Fecal incontinence is the recurrent uncontrolled passage

Transcription

1 GASTROENTEROLOGY 2003;124: Fecal Incontinence ADIL E. BHARUCHA Clinical Enteric Neuroscience Translational and Epidemiological Research (C.E.N.T.E.R.) Program, Division of Gastroenterology and Hepatology, Department of Medicine Mayo Clinic, Rochester, Minnesota Fecal incontinence is the recurrent uncontrolled passage of fecal material, of 1 month or greater duration, in an individual with a developmental age of at least 4 years. 1 Patients who have fecal incontinence are also distressed by involuntary passage of flatus. However, involuntary passage of flatus alone should probably not be characterized as fecal incontinence, partly because it is difficult to define when passage of flatus is abnormal. Although the physical consequences of incontinence (e.g., perianal dermatitis) are modest, the psychosocial consequences are often overwhelming, and include diminished self-esteem, social isolation, and anxiety of having unexpected episodes. Fecal incontinence jeopardizes employment, and may lead to institutionalization. 2 4 Unfortunately, physicians may not always appreciate the devastating consequences of incontinence, perhaps because patients are often embarrassed to discuss the symptom. In addition to increasing awareness of the disorder, newer diagnostic (e.g., magnetic resonance imaging [MRI]), and therapeutic modalities may enhance our understanding of pathophysiology and facilitate management. The limitations of other diagnostic modalities (e.g., pudendal nerve latencies) are being appreciated. Moreover, the overlap between uro-gynecological, and anorectal manifestations of pelvic floor disorders was apparent at a recent consensus conference on urinary and fecal incontinence. Epidemiology Population-based studies of the epidemiology of fecal incontinence avoid the referral bias of single-institution based studies (Table 1). The prevalence ranges from 2.2% in the Wisconsin survey to 15.3% in Olmsted County, Minnesota, suggesting that fecal incontinence is a common problem. Different prevalence rates between studies probably reflect differences in survey methods, in age distribution of population surveyed, and varying definitions of fecal incontinence. Whereas most attention has been focused on fecal incontinence in women, the prevalence in men is comparable to women (Table 1). The prevalence increases steadily with age, from approximately 4% for any incontinence in men and women aged between 40 and 49 years old, to 11.6% in patients aged 80 years and older; incontinence was more frequent and severe in older compared to younger patients. 5 The prevalence in nursing home residents is considerably higher, approaching 47% in a survey of 18,000 nursing home residents in Wisconsin. 6 Incontinence worsened quality of life in more than 50% of those reporting major incontinence, defined as soiling of underwear, outer clothing, furnishings, or bedding several times a month or more often. 5 In another survey, 33% of patients restricted activities due to incontinence. 7 Thus, these surveys document the prevalence, severity, and general risk factors for fecal incontinence (e.g., age older than 65 years, female sex, poor general health, and physical limitations 7 ). However, the incidence, and natural history of fecal incontinence are unknown. Epidemiological studies also hold promise for dissecting the relative contributions of obstetric and nonobstetric risk factors to idiopathic fecal incontinence (see below). Pathophysiology Anatomical factors that prevent fecal evacuation include the pelvic barrier, rectal curvatures, and transverse rectal folds; recto-anal sensation and rectal compliance also help maintain continence. Pelvic Barrier The internal anal sphincter is a thickened extension of the colonic circular smooth muscle layer that maintains approximately 70% of anal resting tone (Figure 1). Tonic sympathetic excitation contributes to internal sphincter tone. 8 The external anal sphincter is a tonically active striated muscle that accounts for the remaining component of resting tone; the external sphincter, puborectalis, and levator ani contract further when necessary to preserve continence, but relax nearly completely during evacuation. External sphincter contraction may be voluntary, or reflex Abbreviations used in this paper: EMG, electromyography; MRI, magnetic resonance imaging; US, ultrasound by the American Gastroenterological Association /03/$30.00 doi: /s (03)

2 May 2003 FECAL INCONTINENCE 1673 Table 1. Epidemiology of Fecal Incontinence: Community-Based Studies Survey Respondents; instrument Response rate (number respondents) Prevalence Drossman 99 US householder marketing list; Mailed questionnaire 66% (5,430) Soiling - 6.9%(F); 7.4%(M) Gross incontinence 0.9% (F); 0.5% (M) Nelson (1993) 7 Wisconsin residents of all ages; Phone interview with 1 member in each household 73% (6,959) Any FI over past year 2.2% (overall) 7.5% (aged 65) Talley (1990) 100 Olmsted County residents 65 years; Mailed questionnaire 66% (328) FI once per week over past year 3.1% (F); 4.5% (M) Reilly (1994) 28 Olmsted Country residents 50 years; Mailed 64% (1,540) Any FI 17.8% (F); 12.8% (M) questionnaire Perry (2002); UK 5 Leicestershire Health Authority patient register; Mailed questionnaire 70% (10,226) Any FI 5.7% (F); 6.2% (M) NOTE. Prevalence rates for males (M) and females (F) are provided separately where available. FI, fecal incontinence. (e.g., when intra-abdominal pressure increases). Anal resting and/or squeeze pressures are generally reduced in fecal incontinence, suggesting sphincter weakness (Table 2). Common causes of anal sphincter weakness include sphincter damage, neuropathy, or reduced input from the cortex or spinal cord (Table 3). Although gastroenterologists tend to focus on the anal sphincters, the levator ani muscles, including the puborectalis, are also important constituents of the pelvic barrier. 9 The puborectalis is a U-shaped component of the levator ani complex, which blends with the upper aspect of the external sphincter, maintaining the rectoanal angle at rest; it contracts further to reduce the angle when patients squeeze. A recent study suggested that the inward traction exerted by the puborectalis was reduced in fecal incontinence, correlated more closely with symptoms than squeeze pressures, and improved after biofeedback therapy. 10 Pelvic floor weakness may also manifest as the descending perineum syndrome, which results from a combination of obstetric trauma, chronic straining, and/or a neuropathy. 11 The natural history of the descending perineum syndrome is characterized initially by constipation-predominant symptoms, and subsequently by fecal incontinence. Constipation, especially excessive straining during evacuation, may contribute to pelvic floor weakness. 12 Ultimately, excessive descent causes a stretch-induced pudendal neuropathy, and makes the anorectal angle obtuse, impairing the flap valve that normally maintains continence when intra-abdominal pressure increases. Sphincter pressures were lower in incontinent patients when compared with continent patients who have descending perineum syndrome. This reinforces the importance of anal sphincters for maintaining continence, particularly when the anorectal angle is obtuse. 11 Rectal Compliance and Sensation The rectal segment above the middle fold is derived from the embryological hindgut, may contain feces, and is free to distend toward the peritoneal cavity (Figure 1). The lower rectum, situated below the middle rectal fold, is derived from the cloaca, surrounded by condensed extraperitoneal tissue, and is empty except during defecation. Stool is often transferred into the rectum by colonic high-amplitude propagated contractions, which tend to occur after awakening or meals. 13 Rectal distention by stool is associated with several processes that serve to preserve continence, or if circumstances are appropriate, proceed to defecation. Rectal distention induces reflex relaxation of the internal anal sphincter. Whereas sigmoid colonic balloon distention is accompanied by abdominal sensation, rectal balloon distention is Figure 1. Diagram of a coronal section of the rectum, anal canal and adjacent structures. The pelvic barrier includes the anal sphincters and pelvic floor muscles.

3 1674 ADIL E.BHARUCHA GASTROENTEROLOGY Vol.124, No.6 Table 2. Etiological factor Anorectal Sensorimotor Disturbances in Fecal Incontinence Anal sphincter pressure Threshold for internal sphincter relaxation Threshold for external sphincter contraction Rectal sensation a Rectal compliance Pelvic floor function Idiopathic or Diabetes mellitus 101 R 2; S Multiple sclerosis 101 R 7; S Elderly patients with R 7; S fecal impaction and incontinence 102 Acute radiation R 2; S2 NA NA 7 2 NA proctitis 103 Chronic radiation NA NA NA 1 2 NA injury 104 Ulcerative colitis 105,106 S 2 incontinent 2 (active NA 1 (active 2 (active NA patients colitis only) colitis only) colitis only) Spinal cord injury R 7; S NA high spinal lesion; (T12 or higher) 37 Low spinal lesion (below T12) R 2; S NA NOTE. Information pertains to patients who have underlying disease and fecal incontinence. 1 Increased; 2 decreased; 7 no change. R, resting; S, squeeze sphincter pressure; NA, not available. a Rectal sensation expressed as volume thresholds for perception; 1 sensation indicates volume threshold for perception was lower compared to normals. perceived as a sensation of rectal fullness, as if the rectum were uncomfortably full of flatus or feces. If defecation is inconvenient, the desire to defecate prompts voluntary contraction of the external sphincter, 14 which wanes together with the sense of urgency as the rectum relaxes, accommodating to hold more stool. The factors that determine whether rectal distention is interpreted as a desire to defecate or to pass flatus are unclear. Sphincter pressures do not always distinguish continent from incontinent patients, emphasizing the importance of rectal compliance and sensation for maintaining continence. Reduced rectal sensation allows stool to enter Table 3. Etiology of Fecal Incontinence Anal sphincter weakness Injury: obstetric trauma, related to surgical procedures (e.g., hemorrhoidectomy internal sphincterotomy, fistulotomy, anorectal infection) Nontraumatic: scleroderma, internal sphincter thinning of unknown cause Neuropathy Stretch injury, obstetric trauma, diabetes mellitus Anatomical disturbances of pelvic floor Fistula, rectal prolapse, descending perineum syndrome Inflammatory conditions Crohn s disease, ulcerative colitis, radiation proctitis Central nervous system disease Dementia, stroke, brain tumors, spinal cord lesions, multiple system atrophy (Shy Drager s syndrome), multiple sclerosis Diarrhea Irritable bowel syndrome, postcholecystectomy diarrhea the anal canal, and perhaps leak before the external sphincter contracts Conversely, exaggerated rectal sensation, perhaps a marker of coexistent irritable bowel syndrome, 15,18 is associated with reduced rectal compliance, repetitive rectal contractions during rectal distention, external sphincter weakness, and exaggerated anal sphincter relaxation during rectal distention. These observations confirm that fecal incontinence is a heterogenous disorder; patients often suffer from more than one deficit (Table 2). 15 Anal sphincter relaxation may occur during or independent from rectal distention, enabling the anal lining to ascertain whether rectal contents are gas, liquid, or stool. 19 Read and Read observed that continence for saline was not affected when the anal canal was anesthetized by lidocaine in 9 healthy patients, suggesting that anal sensation was not critical for continence in healthy patients. However, the anal sampling response may be involved in maintaining continence when other factors predispose to incontinence. Indeed, the sampling response occurred less frequently and anal sensation was reduced in incontinent patients, perhaps depriving them of sensory information. 19 Other Factors In addition to normal anorectal function, continence is preserved by normal stool consistency, mental faculties, and mobility. 20

4 May 2003 FECAL INCONTINENCE 1675 Etiology Table 3 lists conditions associated with fecal incontinence. Fecal incontinence is generally attributable to a combination of impaired pelvic floor continence mechanisms and disordered bowel habits. Before endoanal ultrasound (US) imaging was available, unexplained sphincter weakness was considered idiopathic or attributed to a pudendal neuropathy. Clinically overt anal tears occurred in approximately 3.3% of women after a vaginal delivery. 21 However, endoanal US identified anal sphincter defects in 35% of women after their first vaginal delivery. 22 Internal and external sphincter defects were associated with significantly lower resting and squeeze pressure, respectively. 22 Anal US also identified internal sphincter thinning in fecal incontinence, 23 fostering the concept that idiopathic fecal incontinence was often attributable to anal sphincter injury during vaginal delivery. However, the precise role of obstetric anal sphincter injury to fecal incontinence is unclear. The studies assessing sphincter injury after vaginal delivery did not enroll a consecutive series of patients. 22 The incidence of postpartum anal sphincter defects by 3-dimensional US is only 11%, i.e., lower than reported previously by two-dimensional imaging; an additional 18% had tears of the puboanalis or transverse perinei which may not impair the pelvic barrier. 24 The incidence of postpartum fecal incontinence is much lower; ranging from 0% in one study 25 to around 10% in several reports. 22,26,27 Lastly, the median age of onset of fecal incontinence in women living in the community is 61 years; the significance of clinically occult anal sphincter defects to fecal incontinence occurring 2 to 3 decades after vaginal delivery is unclear. 28 Thus, other factors, as yet poorly defined, including aging, menopause, chronic straining, and disordered bowel habits also likely predispose to fecal incontinence. Although the prevalence of fecal incontinence in the community is comparable in men and women, only a few studies have focused on pathophysiology of fecal incontinence in men. 29 In men, incontinence is frequently associated with, and perhaps attributable to local causes (e.g., hemorrhoids, fistula, poorly healed surgical scars or proctitis after radiotherapy for prostate cancer). Idiopathic fecal soiling or leakage in men may be caused by a long anal sphincter of high pressure that entraps small particles of feces during defecation and subsequently expels them, causing perianal soiling and discomfort. 30 Approximately 5% of patients develop chronic anorectal complications, i.e., fistula, stricture and disabling fecal incontinence after pelvic radiotherapy. 31 Surgical procedures that may contribute to fecal incontinence include sphincterotomy and fistulotomy. Postoperative fecal incontinence affects approximately 45% of patients after a lateral internal sphincterotomy; 6%, 8% and 1% reported incontinence to flatus, minor fecal soiling and loss of solid stool, respectively, 5 years thereafter. 32 Similarly, the risk of fecal incontinence after a fistulotomy approaches 18% to 52%, but is perhaps less with recent modifications. 33 Several neurological disorders are associated with incontinence. Anal sphincter weakness, diminished rectoanal sensation, and diarrhea predispose to fecal incontinence in patients who have diabetic neuropathy; impairment of anorectal function generally parallels the duration of disease. 34 A case-control study observed a similar prevalence of fecal incontinence in type 1 diabetics and controls; the prevalence in type 2 diabetics was not significantly higher compared with controls (4.6% vs. 1.8% for controls). 35 Fifty-one percent of a group of unselected outpatients with multiple sclerosis had fecal incontinence. 36 Constipation is the predominant symptom after supraconal spinal cord injury; anal resting pressure is relatively preserved and fecal incontinence is relatively uncommon. In patients who have spinal cord lesions at or below T 12, resting anal sphincter tone is often reduced and together with blunted rectoanal sensation 37 and laxatives predisposes to fecal incontinence. Clinical Evaluation The clinical evaluation is the cornerstone of arriving at an accurate diagnosis, establishing rapport with the patient, and formulating a logical strategy for diagnostic testing and treatment. Patients who have chronic diarrhea, fecal urgency, constipation, prolapsed hemorrhoids, urinary incontinence, dementia, and diabetes mellitus must be asked whether they have fecal incontinence. Staining, soiling, seepage, and leakage are terms used to reflect the nature and severity of incontinence. Soiling indicates more leakage than staining of underwear; soiling can be specified further (i.e., of underwear, outer clothing, or furnishing/bedding). Seepage refers to leakage of small amounts of stool. All scales for rating severity of fecal incontinence quantify consistency and frequency of stool loss; some scales also incorporate number of pads used, severity of urgency, and the impact of fecal incontinence on coping mechanisms and/or behavioral changes. None of the several available continence scores 38,39 have been universally adopted in clinical trials. The St. Mark s and Pescatori scales emphasize the frequency and physiological aspects

5 1676 ADIL E.BHARUCHA GASTROENTEROLOGY Vol.124, No.6 Table 4. Components of a Comprehensive History in Fecal Incontinence Question Elucidate whether a patient has incontinence Onset, natural history, and risk factors Bowel habits/type of leakage Degree of warning before incontinence Diurnal variation in incontinence Impact of fecal incontinence on quality of life Urinary incontinence - presence and type Evaluate possible causes for Incontinence Rationale Patients may not volunteer the symptom spontaneously Relationship of symptom onset/deterioration to other risk factors may suggest etiology Natural history (e.g., recent symptomatic deterioration) may reveal reason for seeking medical attention Disordered bowel habits are critical to pathogenesis of incontinence Incontinence for solid stool suggests more severe sphincter weakness than for liquid stool Management should be tailored to specific bowel disturbance Urge and passive incontinence are associated with more severe weakness of the external and internal anal sphincter, respectively These symptoms may also reflect rectoanal sensory disturbances, potentially amenable to biofeedback therapy Nocturnal incontinence occurs uncommonly in idiopathic fecal incontinence, and is most frequently encountered in diabetes and scleroderma Critical to ascertain severity of incontinence Association between urinary and fecal incontinence 107 Same therapy may be effective for both conditions Most neurological conditions (Table 2) are associated with other, i.e., non-anorectal manifestations The obstetric history must inquire specifically for known risk factors for pelvic trauma, e.g., forceps delivery, episiotomy, and prolonged second stage of labor Medications, including laxatives, and artificial stool softeners may cause or exacerbate incontinence of incontinence. 38,40 Physicians and patients generally agree that infrequent incontinence for flatus reflects mild, while frequent leakage of liquid and stool reflects severe incontinence. 39 In other areas, physicians and patients may disagree; patients attach greater significance to frequent leakage of gas compared to physicians, while physicians attach greater significance to leakage of solid stool. Consequently, it is vital to ascertain quality-of-life, reflected not only by items connected with coping, behavior, self-perception, and embarrassment, 4 but also practical day-to-day limitations (e.g., the ability to socialize and get out of the house). 41 Patients are affected even by the possibility and unpredictability of incontinence episodes 41 ; conversely, behavioral adjustments may also influence the type and frequency of incontinent episodes. Thus, the type and frequency of incontinent episodes alone will underestimate the severity of the disorder in a patient who is housebound, having withdrawn from professional and social activities. Table 4 outlines a logical sequence of questions designed to obtain a useful history from incontinent patients. The importance of carefully characterizing bowel habits cannot be overemphasized; pictorial stool scales may avoid misinterpretation of the stool form and consistency. 42 Semiformed or liquid stools pose a greater threat to pelvic floor continence mechanisms than formed stools. Conversely, patients who have severe constipation, perhaps accompanied by fecal impaction, often develop spurious diarrhea and incontinence in the setting of laxative use. The awareness of the desire to defecate before the incontinent episode is variable, and may also provide clues to pathophysiology. Patients who have urge incontinence experience the desire to defecate, but cannot reach the toilet on time. Patients who have passive incontinence are not aware of the desire to defecate before the incontinent episode. Patients who have urge incontinence have reduced squeeze pressures 43 and/or squeeze duration 44 whereas patients who have passive incontinence have lower resting pressures. 43 Urge and passive incontinence also suggest increased or reduced rectoanal sensation, respectively. However, the association between symptoms and rectal sensation in fecal incontinence has not been assessed. A multisystem examination guided by the history, and knowledge of underlying diseases is necessary. A complete anorectal examination should be conducted in the left lateral decubitus and seated positions if necessary. The positive predictive value of examination for identifying low resting and squeeze pressure is 67% and 81%, respectively. 45 In addition to assessing the sphincter, the examination must also assess for the normal upward and anterior movement of the puborectalis, a lift, as the patient squeezes, perianal sensation, and the anal wink reflex. Examination in the seated position on a commode may be more accurate than the left lateral decubitus position for characterizing rectal prolapse, pouch of Douglas hernia, or excessive perineal descent.

6 May 2003 FECAL INCONTINENCE 1677 Table 5. Function Anal sphincter-levator ani complex Assessment of Anorectal Functions Maintaining Continence Rectal accommodation Rectal sensation Anal sensation Diagnostic Testing Measurement Anal manometry (saline continence test - anorectal barrier function) Endoanal ultrasound (MRI - sphincter integrity) Evacuation proctography (dynamic MRI - pelvic floor motion) Pudendal nerve latency (sphincter EMG - neural injury) Rectal compliance Rectal perception of distention by manual syringe or barostat Mucosal electrosensitivity, temperature perception NOTE. Italicized tests are used in research studies, but not widely available, or used in clinical practice. The extent of diagnostic testing is tailored to the patient s age, probable etiological factors, symptom severity, impact on quality of life and response to conservative medical management (e.g., loperamide or stool softeners) (Table 5). The availability of tests, and physician biases may also influence testing. The strengths and limitations of these tests will be discussed. Endoscopy Endoscopic assessment of the rectosigmoid mucosa should be considered in most patients, particularly those who have constipation and/or diarrhea; a colonoscopy may be desirable in certain circumstances (e.g., if the differential diagnosis includes colon cancer). Anal Manometry Anal pressures are measured by withdrawing a catheter with perfused or solid-state transducers by the station pull-through method; resting and squeeze pressures are recorded as the transducer is stationary at serial 1 cm intervals from the rectum to the anal verge. Waterperfused transducers are considerably cheaper and probably equivalent in other respects to solid-state manometry for measuring sphincter pressures. Although anal manometry is technically undemanding and widely available, variations in catheter design, definitions, and methods to calculate average and maximum resting and squeeze pressures between centers should be borne in mind when interpreting pressures. Given asymmetry, pressures should be calculated by averaging forces exerted in all 4 quadrants. Tracings should be scrutinized to ensure pressures calculated by automated software programs are accurate. Because normal values are strongly influenced by technique, age, and gender, anal pressures should be compared to normal values obtained using the same technique. 46,47 In one study, maximum squeeze pressure less than 60 mm Hg in females had 60% sensitivity, and 78% specificity for discriminating between asymptomatic controls and fecal incontinence. 48 Another large study of 302 incontinent patients found maximum squeeze pressure had more than 90% sensitivity and specificity for discriminating between controls and fecal incontinence (using mean 2 standard deviation (SD) for controls as the cutoff). 15 The variable utility of anal pressure measurements may relate to methodological limitations, and/or reflect the multifactorial nature of the disorder. Other factors (e.g., disordered stool consistency, rectoanal sensation, and/or rectal compliance) should be considered particularly in incontinent patients with normal sphincter pressures. 49 Endoanal Ultrasound Endoanal US identifies sphincter thinning and defects, 50 which are often clinically unrecognized 22 and/or amenable to surgical repair (Figure 2). 51 Endoanal US identified all 9 surgically verified external sphincter defects in one study 52 ; US also identified thinning of the internal sphincter, 53 interpretation of external sphincter images is more subjective, operator-dependent, and confounded by normal anatomical variations in the external sphincter 50,54 ; there is substantial interobserver variability. 55,56 The cause of internal sphincter thinning is unknown. Because the external sphincter is often asymmetric in the upper anal canal, particularly in women, it may be difficult to distinguish a normal variant from a sphincter defect. 50,57 As discussed previously, the extent to which clinically occult isolated external anal sphincter defects are responsible for fecal incontinence is unclear. Lastly, the external sphincter and perirectal fat are both echogenic and frequently indistinguishable, precluding accurate characterization of external sphincter thickness and identification of external sphincter atrophy. Evacuation Proctography (Defecography) During evacuation proctography, anorectal anatomy and pelvic floor motion are recorded on video with the patient at rest, coughing, squeezing, and straining to expel barium from the rectum; the anorectal angle, and position of the anorectal junction are tracked during these maneuvers. Before dynamic MRI, evacuation proctography was the only modality for identifying excessive perineal descent, internal rectal intussusception, rectocoeles, sigmoidocoeles or enterocoeles; puborectalis dys-

7 1678 ADIL E.BHARUCHA GASTROENTEROLOGY Vol.124, No.6 Figure 2. Endoanal US (A) and magnetic resonance images (B) of anal sphincters in a patient with fecal incontinence. The internal anal sphincter is hypoechoic on the ultrasound image, whereas on the magnetic resonance image the internal sphincter is of higher signal intensity than the external sphincter. Thick and thin arrows indicate normal internal sphincter and tear, respectively (located approximately between 10 and 5 o clock) on ultrasound and magnetic resonance images. Large and small arrowheads indicate normal appearing and partially torn external sphincter (between 10 and 2 o clock), respectively. function during squeeze and evacuation can also be characterized. 58 Two factors have hampered an assessment of its diagnostic utility in fecal incontinence. The methods for test conduct and interpretation are incompletely standardized 59 ; for incontinent patients, a thick barium paste (Anatrast, E-Z-EM, Westbury, NY) is probably preferable to liquid barium. Because pelvic floor prolapse (e.g., rectocoele) is relatively common in asymptomatic older women, findings must be compared with asymptomatic age- and sex-matched patients; only a handful of studies have characterized pelvic floor motion by evacuation proctography in asymptomatic patients. 60 Despite these limitations, evacuation proctography in incontinent patients may be useful, particularly before surgery, when there is a high index of suspicion for excessive perineal descent, a significant rectocoele (e.g., 2 cm in size, need to splint vagina to facilitate rectal emptying), enterocoele, or internal rectal intussusception. These findings may also help educate patients about the nature of the disorder, and reinforce the need for treatment (e.g., pelvic floor retraining). Pelvic Magnetic Resonance Imaging With the advent of rapid MRI sequences, this imaging modality can visualize both anal sphincter anatomy and global pelvic floor motion in real-time, without radiation exposure. 61 The anal sphincters are visualized by axial T2-weighted fast spin-echo images and corresponding T1-weighted spin-echo images with a disposable endorectal colon coil (Figures 2 and 3). 61 The location and extent of anal sphincter defects visualized by endoanal MRI was confirmed at surgery in all 6 patients with postpartum fecal incontinence in one study though surgical dissection was restricted to portions of the external sphincter which appeared abnormal by MRI. 62 Disagreement exists over the best technique for the Figure 3. Axial endoanal fast spin-echo magnetic resonance image (TR/TE 4000/105) shows a normal puborectalis from an asymptomatic subject (A) and puborectalis tear on the right in a patient with fecal incontinence (B). Arrows indicate normal puborectalis.

8 May 2003 FECAL INCONTINENCE 1679 Figure 4. Magnetic resonance fluoroscopic images of the pelvis (A) at rest, (B) during squeeze, and (C) simulated defecation in a 52-year-old asymptomatic patient after filling the rectum with ultrasound gel. At rest, the pelvic floor was well supported; the anorectal angle was relatively obtuse (126 ). Pelvic floor contraction during the squeeze maneuver was accompanied by normal upward and anterior motion of the anorectal junction; the angle declined to 95. During rectal evacuation, the bladder base dropped by 2.5 cm below pubococcygeal line; the 2.8-cm anterior rectocoele emptied completely and was probably not clinically significant; perineal descent (5 cm) was outside the normal range for evacuation proctography. evaluation of the internal sphincter; however, MRI performed the same or better than US for the assessment of the external sphincter (Figure 2). 63,64 In contrast to US, MRI can identify external sphincter atrophy with a sensitivity of 89% and specificity of 94% compared with histology. 65 External sphincter atrophy is a good prognosticator for poor continence after repair of external sphincter defects. 66 Dynamic images are acquired as patients squeeze their sphincters or try and expel US gel from the rectum, providing a unique appreciation of global pelvic floor motion, i.e., in addition to the anorectum, the bladder and genital organs are also visualized (Figures 4 and 5). The examination can be performed using conventional, closed-configuration MRI systems because there is little difference in the detection of clinically relevant findings between supine MRI and seated MRI using open-configuration magnets. The exception is in detecting rectal intussusceptions for which seated MRI was superior. 67 Preliminary studies indicate that dynamic MRI was useful for accurately characterizing perineal descent, puborectalis dyssynergia, and pelvic organ prolapse, i.e., rectocoele, cystocoele, enterocoele, and uterine or vaginal prolapse. 61,68 As with any new diagnostic imaging modality, comparisons to age- and gender-matched asymptomatic patients are necessary to assess the role of static and dynamic MRI as a diagnostic tool in clinical practice. Pudendal Nerve Terminal Motor Latency Delayed pudendal nerve terminal motor latency is used as surrogate marker of pudendal nerve injury and to ascertain whether anal sphincter weakness is attributable to pudendal nerve injury, sphincter defect, or both. 59 The index finger covered by a glove containing stimulating and recording electrodes stimulates the pudendal nerve as it courses around the pelvic brim. Measuring the Figure 5. Pelvic magnetic resonance fluoroscopic images at rest (A) and squeeze (B) ina 57-year-old female patient who has fecal incontinence. During squeeze, the puborectalis indentation on the posterior rectal wall was exaggerated compared to rest, and the anorectal angle declined from 143 at rest to 90 during squeeze; however, the anal canal remained patulous.

9 1680 ADIL E.BHARUCHA GASTROENTEROLOGY Vol.124, No.6 shortest latency between stimulus delivery and recording is critically dependent on close proximity of the examining finger to the nerve. In theory, delayed latencies indicate pudendal neuropathy; initial studies suggested that patients with a pudendal neuropathy would not fare as well after surgical repair of sphincter defects as patients without a neuropathy. However, the utility of measuring pudendal nerve terminal motor latency is questionable because it measures only the fastest conducting fibers in the nerve and there are inadequate normative data; test reproducibility is unknown, and sensitivity and specificity are poor. In contrast to initial studies, recent studies suggest the test does not predict improvement, or lack thereof, after surgical repair of anal sphincter defects. 69 Needle Electromyography of the External Sphincter Endoanal US has supplanted anal sphincter electromyography (EMG) for identifying sphincter defects. With the recognition that pudendal nerve latencies are flawed as markers of pudendal neuropathy, sphincter EMG provides a sensitive measure of denervation (fibrillation potentials) and can usually identify myopathic (small polyphasic motor unit potentials), neurogenic (large polyphasic motor unit potentials), or mixed injury. 70,71 In experienced hands, anal sphincter EMG is not accompanied by severe discomfort. Presently, anal sphincter EMG should be conducted if a proximal neurogenic process, (e.g., in the spinal cord or sacral roots) is suspected on clinical grounds. The use of EMG as a prognosticator of success after repair of sphincter defects deserves further study. Rectoanal Sensation Rectal sensation is assessed by progressively distending a latex balloon manually, or, a polyethylene balloon by a barostat while measuring thresholds for first perception, desire to defecate, and severe discomfort. 18 Alternatively, perception is recorded by a visual analogue scale during phasic distentions of graded intensity by a barostat. 72 Within limits, rectal sensation can be modulated by biofeedback therapy, improving continence. Anal sensation is assessed by determining the perception threshold to an electrical stimulus, or temperature change in the anal canal. The threshold for detecting temperature change is assessed while the temperature of a thermode within the anal canal is varied between normal (37 C), hot (42 C), and cold (32 C) temperatures. 73,74 Electrosensitivity is nonphysiological and does not activate mucosal receptors. 59,75 Anal sensitivity to temperature change is reduced in fecal incontinence. 74 Rectal Compliance Compliance is a measure of rectal distensibility, reflecting rectal ability to hold a larger volume of stool at a given pressure, thereby postponing defecation to a convenient time 20 ; the rectum becomes less compliant with aging, perhaps predisposing to incontinence. 76 Compliance is measured by assessing rectal pressurevolume relationships with a balloon; either by manually inflating a latex balloon with air or water, or more accurately by distending a highly compliant polyethylene balloon with a barostat. With the latter approach, pressure-volume or compliance curves in healthy patients are extremely reproducible within patients on the same day, 77 and on separate days after an initial or conditioning distention. 78 The contribution of abnormal rectal compliance to rectal sensory disturbances in idiopathic fecal incontinence is currently being investigated. Recommendations Endoscopy to identify mucosal pathology is probably necessary for incontinent patients who have significant, particularly recent-onset diarrhea or constipation; the extent of examination (i.e., sigmoidoscopy or colonoscopy) is also guided by the patient s age and comorbidities. The indications for, and extent of diagnostic testing in incontinent patients are evolving. For ambulatory, otherwise healthy patients, reviewing anorectal manometry and endoanal US images is necessary to document severity of weakness and identify abnormal sphincter morphology. Evacuation proctography may be useful to document clinically suspected, but not overt excessive perineal descent or pelvic organ prolapse. Pelvic MRI combines static and dynamic imaging without radiation exposure, is particularly useful for identifying external sphincter atrophy and pelvic organ prolapse, but is relatively expensive, and not widely available. Anal sphincter EMG should be considered for incontinent patients who have an underlying disease associated with a neuropathy (e.g., diabetes mellitus), clinical suspicion of a proximal neurogenic process, or sphincter weakness unexplained by morphology as visualized by US. Pudendal nerve latencies are not useful. These advanced tests (e.g., pelvic MRI, anal sphincter EMG, assessment of rectal compliance, and sensation) are not widely available. Therefore, consideration should be given to referring independently living patients with moderate-severe fecal incontinence to specialist centers for further assessment under the following circumstances: (1) when symptoms cannot be explained by routine diagnostic tests (e.g., when anal sphincter weakness is mild); (2) before considering repair sphincter of

10 May 2003 FECAL INCONTINENCE 1681 Table 6. Management of Fecal Incontinence Intervention Side effects Comments Mechanism of action Incontinence pads a 108 Disposable bodyworns, reusable bodyworns, disposable underpads, and reusable underpads. (Disposable bodyworns are the largest category) Antidiarrheal agents a Loperamide (Imodium) up to 16 mg/day in divided doses Diphenoxylate -5mgqid Enemas b Biofeedback therapy using anal canal pressure or surface EMG sensors b 109 Rectal balloon for modulating sensation Sphincteroplasty for sphincter defects b 109 Skin irritation Constipation Inconvenient; side effects of specific preparations Wound infection; recurrent incontinence (delayed) Disposable products provide superior skin protection than nondisposable products; underpad products were slightly cheaper than bodyworn products Titrate dose for each patient; administer before meals and social events Prerequisites for success include motivation, intact cognition, absence of depression, and some rectal sensation Restricted to isolated sphincter defects without denervation Provide skin protection and prevent soiling of linen; polymers conduct moisture away from the skin Increased fecal consistency decreased urgency; increased anal sphincter tone Rectal evacuation decreases likelihood of fecal incontinence Improved rectal sensation and coordinated external sphincter contraction, possible increased anal sphincter tone; restore sphincter integrity Sacral nerve stimulation b Infection Preliminary uncontrolled trials promising Unclear; increased anal sphincter tone; may modulate rectal sensation Artificial sphincter Gracilis transposition b Device erosion, failure and infection Either artificial device or gracilis transposition with/ without electrical stimulation Restore anal barrier NOTE. Grades A or B are supported by at least 1 randomized controlled trial, or 1 high-quality study of nonrandomized cohorts. a Grade A and b Grade B therapeutic recommendations. Adapted from the chapter on GI dysmotility and sphincter dysfunction in Neurological Therapeutics: Principles and Practice, John H. Noseworthy, ed. London: Martin Dunitz, Ltd. (in press). external sphincter defects in older women, (since surgical repair is not always successful, careful consideration of other factors contributing to incontinence, including pelvic MRI to identify external sphincter atrophy may be useful); (3) for patients who have symptoms of an evacuation disorder and fecal incontinence, biofeedback therapy is more rigorous and specialized than for patients who only have fecal incontinence; and (4) for patients who have combined fecal and urge urinary incontinence, sacral nerve stimulation is an FDA-approved procedure for urge urinary incontinence; bowel symptoms may also respond to sacral nerve stimulation. Management Management of fecal incontinence must be tailored to clinical manifestations, and includes treatment of underlying diseases, and other approaches detailed in Table 6. Bowel Habit Modification Modifying irregular bowel habits is often the cornerstone to effectively managing incontinence. Several seemingly simple approaches are underused because of a perceived lack of benefit. Accurate characterization of bowel habits is necessary to tailor therapy. In addition to reducing diarrhea, loperamide, given at an adequate dose (i.e., 2 to 4 mg, 30 minutes before meals, up to 16 mg/day) slightly increased internal sphincter tone, thereby reducing incontinence. 79 In patients who have constipation and diarrhea, effective dose titration to reduce diarrhea, but avoid constipation can be challenging. By taking loperamide before social occasions, or meals outside the home, incontinent patients may avoid having an accident outside the home, and gain confidence in their ability to participate in social activities. Diphenoxylate is an alternative option for diarrhea 80 ; the serotonin 5-HT 3 antagonist alosetron (Lotronex; Glaxo SmithKline, Research Traingle Park, NC), currently

11 1682 ADIL E.BHARUCHA GASTROENTEROLOGY Vol.124, No.6 available under a restricted use program, may benefit patients who have severe diarrhea and whose symptoms do not respond to other agents. Patients who have constipation, fecal impaction, and overflow incontinence may benefit from a regularized evacuation program, incorporating timed evacuation by digital stimulation and/or bisacodyl/glycerol suppositories, fiber supplementation, and selective use of oral laxatives as detailed in a recent review. 81,82 Pharmacological Approaches Phenylephrine, an 1 -adrenergic agonist applied to the anal canal increased anal resting pressure by 33% in healthy patients and incontinent patients. 83 However, phenylephrine did not significantly improve incontinence scores or resting anal pressure compared with placebo in a randomized double-blind placebo-controlled crossover study of 36 patients with fecal incontinence. 84 Biofeedback Therapy Biofeedback therapy is based on the principle of operant conditioning. 85 Using a rectal balloon-anal manometry device, patients are taught to contract the external anal sphincter when they perceive balloon distention; perception may be reinforced by visual tracings of balloon volume and anal pressure, and the procedure is repeated with progressively smaller volumes. Several uncontrolled studies suggest continence improves in approximately 70% of patients 86 ; controlled studies are in progress. Although resting and squeeze pressures increased to a variable degree after biofeedback therapy, the magnitude of improvement was relatively small and not correlated to symptom improvement; perhaps these modest effects are attributable to inadequate biofeedback therapy, lack of reinforcement, and assessment of objective parameters at an early stage after biofeedback therapy. 87 In contrast, sensory assessments (i.e., preserved baseline sensation) and improved sensory discrimination after biofeedback therapy are more likely to be associated with improved continence after biofeedback therapy. 88,89 A recent study randomized 171 incontinent patients to 4 groups: standard medical/nursing care (i.e., advice only), advice plus verbal instruction on sphincter exercises, hospital-based computer-assisted sphincter pressure biofeedback, or hospital biofeedback plus use of a home EMG biofeedback device. 90 Overall, 75% reported improved symptoms and 5% were cured. Improvement was sustained at 1 year after therapy; symptoms, resting and squeeze pressures improved to a similar degree in all four groups. These results emphasize the importance patients attach to understanding the condition, practical advice regarding coping strategies (e.g., diet and skin care), and nurse-patient interaction. Surgical Approaches Continence improved in up to 85% of patients with sphincter defects after an overlapping anterior sphincteroplasty. For reasons that are unclear, continence deteriorates thereafter; only approximately 50% of patients remain continent up to 5 years after the operation. 91,92 Dynamic graciloplasty involves continuous electrical stimulation of the gracilis muscle transposed around the anal canal; electrical stimulation facilitates anal tone by converting type II (fast switch, fatigueprone) to type I (slow-twitch, fatigue-resistant) muscle fibers. The hardware for dynamic graciloplasty is not approved for use in the United States. The cumulative experience of 383 patients who have undergone dynamic graciloplasty was summarized in a recent review. 93 The procedure is complicated with a mortality of approximately 2% and significant morbidity, including infections (28%), device problems (15%), and leg pain (13%), which may require more surgery. Continence was significantly improved in approximately 60% of implanted patients. The experience with the artificial anal sphincter is similar; 99 of 112 implanted patients from a cumulative multicenter report from the United States had device-related adverse effects. 94 Fifty-one patients required 73 operations to revise the device, and the device was completely explanted in 41 patients (37%). Fecal continence scores and quality of life improved; the outcome was successful in 85% of patients with a functioning device, and the intention to treat success rate was 53%. A colostomy is the last resort for patients with severe incontinence; the impact of a colostomy on quality of life in fecal incontinence has not been formally assessed. 93 Minimally Invasive Approaches Sacral nerve stimulation is an FDA-approved device implanted in more than 3000 patients who have urinary incontinence in the United States. Observations from European studies suggest that sacral nerve stimulation augmented squeeze pressure more than resting pressure, and significantly improved continence Sacral stimulation may also modulate rectal sensation in incontinent patients. 97 Sacral stimulation is conducted as a staged procedure; patients whose symptoms respond to temporary stimulation over approximately 2 weeks proceed to permanent subcutaneous implantation of the device. The procedure for device placement is technically straightforward, and device-related complications are less frequent or significant relative to more invasive artificial sphincter devices discussed above. A multicenter study

12 May 2003 FECAL INCONTINENCE 1683 from the United States assessing sacral nerve stimulation for fecal incontinence is in progress. A small nonrandomized study observed improved continence, quality of life, and lower sensory threshold volumes, but no changes in anal sphincter pressures after applying radiofrequency energy to the anal canal in 10 women who had fecal incontinence. 98 Further studies of safety and efficacy are awaited. Summary Fecal incontinence is a relatively prevalent, multifactorial and multifaceted disorder with significant psychosocial consequences. In patients who do not have an underlying disorder associated with incontinence, anal sphincter trauma caused by obstetric or surgical injury, chronic straining, diarrhea, and aging have been implicated. The relative contributions of obstetric and nonobstetric risk factors and the specific obstetric risk factors for pelvic floor injury must be clarified. The clinical assessment and diagnostic testing are directed toward assessing the cause of the symptom, likely anatomical and physiological disturbances, and impact on quality of life. Treatment should be tailored to patient s symptoms; it is unclear which components of biofeedback therapy contribute to its favorable effects. Surgical therapies should be restricted to carefully selected patients; artificial sphincter procedures and dynamic graciloplasty are associated with significant morbidity. Newer, less invasive therapeutic approaches are being evaluated. It is conceivable that an improved understanding of risk factors will lead to measures to prevent pelvic floor injury in future. References 1. Whitehead WE, Wald A, Norton NJ. Treatment options for fecal incontinence. Dis Colon Rectum 2001;44: ; discussion Crowell M, Schettler-Duncan V, Brookhart K, Barofsky I. Fecal incontinence: impact on psychosocial function and health-related quality of life. Gastroenterology 1998;114:A738 (abstr). 3. O Keefe EA, Talley NJ, Zinsmeister AR, Jacobsen SJ. Bowel disorders impair functional status and quality of life in the elderly: a population-based study. Journals of Gerontology Series A, Biological Sciences and Medical Sciences 1995;50: M184 M Rockwood TH, Church JM, Fleshman JW. Fecal incontinence quality of life scale: quality of life instrument for patients with fecal incontinence. Dis Colon Rectum 2000;43:9 16 discussion Perry S, Shaw C, McGrother C. Prevalence of faecal incontinence in adults aged 40 years or more living in the community. Gut 2002;50: Nelson R, Furner S, Jesudason V. Fecal incontinence in Wisconsin nursing homes: prevalence and associations. Dis Colon Rectum 1998;41: Nelson R, Norton N, Cautley E, Furner S. Community-based prevalence of anal incontinence. JAMA 1995;274: Frenckner B, Ihre T. Influence of autonomic nerves on the internal and sphincter in man. Gut 1976;17: Bartolo DC, Jarratt JA, Read MG, Donnelly TC, Read NW. The role of partial denervation of the puborectalis in idiopathic faecal incontinence. Br J Surg 1983;70: Fernandez-Fraga X, Azpiroz F, Malagelada JR. Significance of pelvic floor muscles in anal incontinence. Gastroenterology 2002;123: Bartolo DC, Read NW, Jarratt JA, Read MG, Donnelly TC, Johnson AG. Differences in anal sphincter function and clinical presentation in patients with pelvic floor descent. Gastroenterology 1983;85: Harewood GC, Coulie B, Camilleri M, Rath-Harvey D, Pemberton JH. Descending perineum syndrome: audit of clinical and laboratory features and outcome of pelvic floor retraining. Am J Gastroenterol 1999;94: Bassotti G, Crowell MD, Whitehead WE. Contractile activity of the human colon: lessons from 24 hour studies. Gut 1993;34: Sun WM, Read NW, Miner PB. Relation between rectal sensation and anal function in normal subjects and patients with faecal incontinence. Gut 1990;31: Sun WM, Donnelly TC, Read NW. Utility of a combined test of anorectal manometry, electromyography, and sensation in determining the mechanism of idiopathic faecal incontinence. Gut 1992;33: Allen ML, Orr WC, Robinson MG. Anorectal functioning in fecal incontinence. Dig Dis Sci 1988;33: Buser WD, Miner PB Jr. Delayed rectal sensation with fecal incontinence. Successful treatment using anorectal manometry. Gastroenterology 1986;91: Whitehead WE, Palsson OS. Is rectal pain sensitivity a biological marker for irritable bowel syndrome? Psychological influences on pain perception. Gastroenterology 1998;115: Miller R, Bartolo DC, Cervero F, Mortensen NJ. Anorectal sampling: a comparison of normal and incontinent patients. Br J Surg 1988;75: Devroede G. Functions of the anorectum: defecation and continence. In: Phillips S, Pemberton J, Shorter R, eds. The large intestine: physiology and disease. New York: Raven, 1991: Samuelsson E, Ladfors L, Wennerholm UB, Gareberg B, Nyberg K, Hagberg H. Anal sphincter tears: prospective study of obstetric risk factors. BJOG: Int J Obstet Gynaecol 2000;107: Sultan AH, Kamm MA, Hudson CN, Thomas JM, Bartram CI. Anal-sphincter disruption during vaginal delivery [see comments]. N Engl J Med 1993;329: Law PJ, Kamm MA, Bartram CI. Anal endosonography in the investigation of faecal incontinence. Br J Surg 1991;78: Williams AB, Bartram CI, Halligan S, Spencer JA, Nicholls RJ, Kmiot WA. Anal sphincter damage after vaginal delivery using three-dimensional endosonography. Obstetr Gynecol 2001;97: Varma A, Gunn J, Gardiner A, Lindow SW, Duthie GS. Obstetric anal sphincter injury: prospective evaluation of incidence. Dis Colon Rectum 1999;42: Abramowitz L, Sobhani I, Ganansia R. Are sphincter defects the cause of anal incontinence after vaginal delivery? Results of a prospective study. Dis Colon Rectum 2000;43: ; discussion Chaliha C, Sultan AH, Bland JM, Monga AK, Stanton SL. Anal function: effect of pregnancy and delivery. Am J Obstet Gynecol 2001;185: Reilly W, Talley N, Pemberton J. Fecal incontinence: prevalence and risk factors in the community. Gastroenterology 1995;108: A32 (abstr). 29. Chen H, Humphreys MS, Kettlewell MG, Bulkley GB, Mortensen N, George BD. Anal ultrasound predicts the response to nonop-