 2: mobilization of sphincter muscle. It is often easiest first to identify normal muscle laterally in the ischiorectal fossa and then to work medially toward the attenuated tissue in the midline. Lateral dissection is extended back on either side until enough healthy muscle is mobilized to allow overlapping without tension. Generally, however, lateral dissection should not extend beyond the midcoronal line, so as not to risk injury to the inferior rectal branches of the pudendal nerves, which cross the ischiorectal fossae posterolaterally. Dissection is then carried out cranially in the rectovaginal septum to the level of the puborectalis. The muscle is divided through its midline scar, but the scar is preserved to help prevent the sutures from tearing through.

Step 3: overlapping repair. The tapes on the buttocks are then re leased, and an overlapping sphincter repair is performed with absorb able mattress sutures [see Figure 7, parts b and c]. A snug plication is universally advocated, but unfortunately, there are no generally accepted objective criteria to define exactly what 'snug' means in this context. Many authorities advise plication of the puborectalis (so-called levatorplasty) at the cranial aspect of the repair to maximize the length of the anal canal.⁴² Others favor individual dissection and repair of the internal and external sphincter muscles, but at present, there is no compelling evidence for the superiority of this approach.

Step 4: restoration of perineal body. The skin incision is closed in a V-Y configuration [see 3:7 Surface Reconstruction Procedures] to restore the perineal body and maximize the distance between the anus and the vaginal introitus. The wound is left partially open or closed loosely over small Penrose drains to minimize the risk of surgical site infection [see Figure 7, part d]. A diverting stoma is not generally indicated but may be considered in special situations (e.g., multiple previous failed repairs, Crohn disease, or various chronic diarrheal states).

Outcome evaluation Overlapping sphincteroplasty yields substantial clinical improvement in approximately 65% to 80% of patients.^43,44 Unfortunately, current data indicate that results deteriorate significantly over time.^45–47 When sphincteroplasty fails, repeat EAUS evaluation should be done to confirm that the muscle wrap is intact, and another repair should be performed after 6 to 12 months if a significant defect persists.⁴⁸ If the muscle wrap is intact, the functional outcome can often be improved by means of biofeedback.⁴⁹

Various surgical options are available for patients in whom sphincteroplasty has failed or who are not candidates for the procedure (e.g., those with pudendal neuropathy and an anatomically intact sphincter). A number of these options are investigational, and further study is needed to determine their eventual role (if any) in incontinence therapy.

Postanal Repair

Sir Alan Parks devised the postanal repair in 1975 to treat patients with incontinence and intact sphincters. The initial results were encouraging but tended to deteriorate over time. Consequently, despite evidence of lasting improvement in some patients, this operation is rarely performed today.^50,51

Injectable Biomaterials

A number of studies have explored the use of injectable biomaterials to provide bulk around the anal sphincter and thereby improve continence. The materials employed have included autologous fat, cross-linked collagen, silicone, and carbon-coated beads.^52,53 Several small, uncontrolled studies have reported promising results, but larger series with longer follow-up times are needed.

Nonstimulated Muscle Transposition

Attempts to restore continence by creating a neosphincter from transposed skeletal muscle date back to the early 20th century. Most such attempts have made use of either the gluteus maximus⁵⁴ or the gracilis.⁵⁵ Good results have frequently been reported, but many authorities believe that the quality of the resulting continence is poor. One of the main limitations of nonstimulated muscle transposition is that patients are typically unable to maintain voluntary contraction of the transposed muscle over the long term.

Stimulated (Dynamic) Graciloplasty

Successful electrical stimulation of a transposed gracilis by means of an implantable pulse generator was first reported in 1988.⁵⁶ Such stimulation has two main effects. First, it converts the fast-twitch, rapidly fatigable gracilis to a slow-twitch, fatigue-resistant muscle that is capable of tonic contraction for prolonged periods.⁵⁷ Second, electrical stimulation maintains tonic muscle contraction without the need for continuous voluntary control on the part of the patient. A small number of centers with particular expertise in dynamic graciloplasty and high patient volumes have reported good results with acceptable morbidities⁵⁸; however, three large multicenter trials have reported less encouraging results with prohibitive morbidities.^59–61 In the United States, dynamic graciloplasty is not available, because it has not been approved by the Food and Drug Administration. Elsewhere in the world, the operation can be considered a salvage option at centers with the requisite expertise and experience.

Artificial Anal Sphincter

Figure 8a. Artificial anal sphincter

Figure 8b. Placement of artificial anal sphincter

The artificial anal sphincter is an implantable system consisting of three parts: an inflatable perianal cuff, a pressure-regulating balloon, and a control pump that is implanted in the scrotum or the labia majora [see Figures 8a and 8b]. Good results have been reported in individual case series,⁶² but device infection has been a problem.^63,64 In a large multicenter trial, 46% of patients required surgical revision of the device, including 25% who required revision or explantation because of infection. Of the patients who underwent implantation, 53% had successful results; among those with a functioning device in place, the success rate was 85%.

Sacral Nerve Stimulation

In SNS, an electrode is inserted through a sacral foramen and used to stimulate the sacral nerves. To date, the procedure has been employed mainly in patients with intact anal sphincters (including those with intact repairs). It is available for treatment of fecal incontinence in Europe but has not yet received FDA approval for this indication in the United States.

SNS is generally carried out in two stages. The first stage, peripheral nerve evaluation (PNE), is performed to confirm a muscular response to stimulation of the sacral nerves, to identify the optimal site for stimulation (S2, S3, or S4) and to determine the clinical response to stimulation with an external pulse generator. In most cases, stimulation of the S3 nerves provides the optimal response.

Figure 9a. Sacral nerve stimulation: lead

Figure 9b. Identification of sacral foramina

Figure 9c. Quadripolar lead in position

PNE is performed with the patient prone and under local anesthesia, with or without sedation. The sacral foramina are located by means of bony landmarks; S3 is typically about 1.5 cm off the midline at the level of the sciatic notch.⁶⁵ Initial testing is performed with an insulated spinal needle and an external pulse generator. Stimulation of each foramen leads to a typical response: S3 causes a bellows-type contraction of the pelvic floor and dorsiflexion of the ipsilateral great toe. Usually, several levels are tested until the optimal site is identified. A temporary pacing wire or a permanent quadripolar lead is then inserted and connected to an external stimulator [see Figures 9a, 9b and 9c].

Patients are asked to provide a baseline continence diary, and a second diary is recorded during the test stimulation period. If continence is significantly improved (e.g., by 50% or more), the second stage of SNS, implantation of a permanent lead (if not already in place) and a pulse generator, is carried out. This second stage is also performed with the patient prone, under local anesthesia, and sedated. The pulse generator is implanted in a subcutaneous pocket on the same side as the stimulating electrode.

Both stages of SNS are performed as outpatient procedures. The pulse generator is activated and its stimulation parameters set by means of a telemetric programmer. If problems (e.g., pain) develop or if the results of stimulation are inadequate, the system can be reprogrammed in a variety of ways: stimulation frequency can be altered, voltage can be increased or de creased, and the configuration of the stimulating electrodes can be modified.

SNS has been shown to be a highly effective treatment for fecal incontinence.^66–68 Unlike dynamic graciloplasty and the artificial anal sphincter, SNS is associated with only minimal morbidity. In a multicenter prospective trial, the frequency of incontinent events dropped from 16.4/wk at baseline to 3.1/wk at 12 months after SNS and 2.0/wk at 24 months. Fecal incontinence-related quality of life was significantly improved.

Because of its high success rate and excellent safety profile, many authorities now consider SNS the salvage procedure of choice for patients with refractory incontinence. If SNS fails, more aggressive treatments may still be tried at a later time.

Colostomy

Although creation of a colostomy does not restore continence, it does provide a degree of bowel control in a manner that allows patients to resume their normal activities without fear of accidents. Surprisingly few data are available regarding colostomy for incontinence; however, one questionnaire study of patients who underwent colostomy for incontinence reported extremely high levels of patient satisfaction and marked improvements in subjective quality of life.⁶⁹ In most cases, a simple end sigmoid colostomy with a Hartmann pouch is the appropriate procedure, and it can often be performed with relatively little operative trauma by using a laparoscopic or minilaparotomy technique. Patients should receive preoperative counseling from an enterostomal therapist, and the optimal stoma site should be marked before the procedure is initiated.