Chapter  2:  Urinary Incontinence in Adults: Acute and Chronic Management: Clinical Practice Guideline Number 2 (1996 Update)

Functional Incontinence

Urine loss may be caused by factors outside the lower urinary tract such as chronic impairment of physical or cognitive functioning, or both, a condition commonly termed "functional incontinence" Williams and Pannill, 1982). This diagnosis should be one of exclusion, however, because some immobile and cognitively impaired individuals have other types and causes of UI that may respond to specific therapies (Skelly and Flint, 1995). For example, most nursing home residents have not only functional limitations that interfere with their ability to toilet but also urge UI with detrusor instability or DH. In addition, UI can often be improved or "cured" by improving the patient's functional status, treating other medical conditions, discontinuing certain types of medication, adjusting the hydration status, reducing environmental barriers, or all of the above -- even if a lower urinary tract abnormality is present.

Unconscious or Reflex Incontinence

Urine loss may also occur without any warning or sensory awareness such as in paraplegics and some patients without overt neurologic dysfunction. This condition, which also has been termed unconscious incontinence, presents as either postmicturitional or constant (continuous) dribbling. Extraurethral UI (i.e., fistulas), UI in children, and neuropathic conditions are not specifically addressed in this guideline.

Another urodynamic finding that may be associated with UI is decreased bladder compliance. This abnormal bladder condition may result from radiation cystitis, inflammatory bladder conditions such as chemical cystitis or interstitial cystitis, and some neurologic bladder disorders such as those that occur following radical pelvic surgery and in myelomeningocele. Many of the patients with a nonneurogenic etiology for their decreased bladder compliance (e.g., chemical or radiation cystitis) have severe urgency associated with bladder hypersensitivity and with no demonstrable DI. This is termed "sensory urgency." Both loss of bladder wall elasticity and lack of bladder accommodation produce a steep rise in intravesical pressure during bladder filling without detrusor contraction. In patients in whom the urethral sphincter mechanism may already be compromised by the condition that produced the inelastic bladder (e.g., radiation, neurologic dysfunction), the abnormal increase in bladder pressure may overcome the urethral pressure and produce UI. A major concern in patients with a poorly compliant bladder, especially those who are neurologically impaired, is the development of vesicoureteral reflux and hydronephrosis.

History

• Duration and characteristics of UI (stress, urge, dribbling, others).

• Most bothersome symptom(s) to the patient (which may be especially important in guiding therapy and determining response).

• Frequency, timing, and amount of continent voids and incontinent episodes.

• Precipitants of incontinence (e.g., situational antecedents, cough, certain types of exercises, surgery, injury, previous pelvic radiation therapy, trauma, new onset of diseases, new medications).

• Other lower urinary tract symptoms (e.g., nocturia, dysuria, hesitancy, poor or interrupted stream, straining, hematuria, suprapubic or perineal pain).

• Fluid intake pattern, including caffeine-containing or other diuretic fluids.

• Alterations in bowel habits or sexual function.

• Previous treatments and their effects on UI.

• Amount and types of pads, briefs, and protective devices.

• Expectations for outcomes of treatment.

• A bladder record.

• A mental status evaluation and assessment of mobility, living environment, and social factors, especially in elderly patients.

Bladder records are helpful supplements to the history in many patients (Diokno, Wells, and Brink, 1987; Larson and Victor, 1992; Wyman, Choi, Harkins, et al., 1988). These written records are used to determine the frequency, timing, and amount of voiding, as well as other factors associated with UI. Such a record can be kept by the patient or a caregiver for a few days before the basic evaluation. The record may provide clues about the underlying cause of UI, as well as information on fluid intake and voiding patterns that may be helpful for behavioral interventions, and can serve as a baseline to gauge severity and treatment efficacy (see sample records, Attachments B and C.

A mental status examination should be performed when appropriate. Assessment of mobility and living environment is also especially important in certain individuals. Questions should be asked about access to toilets or toilet substitutes, and about social factors such as living arrangements, social contacts, or caregiver involvement (Williams and Gaylord, 1990).

Physical Examination

The physical examination should include the following:

• General examination if indicated to detect conditions such as edema that may contribute to nocturia and nocturnal UI; to detect neurologic abnormalities that may suggest multiple sclerosis, stroke, spinal cord compression, or other neurologic conditions; and to assess mobility, cognition, and manual dexterity related to toileting skills among frail and functionally impaired patients Williams and Gaylord, 1990).

• Abdominal examination to check for diastasis rectii, organomegaly, masses, peritonitis, fluid collections, and so on. Abnormality of abdominal contents may influence intraabdominal pressure and detrusor physiology.

• Rectal examination to test for perineal sensation, sphincter tone (both resting and active), fecal impaction, or rectal mass, and to evaluate the consistency and contour of the prostate in men. The size of the prostate on digital examination does not exclude or imply obstruction and is usually helpful only for the surgeon in determining the surgical approach if an operation is contemplated.

• Genital examination in men to evaluate skin condition and detect abnormalities of the foreskin, glans penis, and perineal skin.

• Pelvic examination in women to assess perineal skin condition, genital atrophy, pelvic organ prolapse (cystocele, rectocele, uterine prolapse), pelvic mass, paravaginal muscle tone, or other abnormalities. Palpation of the anterior vaginal wall and urethra may elicit urethral discharge or tenderness that suggests a urethral diverticulum, carcinoma, or inflammatory condition of the urethra. Pelvic organ prolapse may not relate to urinary symptoms, especially in the elderly (Ouslander, Hepps, Raz, et al., 1986).

• Direct observation of urine loss using the cough stress test. Observation of urine loss can be performed by having the individual cough vigorously while the examiner observes for urine loss from the urethra. Optimally, testing should be done when the patient's bladder is full but before the patient has a precipitant urge to void Kadar, 1988). In women who are being evaluated for specific treatments for stress incontinence, this test is important for objective demonstration of urine loss and identification of provoking factors. In other patients, particularly those with physical or cognitive impairment, direct observation may be difficult to carry out, and this documentation may not be critical in determining initial treatment of incontinence. If an instantaneous leakage occurs with cough, then SUI is likely; if leakage is delayed or persists after the cough, DI should be suspected. If the test is initially performed in the lithotomy position and no leakage is observed, the test should be repeated in the upright position (Kadar, 1988). If bladder filling is needed to perform stress testing, it may be conveniently performed in conjunction with the catheterization method of measurement of PVR volume.

Estimation of PVR Volume

Accurate measurement of PVR can be accomplished either by catheterization or by pelvic ultrasound (Coombes and Millard, 1994; Haylen, Frazer, and MacDonald, 1989; Ireton, Krieger, Cardenas, et al., 1990). Some health care providers may use abdominal palpation or bimanual pelvic examination to suspect elevated PVR; however, the exact amount of volume or confirmation usually requires other methods (catheterization, ultrasound) (Norton, Peattie, and Stanton, 1989).

Before PVR is measured, the patient should void in the most comfortable and private environment possible. Voiding can be observed at this time to detect signs of hesitancy, straining, or slow or interrupted stream, which may indicate urethral obstruction, a bladder contractility problem, or both.

Measurement of PVR volume is generally done within a few minutes after voiding by catheterizing the patient or by pelvic ultrasound. Review of the literature failed to indicate a specific maximum PVR volume considered normal, nor is there any documentation of the minimal PVR considered abnormal. In general, PVRs of less than 50 mL are considered adequate bladder emptying. Repetitive PVRs ranging from 100 to 200 or higher are considered inadequate emptying. Clinical judgment must be exercised and all other clinical information included in interpreting the significance of PVR volume, especially in the intermediate range of 50-199 mL.

The PVR urine may be influenced by such factors as the volume voided before PVR measurement, whether the patient is "ready" to void or strains to void, the efforts made to drain the bladder completely, and the environment or clinical setting. Because PVR volume may vary, one measurement of PVR may not be sufficient and repeated measurements may be of value in some patients.

Urethral catheterization in men with prostate obstruction may cause urinary tract infection. Therefore, catheterization in men should be performed only with clear indication and by a health care provider who is prepared to manage abnormal findings.

An alternative to catheterization is the use of pelvic ultrasound. Portable ultrasound devices that estimate PVR with reasonable accuracy are now available (Ireton, Krieger, Cardenas, et al., 1990; Ouslander, Simmons, Tuico, et al., 1994; Topper, Holliday, and Fernie, 1993).

Urinalysis

This test is used to detect conditions that are associated with or contributing to UI such as hematuria (suggestive of infection, cancer, or stone), glucosuria (which may cause polyuria and contribute to UI symptoms), pyuria, and bacteriuria, as well as glycosuria and proteinuria. If catheterization is performed for PVR measurement, a sample of the residual urine can be used for the urinalysis and microscopic examination. Careful cleaning of the glans penis in men and the periurethral area in women with an antiseptic solution allows collection of a spontaneously voided specimen that accurately reflects bladder urine.

Dipstick methods are available to detect bacteriuria and pyuria. The diagnostic accuracy of these methods varies considerably among methods and different patient populations (Lachs, Nachamkin, Edelstein, et al., 1992). Therefore, urine cultures should be obtained in incontinent patients when dipstick tests indicate infection or when symptoms suggest infection.

Among chronically incontinent nursing home residents, eradication of bacteriuria (with or without pyuria) has no effect on morbidity, mortality, or the severity of UI (Boscia, Kobasa, Abrutyn, et al., 1986). Thus, unless UI is of recent onset, has recently worsened, or is accompanied by other symptoms of infection (e.g., dysuria, hematuria, fever, sudden decline in mental or physical functioning), bacteriuria (with or without pyuria) does not need to be treated in this patient population. Among noninstitutionalized patients, the relationship of bacteriuria (with or without pyuria) to UI is unclear. Further research is needed to clarify this relationship. Until such data are available, infection should be treated when the incontinent patient is initially evaluated and the effect observed before further diagnostic or therapeutic interventions are undertaken (Ouslander, 1989).

Supplementary Assessments

The use of blood tests (BUN, creatinine, glucose, and calcium) and urine cytology in the assessment of the incontinent patient was also evaluated.

Blood tests that may be helpful in the basic evaluation of the incontinent patient include creatinine levels in patients suspected of having obstruction, noncompliant bladders, or urinary retention. However, normal creatinine levels do not rule out hydronephrosis. Patients with polyuria in the absence of diuretic agents should be evaluated for excess intake (diabetes insipidus), hyperglycemia, and hypercalcemia if indicated.

There is no evidence in the literature to support urine cytology testing in patients with UI. Individuals with incontinence are at no higher risk of bladder cancer than the continent population Mohr, Offord, Owen, et al., 1986 ). Urine cytology is a poor screening test for bladder cancer, with sensitivities ranging from approximately 20-70 percent (Curling, Broome, and Gendry, 1986; Murphy, 1990; Rife, Farrow, and Utz, 1979). Patients with hematuria (2-5 RBC/hpf) or acute onset of irritative voiding symptoms in the absence of urinary tract infection (UTI) require cystoscopy and cytology to exclude bladder neoplasm (Utz and Zincke, 1974).

Urodynamic Tests

These tests are designed to determine the anatomic and functional status of the urinary bladder and urethra. The tests are performed by qualified professionals trained in the specific definitions and procedures.

Cystometry is a test of detrusor function. Depending on the technique used, cystometry can be used to assess bladder sensation, capacity, and compliance, and to determine the presence and magnitude of both voluntary and involuntary detrusor contractions. The patient's symptoms should be reproduced at the time of cystometry, because involuntary DI may be observed in asymptomatic patients and symptomatic patients may show no evidence of involuntary ventrusial contractions. On the other hand, cystometry may be falsely negative in a patient with a genuinely overactive bladder (McGuire, 1994). This is caused by heightened psychological inhibition of reflex activity or lack of measurable increase of detrusor pressure, urethral pressure, or electromyographic activity, which may be dissipated by poor urethral resistance and must therefore be examined closely.

Simple cystometry can be used to detect detrusor contractions and abnormalities of bladder compliance, and to measure PVR and determine bladder capacity (Ouslander, Leach, and Staskin, 1989; Sand, Brubaker, and Novak, 1991). This procedure is performed by filling the bladder to capacity by means of a urethral catheter. Because a rectal or vaginal catheter is not used to monitor the abdominal pressure, results must be interpreted with caution, especially in uncooperative or demented patients. Compared with multichannel studies, simple cystometry has a reported 75-100 percent sensitivity, 69-89 percent specificity, and 74-91 percent positive predictive value for the diagnosis of DI (Fonda, Brimage, and D'Astoli, 1993; Sand, Brubaker, and Novak, 1991; Sutherst and Brown, 1984).

A multichannel or subtracted cystometrogram (CMG) with simultaneous measurement of intra-abdominal, total bladder, and true detrusor pressures can differentiate an involuntary detrusor contraction or deversed bladder compliance from an increase in intra-abdominal pressure. However, few data are available comparing multichannel cystometrics with other measures of intravesical pressure, and it is suspected that false positives may be a problem, especially in elderly populations (Diokno, Brown, Brock, et al., 1988). Likewise, ambulatory continuous monitoring may reveal detrusor contractions missed during a provocative subtracted CMG (McGuire, 1994).

A voiding CMG or pressure flow study can measure detrusor contractility and detect outlet obstruction if the patient is able to void. Simultaneous measurement of detrusor and urethral pressures during voiding is especially helpful in diagnosing urodynamic obstruction (Diokno, 1990).

Uroflowmetry measures the urine flow rate visually, electronically, or with the use of a disposable unit. An electronically generated flow curve is considered helpful in identifying abnormal voiding patterns. Uroflowmetry is not helpful in diagnosing the types of incontinence found in women (Diokno, Normolle, Brown, et al., 1990; Fantl, Smith, Schneider, et al., 1983), but it may be helpful in patients who have difficulty with bladder emptying. However, this test cannot distinguish between obstruction and detrusor weakness without a simultaneous measurement of detrusor function.

Urethral pressure profilometry (UPP). The urethral function test measures resting and dynamic pressures in the urethra. Passive measurement of urethral pressures has been used by some investigators to help identify ISD, especially in women who have had previous operations. However, no specific measurement to date has been found to be discriminatory, especially in the elderly because of the normal decline of urethral pressures with age (Diokno, Normolle, Brown, et al., 1990). Dynamic measurements of urethral and bladder pressures may be used to measure the effect of exertion on the urethral closure mechanism. Sphincter function can also be assessed by measuring the abdominal or vesical pressure needed to overcome the urethral resistance (McGuire, Fitzpatrick, Wan, et al., 1993; McGuire, Woodside, Borden, et al., 1981; Wan, McGuire, Bloom, et al., 1993). The relative usefulness of the UPP versus that of Abdominal Leak Point Pressure (ALPP) has not been adequately studied.

Electromyography (EMG) of the striated urethral sphincter measures the integrity and function of its innervation. Both needle and surface EMG, in conjunction with CMG, may be helpful in diagnosing DSD Diokno, Koff, and Anderson, 1976 ). However, this condition should be diagnosed only after common artifacts such as volitional tightening of the sphincter during involuntary detrusor contraction are excluded.

Endoscopic Tests

A literature search revealed that in studies involving more than 600 incontinent patients selectively cystoscoped, only 11 (less than 2 percent) metaplastic or neoplastic lesions were identified (Awad, Gajewski, Katz, et al., 1992; Benson, 1985; Fischer-Rasmussen, Hansen, and Stage, 1986; Ouslander, Hepps, Raz, et al., 1986; Ouslander, Leach, Staskin, et al., 1989; Vehkalahti, Kivela, and Seppanen, 1986). Even in elderly, incontinent populations, the yield from selected cystoscopy is less than 1 percent Ouslander, Hepps, Raz, et al., 1986). Therefore, cystoscopy should not be performed routinely in incontinent patients to exclude neoplasm.

The precise role of cystoscopy in the evaluation of incontinent patients is controversial. Data are scarce in the literature on this issue, and studies have a referral bias. Furthermore, urethroscopy is not as useful as urodynamic tests in diagnosing SUI Scotti and Ostergard, 1993). Because this test is not performed during voiding, its usefulness in corroborating or excluding outlet obstruction is limited. Given the large numbers of patients with UI, and given the fact that the majority of UI patients are currently not cystoscoped, cystoscopy cannot be recommended as part of the routine evaluation until data are available to support its performance (Ouslander, Leach, and Staskin, 1989; Fischer-Rasmussen, Hansen, and Stage 1986).

Imaging Tests

Upper tract imaging is not a routine test to evaluate UI. Ultrasound of the kidneys, bladder, or both can help identify dilation of the upper urinary tract and renal pathology, especially in patients with urinary retention, abnormal renal function, or poorly compliant bladders. Excretory urography (i.e., intravenous pyelography) or other imaging modalities are indicated for incontinent patients with sterile hematuria or for further evaluation of upper tract obstruction or other pathology identified by ultrasound (Kumar and Schreib, 1985).

Lower tract imaging with and without voiding is helpful in examining the anatomy of the urinary bladder and urethra. Nonvoiding lateral cystourethrography in the resting and straining view can identify mobility or fixation of the bladder neck, funneling of the bladder neck and proximal urethra, and degree of cystocele. The voiding component can identify a urethral diverticulum, obstruction, and vesicoureteral reflux. Ultrasonography for assessing the dynamics of the bladder neck and urethra is still under investigation, and its clinical utility remains to be confirmed (Benson and Sumners, 1990; Mouritsen, Standberg, Jensen, et al., 1993; Quinn, Beynon, Mortensen, et al., 1988; Richmond and Sutherst, 1989b).

Videourodynamics is a technique that combines the various urodynamic tests with simultaneous fluoroscopy. The technique is helpful in sorting out causes of complex incontinence problems.

Pelvic Muscle Exercise

PMEs, also called Kegel exercises and pelvic floor exercises, are performed to strengthen the voluntary periurethral and perivaginal muscles (i.e., voluntary urinary sphincters and levator ani) that contribute to the closing force of the urethra and to the support of the pelvic visceral structures.

The first step in pelvic muscle re-education is to establish better awareness of pelvic muscle function. PMEs are performed by "drawing in" or "lifting up" of the perivaginal muscles and anal sphincter as if to control urination or defecation with minimal contraction of abdominal, buttock, or inner thigh muscles (Rose, Baigis-Smith, Smith, et al., 1990). Patients are generally told to sustain a contraction for at least 10 seconds, followed by an equal period of relaxation. The exercises should be performed about 30-80 times a day for at least 8 weeks and may need to be continued indefinitely. Elderly patients may require a longer time to train. In general, an individualized program of exercises and repetitions should be tailored to enhance muscle strength progressively (Ferguson, McKey, Bishop, et al., 1990). To condition the muscle to contract with increases in intra-abdominal pressure, patients should be taught to contract the pelvic muscles before and during situations when leakage may occur.

The specific effects of PMEs on actual lower urinary muscle function is not completely understood; some studies show a relationship between changes in various measures of pelvic floor strength, such as anal sphincter strength or maximum urethral closure pressure, and reduction in incontinence (Benvenuti, Caputo, Bandinelli, et al., 1987; Bo, Hagen, Kvarstein, et al.,1990; Ferguson, McKey, Bishop, et al., 1990; Dougherty, Bishop, Mooney, et al., 1993).

PMEs are indicated for women with stress incontinence and can reduce urgency and prevent urge UI ( Burgio and Engel, 1990; Burgio, Robinson, and Engel, 1986; Burgio, Stutzman, and Engel, 1989; Burton, Pearce, Burgio, et al., 1988; Castleden, Duffin, Asher, et al., 1985; Klarskov, Nielsen, Kromman-Andersen, et al., 1991; McDowell, Burgio, Dombrowski, et al., 1992). They may be effective in reducing incontinence following prostatic surgery in men, but to date have only been tested for postprostatectomy in conjunction with a biofeedback component (Burgio, Stutzman, and Engel, 1989). PMEs are effective in reducing UI even after multiple surgical repairs in women ( Baigis-Smith, Smith, Rose, et al., 1989; Brink, Wells, and Diokno, 1987; Burgio and Engel, 1990; Castleden, Duffin, and Mitchell, 1984; Ferguson, McKey, Bishop, et al.,1990; Heller, Whitehead, and Johnson, 1989; Henalla, Kirwan, Castleden, et al., 1988; Pearson and Droessler, 1988; Rose, Baigis-Smith, Smith, et al., 1990; Stoddart, 1983).

The effects of PME alone have been well documented in the medical literature. A summary of outcome findings from well-conducted studies follows.

Dougherty, Bishop, Mooney, et al. 1993) studied 65 women 35-75 years of age (mean age 51.3) and obtained a 62-percent reduction in UI episodes, and Ferguson, McKey, Bishop, et al. (1990)studied 20 women and demonstrated a 56- to 58-percent reduction in leakage as measured by 24-hour pad test using audiotape instruction in PME. Telephone contact and ongoing use of a bladder record may have contributed to the greater success found in these studies compared with others using a one-time instruction session.

To date, the most effective use of PME has been reported in an uncontrolled study by Benvenuti, Caputo, Bandinelli, et al. 1987), who provided 20 sessions of PME and other behavioral strategies and an intensive home program for a total of 10 contact hours to 20 women aged 35-65 years (mean age 50.8). In addition to a 95-percent reduction in incontinent episodes, significant changes were also reported FUL and MUCP at rest and during maximal voluntary contraction. A significant clinical benefit was consistent with changes in several physical measures of bladder function.

In a randomized clinical trial comparing PME and phenylpropanolamine hydrochloride (PPA), PME was found to be an effective alternative treatment for SUI comparable to PPA (Wells, Brink, Diokno, et al., 1991). Of the 54 subjects who completed the exercise program and the 64 who completed the drug protocol, 77 percent of the exercise subjects and 84 percent of the drug subjects reported improvement. The PME protocol involved 6 months of active PME taught with only written instruction, followed by monthly monitoring visits. The PPA was administered in doses of 50 mg daily for 2 weeks, increasing to twice daily if wetting continued. Adherence to the drug treatment was greater than to the exercise protocol.

Indications are that the intensity of the exercise program affects physiological and functional outcomes. Bo, Hagen, Kvarstein, et al. (1990) found that 52 subjects of mean age 45.9 years (range 24-64) randomized into a group receiving ongoing guidance in performing maximum contractions of the pelvic muscle that increased in intensity over 6 months reportedsignificantly greater reduction in incontinence and changes in physiological measures of pelvic floor strength compared with a group that received only a single session of instruction and a home exercise program. The group receiving the ongoing instruction reported a 60.1-percent cure/improvement rate compared with 17.3 percent attained in the home exercise group. This work also indicates a possible systematic relationship between symptom reduction and objective physical changes. Based on the data presented (Benvenuti, Caputo, Bandinelli, et al., 1987; Elia and Bergman, 1993; Mouritsen, Frimodt, and Moller, 1991), there is evidence that pelvic muscle re-education has the potential to change muscle physiology. Standards for assessment of change in pelvic muscle function have yet to be established, however.

PME also appears to be effective in the treatment of older adults. Flynn, Cell, and Luisi (1994) provided treatment for transient UI in 37 older adults receiving home nursing care (mean age 76) and behavioral strategies that included education in the maintenance of bowel regularity, bladder and habit training, fluid intake management, and PME. An 82.4-percent decrease in UI was attained with an average of five nursing visits. Because those patients with transient UI were not separated in the final analysis, the specific effect of the behavioral treatment is difficult to determine. Nonetheless, this report and others (Scheve, Engel, McCormick, et al., 1991) demonstrate the overall effectiveness of a well-directed nursing program, which includes PME for UI in the homebound and long-term chronically ill patients.

Summary of findings. Evidence demonstrates that patients require repeated guidance over an extended period of time to derive optimal benefit from PME. Kegel himself recommended that patients be seen weekly to ensure that the exercises were being performed correctly. Further controlled studies are needed to demonstrate the most efficient method of teaching PME, exercise prescription, and the conditions in which biofeedback provides an added benefit to PME alone. It should also be noted that the majority of the research on PME is in women only.

PME and Bladder Inhibition Augmented by Biofeedback Therapy

Some type of biofeedback device is often used to assist patients to gain function and pelvic muscle awareness { Burns, Pranikoff, Nochajski, et al., 1990). The aim of biofeedback therapy, which uses electronic or mechanical instruments to relay information to patients about their physiologic activity, is to improve bladder dysfunction by teaching people to change physiologic responses that mediate bladder control (Burgio and Engel, 1990). Auditory or visual display of this information forms the core of biofeedback procedures (Schwartz, 1995). Biofeedback for UI typically uses single measurement (surface, needle, vaginal, or anal probe) EMG or manometric methods. Biofeedback using multimeasurement feedback methods involves simultaneous measurement of pelvic and abdominal/detrusor muscle activity. Biofeedback should be used in conjunction with other behavioral techniques such as PME and bladder training. As with all of the behavioral techniques, successful application of biofeedback depends greatly on the knowledge and skill of the health care provider, whose knowledge must include familiarity with evaluation techniques, anatomic and physiologic correlates of the different forms and symptoms of bladder dysfunction, instrumentation, and behavioral principles that guide the procedure.

Studies on the various applications of biofeedback combined with behavioral treatment report a range of 54-87 percent improvement in incontinence across various patient groups using different biofeedback and behavioral procedures. Some biofeedback protocols use only one measure for reinforcement of pelvic muscle contraction, whereas others use up to three, and include measures of abdominal and detrusor activity. The biofeedback protocol that has been associated with the largest and most consistent symptom reduction is one that reinforces pelvic muscle contraction concurrently with inhibition of abdominal and detrusor contraction. Reports using this multimeasurement method show a 75.9-82 percent reduction in UI across six studies involving 166 subjects Burgio, Robinson, and Engel, 1986; Burgio, Stutzman, and Engel, 1989;(Burgio, Whitehead, and Engel, 1985; Burton, Pearce, Burgio, et al., 1988; McDowell, Burgio, Dombrowski, et al., 1992). The presumed benefit of the multimeasurement procedure is that it reinforces pelvic floor contraction directly with moment-to-moment feedback, which characterizes for the patient the quality and intensity of the contraction. Without biofeedback, weak pelvic muscles may provide limited kinesthetic feedback to the desired contraction, and as a result the patient uses an attenuated internal reference to upgrade muscle contractions. Combining bladder and sphincter biofeedback also allows teaching pelvic muscle contraction in response to increasing bladder volume and observed detrusor activity.

In a controlled study, Burgio, Robinson, and Engel (1986) found that 13 subjects (mean age 47.9) receiving multimeasurement biofeedback reduced incontinence by 75.9 percent compared with a 51-percent reduction obtained in 11 subjects (mean age 40.7) who received only verbal feedback and instruction for PME with digital palpation. The multimeasurement biofeedback method has also been used successfully in the treatment of UI in postprostatectomy patients with urge and intermittent stress incontinence (Burgio, Stutzman, and Engel, 1989). However, another study demonstrated that when subjects complained primarily of urge incontinence, no benefit was obtained with the addition of biofeedback to behavioral training and PME, which was provided by a well-trained clinician over an average of five clinic sessions (Burton, Pearce, Burgio, et al., 1988). The results of this study should be interpreted with caution, however, because the groups differed in severity before treatment.

Several studies have also demonstrated significant reductions in UI associated with neurologic disease and in the frail elderly using a combination of multimeasurement biofeedback and other behavioral techniques such as bladder training (McDowell, Engberg, Weber, et al., 1994; Middaugh, Whitehead, Burgio, et al., 1989; O'Donnell and Doyle, 1991). Because multimeasurement biofeedback can provide specific reinforcement for pelvic muscle contraction that is isolated from counterproductive abdominal contraction, it is assumed that awareness of pelvic muscle contraction can be achieved more efficiently than from vaginal palpation alone.

In contrast to multimeasurement methods, the reduction in number of pad changes per 24 hours ranged from 43 to 54 percent when single-measurement biofeedback is used, when excluding a subgroup who had received only one biofeedback session (Wilson, Faragher, Butler, et al., 1987). This improvement was attained in an approximate average of 11 sessions (0.5 to 1.5 hours each). Another study (Susset, Galea, and Read, 1990) used single-channel biofeedback over six weekly clinic visits and a home trainer with which the subjects practiced daily. These subjects demonstrated an 87-percent reduction of leakage on pad test, suggesting that a home training device may provide an added benefit to clinic biofeedback visits, especially when only single-channel biofeedback is used.

In a randomized controlled study, Burns, Pranifoff, Nochajski, et al. (1993) studied 135 women (age range 55-75 years) with primary stress incontinence in three groups. One group received PME, and another group received single-measurement EMG biofeedback to perivaginal contraction for 20 minutes per week. The two treatment groups demonstrated a 54-61 percent reduction in incontinent episodes, compared with a 6-percent reduction of incontinence in the control group, but no difference between the two treatment groups was found. The improvements were maintained over a 6-month followup, and patients with moderate-to-severe symptoms showed even more improvement in the posttreatment phase. This study demonstrated that PME reduces incontinence and provided evidence that symptom severity has a role in response.

Further research is needed to determine which biofeedback protocols ensure that optimal outcomes are achieved in individual conditions and what methods provide the most valid measures of pelvic muscle function. The multimeasurement biofeedback appears to produce greater reduction in incontinence compared with PME alone or single measurement biofeedback. It is not known, however, to what degree detrusor or intra-abdominal pressure biofeedback individually contributes to the outcomes reported.

Summary of findings. Overall, the literature indicates that PME and other behavioral strategies, with or without biofeedback, can "cure" or reduce incontinence. Maximum benefit is derived from any pelvic muscle rehabilitation and education program when ongoing reinforcement and guidance are provided. Also, the intensity of the exercise program seems to influence both functional and physiological outcomes, and multimeasurement biofeedback protocols seem to yield the greatest and most consistent reductions in UI (Bo, Hagen, Kvarstein, et al., 1990; Burgio, Robinson, and Engel, 1986; Elia and Bergman, 1993; McDowell, Burgio, Dombrowski, et al., 1992).

Pelvic Muscle Exercises Augmented with Vaginal Weight Training

Specially designed vaginal weights for strengthening the pelvic muscles can augment PME. The patient receives a set of vaginal weights of identical shape and volume but of increasing weight (20-100 grams). As part of a structured progressive resistive exercise program, women insert the weight intravaginally, with the tapered portion resting on the superior surface of the perineal muscle and attempt to retain it by contracting the pelvic muscles up to 15 minutes. The weight is worn while the patient is ambulatory, and the exercise is done twice daily. The hypothesized mechanism of action is that the sustained contraction required to retain the weight increases the strength of the pelvic muscles, and the weight is assumed to provide heightened proprioceptive feedback to desired pelvic muscle contraction.

Available literature on this technique includes observations made in premenopausal women with SUI. Initial observations from four studies including 103 premenopausal women indicate subjective "cure" or greatly improved status of 68-80 percent after 4-6 weeks of treatment (Bridges, Denning, Olah, et al., 1988; Olah, Bridges, Denning, et al., 1990; Peattie and Plevnik, 1988; Peattie, Plevnik, and Stanton, 1988; Wilson and Borland, 1990). Objective outcome measures included reduction of urine loss on pad test, improvement in ability to hold heavier weights intravaginally, increased pelvic muscle strength (perineometer), and significant reduction in incontinence episodes. There were minimal or no adverse reactions. However, in several studies, PMEs were performed at the same time.

Although vaginal weight training may be useful in the treatment of stress incontinence, issues of applicability to other populations, particularly postmenopausal women with pelvic organ prolapse or other comorbid conditions, must be evaluated in terms of treatment protocols and long-term effects.

Pelvic Floor Electrical Stimulation

Pelvic floor electrical stimulation (nonimplantable) produces a contraction of the levator ani, external urethral and anal sphincters, accompanied by a reflex inhibition of the detrusor; this activity depends on a preserved reflex arc through the sacral micturition center (Vodusek, Plevnik, Vrtacnik, et al., 1988). Nonimplantable pelvic floor electrical stimulation uses vaginal or anal sensors or surface electrodes (Vodusek, Plevnik, Vrtacnik, et al., 1988). Adverse reactions are minimal and include pain and discomfort. Studies vary regarding the type and placement of electrodes; frequency, duration, and amplitude of voltage; and whether the stimulation was phasic, intermittent, or continuous. Several of these studies address long-term followup with reports that the effects for cured or improved patients ranged from 54 to 77 percent (Bent, Sand, Ostergard, et al., 1993; Bergmann and Eriksen, 1986; Blowman, Pickles, Emery, et al., 1991; Caputo, Bensen, McClellan, et al., 1993; Eriksen, 1990; Eriksen, Bergmann, and Mjolnerod, 1987; Eriksen, Bergmann, and Eik-Nes, 1989; Esa, Kiwamoto, Sugiyama, et al., 1991; Fall, Ahlstrom, Carlsson, et al., 1986; Fossberg, Sorensen, Ruutu, et al., 1990; Green and Laycock, 1990; Hahn, Sommar, and Fall, 1991; Lamhut, Jackson, and Wall, 1992; Leach and Bavendam, 1989; Lose, Andersen, and Kristensen, 1986; McIntosh, Frahm, Mallett, et al., 1993; Meyer, Dhenin, Schmidt, et al., 1992; Eriksen and Mjolnerod, 1987; Nakamura, Sakurai, Tsujimoto, et al., 1986; Nakamura, Sakurai, Sugao, et al., 1987; Ohlsson, Fall, and Frankenberg-Sommar, 1989; Plevnik, Joney, Vrtacnik, et al., 1986; Sand, Richardson, Staskin, et al., 1995; Wilson, 1990; Zollner-Nielsen and Samuelsson, 1992).

Two randomized controlled trials have been conducted. Using active and placebo perianal surface patch neurostimulation for SUI in patients, Blowman, Pickles, Emergy, et al. 1991) reported a "cure" or improvement rate of 86 percent in the active group (N=7) and 33 percent in the placebo group (N=6). Electrical stimulation was not used as a single therapy, however; both groups also received instruction in PME. Using active and inactive vaginal plug devices in 52 women with SUI, Sand, Richardson, Staskin, et al. (1995) reported objective "cure" or improvement in 48 percent of the active device group and in 13 percent of the placebo group. Active device patients had significant improvements in UI episodes, leakage volume, and vaginal muscle strength and in subjective improvement measures when compared with the placebo group.

In other studies with similar settings, anal or vaginal plug devices were used for maximal electrical stimulation for 4 weeks to 3.5 months, 20 minutes to 20 hours/day. "Cure" or improvement rates ranged from 48 to 94 percent in 842 patients with stress, urge, or mixed incontinence. The effects of therapy were sustained 6 weeks to 2 years in 54-77 percent of patients, especially if patients continued to do PME after treatment (Bergman and Eriksen, 1986; Blowman, Pickles, Emery, et al., 1991; Bridges, Denning, Olah, et al., 1988; Caputo, Bensen, McClellan, et al., 1993; Eriksen, Bergmann, and Eik-Nes, 1989; Eriksen, Bergmann, and Mjolnerod, 1987; Eriksen and Eik-Nes, 1989; Esa, Kiwamoto, Sugiyama, et al., 1991; Fossberg, Sorensen, Ruutu, et al., 1990; Kunkle, Payne, and Whitmore, 1993; Meyer, Dhenin, Schmidt, et al., 1992; Plevnik, Joney, Vrtacnik, et al., 1986; Sand, Richardson, Staskin, et al., 1995; Zollner-Nielsen and Samuelsson, 1992). One study of cognitively impaired patients in a nursing home showed no significant effect of stimulation, with a trend toward increased wetness (Lamhut, Jackson, and Wall, 1992).

Summary of findings. Research indicates that pelvic floor electrical stimulation can significantly reduce UI in women with SUI, and may be effective in men and women with mixed and urge UI. Stimulation may be effective when augmented with other pelvic muscle rehabilitation therapies. Minimal adverse side effects occur with this treatment. Treatment using stimulation requires monitoring by a health care provider to determine effectiveness. Further research is needed to determine the efficacy of pelvic floor stimulation used alone or in combination with other therapies. Standardization of the parameters of the techniques used, such as that proposed by the International Continence Society, is necessary to allow further comparison of study results.

Anticholinergic Agents

Anticholinergic agents are effective as treatment for UI because they block contraction of the normal bladder and probably the unstable bladder as well. All anticholinergic drugs are contraindicated in patients with documented narrow-angle, but not wide-angle, glaucoma.

Oxybutynin has both anticholinergic and direct smooth muscle relaxant properties. Seven randomized controlled studies of the use of this agent for UI were identified. Oxybutynin proved superior to placebo in six studies of middle-aged outpatients, reducing incontinence frequency by 15-58 percent over the response to placebo (Holmes, Montz, and Stanton, 1989; Moore, Hay, Imrie, et al., 1990; Riva and Casolati, 1984; Tapp, Cardozo, Versi, et al., 1990; Thuroff, Bunke, Ebner, et al., 1991; Zeegers, Kiesswetter, Kramer, et al., 1987). In two of these studies, 43 percent and 67 percent of subjects became subjectively continent, a much better result than occurred with placebo (Moore, Hay, Imrie, et al., 1990; Riva and Casolati, 1984). Only one study reported a benefit with a dosage of 15 mg tid-qid Holmes, Montz, and Stanton, 1989 ), whereas a second showed no improvement over placebo (Thuroff, Bunke, Ebner, et al., 1991). In the only trial that failed to note a benefit, oxybutynin was used less frequently (5 mg bid) in elderly nursing home residents (Zorzitto, Holliday, Jewett, et al., 1989). Oxbutynin has also been effective and well tolerated when instilled into the bladder of study patients, primary those with neuropathic bladder (Mizunaga, Miyata, Kaneko, et al., 1994).

Side effects were noted in all studies and included dry skin, blurred vision, change in mental status, nausea, constipation, and marked xerostomia (dry mouth). Severity increased with dosage, with severe xerostomia occurring in 84 percent of subjects receiving 5 mg of oxybutynin four times per day (Tapp, Cardozo, Versi, et al., 1990).

Propantheline is the prototype for anticholinergic agents used for urologic conditions. No agent better approximates atropine's effect on the bladder in vitro, although its central nervous system side effects are less marked. Moreover, propantheline is inexpensive and has been widely used over time.

Despite propantheline's success in uncontrolled case series, only five adequately controlled trials could be identified Blaivas, Labib, Michalik, et al., 1980; Dequeker, 1965; Holmes, Montz, and Stanton, 1989; Thuroff, Burke, Ebner, et al., 1991; Zorzitto, Jewett, Fernie, et al., 1986), three of which included only elderly nursing home residents with advanced dementia. Only one study reported a benefit with a dosage of 15 mg tid-qid (Holmes, Montz, and Stanton, 1989). Two of the nursing home studies evaluated higher dosages of propantheline (30 mg qid; 15 mg tid + 60 mg at bedtime (hs)) and found that incontinence frequency was reduced by 13-17 percent over placebo -- a small but statistically significant improvement in both studies Dequeker, 1965; Zorzitto, Jewett, Fernie, et al., 1986). In one of these trials, side effects were reported in half of the subjects, one-fifth of whom withdrew because of severe side effects Zorzitto, Jewett, Fernie, et al., 1986); no side effects were reported in the trial using the 60 mg hs dosage. In a dose titration study, dosages up to 60 mg qid were associated with complete response in 25 of 26 patients (Blaivas, Labib, Michalik, et al., 1980).

In addition to urinary retention, side effects associated with propantheline, as with all anticholinergic agents, include visual blurring, xerostomia, nausea, constipation, tachycardia, drowsiness, and confusion; the most common of these was xerostomia.

Despite the lack of adequate trials, consensus among experts is that at least for less impaired patients who can tolerate full dosages, propantheline is effective (Andersson, 1988; Wein, 1990).

Dicyclomine hydrochloride is an anticholinergic agent with smooth muscle relaxant properties. Studies on the use of this drug for UI are limited, and the populations that have been studied were small. One pilot trial and one larger randomized controlled trial were identified Beck, Arnusch, and King, 1976; Castleden, Duffin, and Millar, 1987). In the larger study, 62 percent of subjects improved on dicyclomine 10 mg tid (90 percent of whom became continent) compared with 20 percent on placebo (65 percent of whom became continent); statistical significance was not specified, and side effects were not mentioned Beck, Arnush, and King, 1976). Results were similar in the smaller pilot study (Castleton, Duffin, and Millar, 1987).

Studies comparing dicyclomine with other anticholinergic agents are not available. This agent may be used as an alternative treatment for UI, however, pending further research. The dosage is 10-20 mg taken orally three times daily.

Flavoxate is a tertiary amine that has demonstrated smooth muscle relaxant properties in vitro. Although flavoxate is widely used for incontinence, only four randomized controlled studies of its efficacy for UI could be identified (Meyhoff, Gerstenberg, and Nordling, 1983; Robinson and Brocklehurst, 1983; Zeegers, Kiesswetter, Kramer, et al., 1987; Chapple, Parkhouse, Gardener, et al., 1990). None demonstrated a significant benefit. At present, this drug is not recommended for the treatment of UI.

Hyoscyamine and other oral anticholinergics are known to be used in clinical practice in the treatment of DI; however, no scientific literature that met the panel's criteria addressed the use of these agents for patients with UI. The panel therefore is not making recommendations regarding these drugs.

Calcium Channel - Blocking Agents

Influx of extracellular calcium is important for detrusor muscle contraction and can be blocked by calcium channel antagonists. Although such agents are often advocated for bladder storage disorders, only a few small, but favorable, case series were identified in the literature. No controlled studies could be found for marketed agents (e.g., nifedipine, diltiazem, verapamil). One placebo-controlled study of flunarizine that showed a positive effect was identified; however, in a subsequent 1-month controlled trial by the same investigators, its efficacy diminished, and the high (but unreported) rate of side effects led the investigators to temper their original enthusiasm Palmer, Worth, and Exton-Smith, 1981). At this time, these agents are not recommended for general use for the treatment of detrusor instability.

Terodiline. An investigational drug in the United States, terodiline possesses calcium channel - blocking and anticholinergic properties. It has demonstrated in vivo activity in the treatment of DI. However, recent reports of serious ventricular arrhythmias have resulted in its temporary removal from the European market and discontinuation of trials in the United States.

Tricyclic Agents

Although TCAs are widely used, they produce many adverse effects (e.g., cardiac and anticholinergic), with the risk more likely in elderly patients. Limited research on the use of TCAs is available. Only three randomized controlled studies were identified. One conducted with psychiatric inpatients in whom the type of UI was not established showed that imipramine, desipramine, and nortriptyline each resulted in a statistically significant reduction in the number of "wet nights" (Milner and Hills, 1968). Side effects were rare and minor except for paralytic ileus in one case.

The other two studies, one using doxepin Lose, Jorgensen, and Thunedborg, 1989) and the other imipramine (Castleden, Duffin, and Gulati, 1986), documented decreases in incontinence frequency. However, the results were statistically significant only for reducing nocturnal incontinence and overall patient preference for doxepin. Side effects in these studies included fatigue, xerostomia, dizziness, and blurred vision in the doxepin group and nausea and insomnia in the imipramine group.

Nonsteroidal Anti-Inflammatory Drugs

NSAIDs are theorized to be effective for UI because of their inhibition of prostaglandin "synthetase," which interferes with prostaglandin-mediated bladder contractions. However, limited research on the use of NSAIDs for patients with UI is available. Two controlled trials in women have been conducted (Cardozo and Stanton, 1979; Cardozo, Stanton, Robinson, et al., 1980). In a non-placebo-controlled trial, patients receiving indomethacin 100 mg po bid experienced comparable resolution of UI compared with those receiving bromocriptine (Cardozo and Stanton, 1979). A placebo-controlled trial demonstrated 55-percent continence with flurbiprofen 50 mg po tid versus 29 percent with placebo (Cardozo, Stanton, Robinson, et al., 1980). In both studies, side effects were frequent, although no patient stopped use of NSAIDs. When NSAIDs are used, close monitoring for drug toxicity (e.g., gastrointestinal hemorrhage, renal dysfunction) should be performed, especially in elderly patients.

Other Drugs of Possible Benefit

Other drugs used for detrusor instability include a beta-adrenergic agonist (terbutaline) (Lindholm and Lose, 1986) and a spinal synaptic inhibitor (baclofen) (Taylor and Bates, 1979). Limited studies and clinical experience with these agents suggest that further studies must be done before they can be recommended for general use.

Summary of findings. Pharmacotherapy, at least in short-term trials, appears to benefit some patients with UI due to detrusor instability. However, regardless of the agent chosen, involuntary bladder contractions are usually not abolished, the "warning time" between appreciation of the need to void and the onset of bladder contraction is usually not affected, the degree of improvement is modest, and "cure" is uncommon. In addition, many of these medications are costly, and most have bothersome side effects. These drugs should be used only in conjunction with a voiding schedule or behavioral intervention and only after other factors contributing to incontinence have been addressed.

To date, oxybutynin possesses the most favorable efficacy, safety, and pharmacokinetics profile of the agents for DI. However, selection of an agent must be individualized and based on patient considerations (i.e., contraindications) and cost as well as on drug considerations (i.e., kinetics, drug interactions, and side effects). For example, an antidepressant may be the preferred drug for UI in a depressed patient but not for a patient with orthostatic hypotension who is prone to falling. A drug with a rapid onset of action may be useful for a patient who desires protection only for certain events (e.g., going out for the afternoon), whereas a drug with a longer half-life may be preferable for a patient who desires continuous protection.

Regardless of the agent selected, the initial dosage should be relatively low. The dosage should be increased slowly, consistent with the drug's half-life, and titrated to balance efficacy with side effects. Patients must be closely monitored for urinary retention. Combining agents with different mechanisms of action may increase their efficacy and reduce side effects, but few relevant data are available. Although the long-term efficacy of bladder relaxants is largely unknown, Moore, Hay, Imrie, et al. (1990) have reported that once incontinence responded to pharmacologic and behavioral interventions, improvement persisted even if the medication was withdrawn after a few months.

Alpha-Adrenergic Agonist Drugs

PPA in sustained-release form is the major alpha-adrenergic agonist drug studied in women with stress incontinence. Seven prospective randomized controlled studies of middle-aged, normotensive women with stress incontinence were reviewed. In three studies (Collste and Lindskog, 1987; Fossberg, Beisland, and Lundgren, 1983; Lehtonen, Rannikko, Lindell, et al., 1986) no patients became continent, but 31-45 percent of patients given PPA had decreased incontinence over placebo response (PPA minus placebo "improved" response). In four studies (Ek, Andersson, Gullberg, et al., 1978; Hilton, Tweddell, and Mayne, 1990; Walter, Kjaergaard, Lose, et al., 1990; Wells, Brink, Diokno, et al., 1991), 9-14 percent of women became dry, and 19-60 percent of patients experienced significant reduction of incontinence over the number of similar responses in the placebo group. In studies in which the improvement rate was reported, the reduction in leakage with PPA ranged from 31 to 60 percent over the response with placebo.

Side effects were minimal and included nausea, xerostomia, insomnia, rash, itching, and restlessness. PPA did not cause significant increases in blood pressure during the evaluation period.

Summary of findings. Pharmacologic therapy of incontinence caused by sphincter insufficiency (stress incontinence) using PPA appears to result in few cures or dryness (0-14 percent) but may cause subjective improvement in 20-60 percent of patients over placebo response. It is unclear whether patient age or the severity of leakage affects the likelihood of response. Possible side effects from alpha-adrenergic agonist drugs include anxiety, insomnia, agitation, respiratory difficulty, headache, sweating, hypertension, and cardiac arrhythmias, all of which may occur more commonly in elderly patients. The risk of PPA use in hypertensive women and its efficacy in women taking antihypertensive drugs have not been determined. PPA should be used with caution in patients with hypertension, hyperthyroidism, cardiac arrhythmias, and angina. The sustained-release form is advocated because of its routine use by investigators in clinical trials. Whether the immediate-release form produces similar results has not been established.

Estrogen Therapy

Because the vagina and urethra are of similar embryologic origin, estrogen supplementation in postmenopausal women may restore urethral mucosal coaptation and increase vascularity, tone, and the alpha-adrenergic responsiveness of urethral muscle, which in turn may increase bladder outlet resistance and decrease stress incontinence. However, the exact role of estrogen, as well as its mechanism of action, is still unknown and deserves further research. Most clinical trials have been conducted with estrogen products available abroad; but their data are included in this report because of the paucity of research using marketed products in the United States and the similarities between estrogens.

Estriol derivatives. Two studies of postmenopausal women with stress incontinence compared effects of placebo and estriol, 4 mg/day, on continence ( Samsioe, Jansson, Mellstrom, et al., 1985; Walter, Kjaergaard, Lose, et al., 1990). Results showed that 0-14 percent of patients on estriol became dry, and 29 percent of patients experienced improved continence compared with patients who received placebo. One study showed that the response to estriol was significantly better than to placebo (Walter, Kjaergaard, Lose, et al., 1990), whereas the other study showed that the response to active drug was not significantly better than to placebo (Samsioe, Jansson, Mellstrom, et al., 1985). In the latter study, 66 percent of patients with UI reported improvement.

A trial of quinestradol, an estriol derivative, in female incontinent nursing home residents did not result in continence, but 89 percent of patients had a significant reduction of leakage episodes while taking quinestradol compared with 0 percent while on placebo (Judge, 1969).

Estrone derivatives. A 3-month prospective study of postmenopausal stress-incontinent women who received placebo or piperazine estrone sulfate for 3 weeks with one drug-free week per cycle revealed that subjective and objective improvements at 3 months were greater, but not statistically significant, while on estrogen (Wilson, Faragher, Butler, et al., 1987).

Fantl, Cardozo, McClish, et al. (1994) provided a meta-analysis from 6 randomized controlled trials and 17 uncontrolled clinical series. Favorable effects on incontinence in postmenopausal women were found with estrogen therapy (P<0.01 for all subjects and P< 0.05 for those with SUI).

Summary of findings. Limited evidence suggests that estrogen therapy by oral or vaginal administration may benefit some patients with SUI and mixed UI. Its effect on urge UI has not been reported in most of the available literature.

Other beneficial effects of long-term estrogen may include decreased risk of stroke, ischemic heart disease, and osteoporosis. Estrogen replacement should be given with a progestin when the uterus is present to avoid unopposed estrogen stimulation of the endometrium, particularly if prolonged therapy is anticipated. The dosages of estrogen and progestin used vary. Risks of estrogen therapy given in this manner include neoplasia of estrogen-responsive organs. Thus, estrogen therapy is contraindicated in patients with known or suspected cancer of the breast or uterus.

Alpha-Adrenergic Agonist and Estrogen Supplementation

The rationale supporting combined estrogen supplementation and alpha-adrenergic agonist therapy in postmenopausal women with incontinence due to sphincteric insufficiency is based on an estrogen-induced increased number, sensitivity of alpha-adrenergic receptors in the urethra, or both, which potentiates the alpha-adrenergic contractile response to drug stimulation.

Review of four controlled studies that combined estrogen and alpha-adrenergic agonist therapy suggests that the combined therapy may be more effective than alpha-adrenergic agonist therapy alone (Ek, Andersson, Gullberg, et al., 1978, 1980; Hilton, Tweddell, and Mayne, 1990; Walter, Kjaergaard, Lose, et al., 1990). However, the studies are limited, and based on these results, combination therapy should be considered when initial single-drug therapy fails. The possible risks of alpha-adrenergic agonist drugs and estrogen do not appear to be increased when used together. Patients should be carefully monitored when combination therapy is prescribed, however.

Other Drugs of Possible Benefit

Imipramine. A tricyclic antidepressant that possesses both alpha-adrenergic agonist activity (presumably mediated by blocking reuptake of norepinephrine) and anticholinergic properties has been reported to benefit women with stress incontinence. In a nonrandomized, uncontrolled study, 30 women (28-64 years old) with pure stress incontinence received imipramine at 75 mg daily for 4 weeks; 21 of the 30 (70 percent) claimed continence (Gilja, Radej, Kovacic, et al., 1984). Side effects reported included nausea, insomnia, weakness, fatigue, and postural hypotension. Prospective controlled studies are not available, and thus further research is necessary.

Propranolol. Limited research on the use of beta-blockers for treatment of SUI is available. In one uncontrolled study, a beta-adrenergic-blocking drug was reported to improve symptoms of SUI (Gleason, Reilly, Bottaccini, et al., 1974). The use of this and other beta-blockers cannot be recommended for treatment of incontinence at this time.

Antidiuretic hormone. Desmopressin (desmopressin diacetate arginine vasopressin (DDAVP)) is used mainly in the treatment of nocturnal enuresis and night-time polyuria. Available research on this drug has been conducted mainly in children and in patients with neuropathic conditions. Because few data exist in adults, DDAVP is not discussed in detail in this guideline.

Preoperative Evaluation

A preoperative evaluation is important to ensure proper patient selection and to determine the appropriate surgical procedure. The evaluation should include a comprehensive history, physical examination, urinalysis, urine culture, and measurement of PVR volume. It is important to document the incontinence objectively by direct observation of a positive stress test (direct visualization). Some multiparous women may have stress incontinence on physical examination but are not troubled by this; therefore, the patient should be asked if the incontinence visible on examination is the type of incontinence that led her to seek care. Because CMGs may be falsely negative in many women with DI, simply using a negative CMG to diagnose SUI is not acceptable. When further corroboration is needed, cystoscopy, a cystogram with straining, measurement of Valsalva leak point pressure, dynamic UPP, or a combination may be used.

Although patients with symptoms of DI may be more likely to remain incontinent after surgery, it is not clear whether a preoperative CMG to search for asymptomatic instability is necessary. If symptomatic instability is present, some health care providers may attempt behavioral or pharmacologic treatment before surgical correction of the stress incontinence. Others may not because many patients experience resolution of DI once stress incontinence is corrected. This decision must be made on an individual basis, using clinical judgment and considering the patient's informed preference.

To select the appropriate surgical procedure, one must assess the position of the urethra and the degree of axial urethral mobility by physical examination, cystogram, ultrasound, or cystoscopy. Factors that may be associated with an increased risk of failure, such as symptomatic DI, obesity, prior surgery, hypoestrogenism, chronic cough, strenuous physical activity, previous radiation therapy, advanced age, or poor nutrition, must be identified. Pelvic organ prolapse or other pathology that would require surgical treatment at the same time as the anti-incontinence procedure should also be identified. The approach required to treat the concurrent pathology will influence the choice of approach for the anti-incontinence procedure. The extent of the evaluation depends on the complexity of the presentation.

The degree of ISD should be assessed preoperatively using a combination of assessment tools. In the history, patients who have only hypermobility describe leakage with vigorous activity, whereas patients with ISD tend to have leakage with minimal activity. On physical examination, patients with incontinence due to hypermobility have descent of the urethra and bladder neck during stress maneuvers. On the other hand, reproducible demonstration of stress incontinence through a normally supported urethra is suggestive of ISD. Corroboration of ISD may be established by a variety of tests, including leak point pressure, passive urethral profilometry, imaging techniques such as fluoroscopy, and on occasion cystoscopic assessment. Most of these assessment tools do not provide a definitive diagnosis alone, especially if the patient has coexistent DI, which can cause a functional opening of the bladder neck at rest. Therefore, the entire clinical presentation must be considered.

Procedures for Hypermobility

After complete evaluation, if the primary pathophysiologic defect appears to be urethral hypermobility or displacement, three main types of procedures are used:

• Retropubic suspension.

• Needle bladder neck suspension.

• Anterior vaginal repair.

Review of the available literature shows that in general retropubic and needle suspension procedures produce a superior result to that of anterior repair in "curing" UI. This analysis of results usually focuses exclusively on the symptom of stress incontinence and does not account for postoperative problems such as the new development of an enterocele or rectocele large enough to require surgery, persistent voiding difficulty, or DI. Therefore, although the reported cure rate for stress incontinence is high, the number of women with an entirely satisfactory outcome may not be. Treatment must be individualized for each patient as well as standard recommendations about which operation will be best. The option selected depends on the surgeon's training and expertise and on the presence of concurrent pathology that would require correction by a vaginal or abdominal approach.

For hypermobility with coexisting ISD the surgical procedure should stabilize the anatomic support and compress the urethra, which invariably means using one of the sling procedures (see Procedures for ISD below). The choice must be individualized for each patient. Women who have severely damaged urethras require special procedures such as urethral or bladder neck reconstruction, urethral substitution, continent vesicotomy, or urinary diversion.

Management must be individualized for women whose primary problem is prolapse of the pelvic organs but who have SUI or in whom temporary replacement of the prolapse during preoperative evaluation uncovers SUI. In these circumstances, a decision about the most appropriate anti-incontinence procedure to be combined with repair of the prolapse should account for the type of SUI present (hypermobility or ISD), the operation chosen to repair the prolapse (abdominal or vaginal), and additional morbidity from the continence operation weighed against the degree of incontinence. Preoperative urodynamic examination should be performed if this information is needed to choose the type of operation to be performed.

Retropubic suspension. Retropubic suspension procedures include several different techniques performed through a low abdominal incision (i.e., retropubic approach). All techniques have in common elevation of the lower urinary tract, particularly the urethrovesical junction, within the retropubic space. The procedures differ according to what structures are used to achieve the elevation. For the Marshall-Marchetti-Krantz procedure, the periurethral tissue is approximated to the symphysis pubis. For the Burch colposuspension, the vaginal wall lateral to the urethra and bladder neck is elevated toward Cooper's ligament. The paravaginal repair involves reapproximating the endopelvic fascia to the pelvic wall at the arcus tendineus.

To evaluate this category of procedures, 45 studies incorporating 3,882 patients were reviewed (Baker and Drutz, 1992; Beck, Thomas, and Maughan, 1968; Bergman, Ballard, and Koonings, 1989; Bergman, Koonings, and Ballard, 1989a; Bergman, Koonings, and Ballard, 1989b; Bergman, Koonings, and Ballard, 1989c; Brieger and Korda, 1992; Duncan, Nurse, and Mundy, 1992; Eriksen, Hagen, Eik-Nes, et al., 1990; Fall, Erlandson, and Pettersson, 1984; Ferriani, de Sa, de Moura, et al., 1990; Francis, Sand, Hamrang, et al., 1987; Goodno and Powers, 1992; Green, McGuire, and Lytton, 1986; Hilton and Stanton, 1983; Iosif, 1985; Jouppila, Kauppila, Ylikorkala, et al., 1977; Kiilholma, Makinen, Chancellor, et al., 1993; Kil, Hoekstra, Van der Meijden, et al., 1991; Korda, Ferry, and Hunter, 1989; Kujansuu, 1983; Langer, Ron-El, Newman, et al., 1988; Langer, Ron-El, Bukovsky, et al., 1988; Lockhart, Maggiolo, and Politano, 1983; Lose, Jorgensen, and Johnsen, 1988; Milani, Scalambrino, Quadri, et al., 1985; Morgan, 1973; Nielsen and Lundvall, 1973; Park and Miller, 1988; Penttinen, Kaar, and Kauppila, 1989; Penttinen, Lindholm, Kaar, et al., 1989; Pow-Sang, Lockhart, Suarez, et al., 1986; Richardson, Ramahi, and Chalas, 1991; Richmond and Sutherst, 1989a; Sand, Bowen, Ostergard, et al., 1988; Shull and Baden, 1989; Spencer, O'Conor, and Schaeffer, 1987; Stanton, Williams, and Ritchie, 1976; Stanton and Cardozo, 1979; Steel, Cox, and Stanton, 1985; Thunedborg, Fischer-Rasmussen, and Jensen, 1990; Van Geelen, Theeuwes, Eskes, et al., 1988; Vordermark, Brannen, Wettlauffer, et al., 1979; Wheelahan, 1990). Total "cure" rates averaged 79 percent, and 84 percent were "cured" or improved.

Overall complication rates reported for retropubic operations average 18 percent (6-57 percent). In the recent literature, complications reported among the studies of retropubic urethral suspensions were sparse, being included in only 6 of 13 papers with insufficient information and consistency to allow accurate statistical reporting. Wound infections, urinary retention, de novo DI, and dyspareunia were all noted in the 3-15 percent range. The cause of postoperative symptomatic enterocele/rectocele with anterior traction on the anterior wall seems to be most prominent after the Burch operation, occurring in up to 12 percent of patients (Kiilholma, Makinen, Chancellor, et al., 1993). Recently, some of these retropubic procedures have been performed laparoscopically.

Needle bladder neck suspension. Another type of anatomic correction employs needle suspension of the bladder neck. Variations of this procedure are all performed through a vaginal approach, and most utilize small suprapubic skin incisions (Cobb and Ragde, 1978; Gittes and Laughlin, 1987; Pereyra, Lebherz, Growdon, et al., 1982; Stamey, 1980; Raz, 1981). Anchoring tissues adjacent to the urethra and bladder neck are held by suspending sutures.

Of the 3,015 patients studied in the reviewed research, most had the Stamey procedure; the others had the Pereyra-Raz procedure Abbassian, 1989; Ashken, 1990; Benderev, 1992; Benson, Agosta, and McClellan, 1990; Bergman, Ballard, and Koonings, 1989; Bergman, Koonings, and Ballard, 1989a; Bergman, Koonings, and Ballard, 1989c; Bhatia and Bergman, 1985a; Bosman, Vierhout, and Huikeshoven, 1993; Cobb and Ragde, 1978; Dijkman and Friese, 1984; Duncan, Nurse, and Mundy, 1992; Foster and O'Reilly, 1989; Ganabathi, Abrams, Mundy, et al., 1992; Gaum, Ricciotti, and Fair, 1984; Griffith-Jones and Abrams, 1990; Hilton, 1989; Huland and Bucher, 1984; Jones, Shah, and Worth, 1989; Juma, Little, and Raz, 1992; Kil, Hoekstra, et al., 1991; Kirby and Whiteway, 1989; Kursh, 1992; Kursh, Angell, and Resnick, 1991; Leach and Raz, 1984; Lopez Lopez, Valdivia Uria, and Uca Terran, 1992; Loughlin, Whitmore, et al., 1990; Mundy, 1983; Netto, Lemos, Palma, et al., 1988; Nitti, Bregg, et al., 1993; Parra and Shaker, 1990; Penttinen, Kaar, and Kauppila, 1989; Pereyra, Lebherz, Growdon, et al., 1982; Pope, Shaw, Coptcoat, et al., 1990; Pow-Sang, Lockhart, Suarez, et al., 1986; Ramon, Mekras, and Webster, 1990; Raz, Sussman, Erickson, et al., 1992; Richardson, Ramahi, and Chalas, 1991; Shah and Holder, 1989; Spencer, O'Conor, and Schaeffer, 1987; Stamey, 1980; Varner, 1990; Walker and Texter, 1992; Wujanto and O'Reilley, 1989). For the combined series, 74 percent were continent postoperatively and 84 percent were "cured" or improved; followup varied. Complications included UTI, urinary retention longer than 3 weeks' duration, obstructive symptoms, suture abscess, wound infection or vaginal granuloma, vesicocutaneous fistula, hematoma, sutures pulling out of the vaginal fascia, sepsis, new onset of symptomatic DI, and prolonged suprapubic pain and de novo pelvic floor defects.

Anterior vaginal repair. The anterior vaginal repair category of treatments includes several modifications of the original Kelly plication. All techniques include some degree of dissection of the anterior vaginal wall from the overlying bladder base and urethra, and plication of the pubocervical fascia. The extent of the dissection and the location and extent of the placating (elevating) sutures vary substantially among these techniques. Although the success of these operations as a group is somewhat lower than retropubic or needle suspensions, some specific techniques achieve excellent success, on objective followup indicating the need to discriminate the actual technical details of each of these different operations before making generalizations (Beck, McCormick, and Nordstrom, 1991).

Review of 11 studies incorporating 957 patients reveals an overall "cure" rate of 65 percent (range = 31-91 percent) and a "cure" or improvement rate of 74 percent (range = 31-98 percent); followup varied (Beck, McCormick, and Nordstrom, 1991; Beck, Thomas, and Maughan, 1968; Bergman, Ballard, and Koonings, 1989; Bergman, Koonings, and Ballard, 1989c; Jouppila, Kauppila, Ylikorkala, et al., 1977; Kujansuu, 1983; Lose, Jorgensen, and Johnsen, 1988; Park and Miller, 1988; Stanton, Hilton, Norton, et al., 1982; Thunedborg, Fischer-Raumussen, and Jansen, 1990; Van Geelen, Theeuwes, Eskes, et al., 1988). Complication rates were provided for only two studies; the average was 14 percent.

Procedures for Intrinsic Sphincter Deficiency

Procedures for management of ISD include

• Sling procedures.

• Periurethral bulking injections.

• Placement of an artificial sphincter.

Sling procedures. The various sling procedures all involve placing a sling, made of either autologous or heterologous material, under the urethrovesical junction and anchoring it to retropubic or abdominal structures or both. The operation can be performed through an abdominal approach, a vaginal approach, or a combined abdominal and vaginal approach. Sling operations are often performed in women with complicated incontinence, many of whom have failed previous attempts at surgical correction. The success and complication rate should be viewed with this fact in mind.

Data from a series of 9 studies of 434 patients with fascial slings indicated that 89 percent were "cured," and 92 percent were "cured" or improved (Beck, McCormick, and Nordstrom, 1988; Deppe, Castro-Marin, Nachamie, et al., 1978; Low, 1969; McGuire and Lytton, 1978; McIndoe, Jones, and Grieve, 1987; Narik and Palmrich, 1962; Ogundipe and Rosenzweig, 1992; Richmond and Sutherst, 1989a; Ridley, 1966). Combined analysis of 298 patients from six studies that used a synthetic sling indicated that 78 percent were "cured," and 84 percent were "cured" or improved (Brieger and Korda, 1992; Bryans, 1979; Kersey, 1983; Korda, Peat, and Hunter, 1989 ).

Although the total complication rate given for the fascial sling series was higher than that for the synthetic sling series, the synthetic sling caused more severe complications, many of which were directly attributable to local effects of the sling (i.e., erosion, nonhealing of the vaginal wall, abscess, vesicovaginal fistula). Fascial slings are preferable to synthetic slings because of the lower rate of local complications.

Data from 32 patients treated with the vaginal wall sling showed that 81 percent were dry, 9 percent were improved, and 9 percent showed no improvement (Ogundpipe and Rosenzweig, 1992; Ridley, 1966). Complications included urinary retention and new onset of irritative voiding symptoms.

Periurethral bulking injections involve the injection of materials such as polytetrafluoroethylene (PTFE), collagen, or autologous fat under cystoscopic guidance into an incompetent periurethral area. Patients being considered for periurethral collagen injection require a skin test for sensitivity to the material. Injection of the material appears to be technically easier because the number of patients requiring anesthesia or sedation is less than that reported with PTFE. The longevity of PTFE versus collagen has not been studied. Urinary tract infection and transient urethral irritation are the most common side effects after periurethral collagen injection. The procedures for periurethral bulking injections are described below in the section on male incontinence caused by ISD. In women, these injections are easily performed under local anesthesia. Combined data from 15 studies of 528 women indicate that after followup for up to 2 years 49 percent of patients were "cured" (range = 8-100 percent), and 67 percent were either "cured" or improved (Beckingham, Wemyss-Holden, and Lawrence, 1992; Berg, 1973; Deane, English, Hehir, et al., 1985; Eckford and Abrams, 1991; Harrison, Brown, and O'Boyle, 1993; Kiilhoma and Makinen, 1991; Lewis, Lockhart, and Politano, 1984; Lotenfoe, O'Kelly, Helal, et al., 1993; McGuire and Appell, 1994; O'Connell, McGuire, Aboseif, et al., in press; Santarosa and Blaivas, 1994; Schulman, Simon, Wespes, et al., 1983; Smart, 1991; Stricker and Haylen, 1993; Vesey, Rivett, and O'Boyle, 1988). Complications included urgency, UTI, and urinary retention.

Although documentation of teflon migration to the lung persisted (Claes and Stroobants, 1989) adverse reactions to this material in other sites has not been found. Increasing experience with collagen has established its efficacy in the short term, but, as with Teflon injection, long-term results beyond 5 years are not available.

Artificial sphincter. Artificial sphincter placement, described below in the section on ISD in men, has been used for women with ISD. Combined data from 8 studies of 192 women with ISD treated with artificial sphincter placement indicated that 77 percent were dry, and 80 percent were "cured" or improved (Appell, 1988; Light and Scott, 1985; Diokno, Hollander, and Alderson, 1987; Duncan, Nurse, and Mundy, 1992; Karram, Rosenzweig, and Bhatia, 1993; Motley and Barrett, 1990; Parulkar and Barrett, 1990; Webster, Perez, Khoury, et al., 1992). Complications included fluid leak, loose cuff, erosion or atrophy of the cuff site, tubing kink, and infection.

Periurethral Bulking Injections

Periurethral bulking injections can improve urinary loss in men with stress incontinence. The mechanism for improvement after injection therapy is still unclear but may reflect an improvement in urethral coaptation and possibly compression. Periurethral injections are less likely to succeed in male than in female patients and in all patients who have undergone pelvic radiation therapy or who have extensive periurethral scarring. Success is more common in patients who have stress incontinence after transurethral or open prostatectomy than in those after radical prostatectomy. The literature does not support the use of bulking agents in men with severe postprostatectomy incontinence (Appell, 1994). Experience and followup are limited for treatment by injections with collagen and fat, which are absorbed by patients over time. There are no randomized studies comparing the efficacy of different materials or of injection therapy with other forms of treatment. The analysis included 9 studies of 1,005 men treated with periurethral injection (Kaufman, Lockhart, Silverstein, et al., 1984; Corrie, Rodriguez, and Thompson, 1989; Deane, English, Hehir, et al., 1985; McGuire and Appell, 1994; Osther and Rohl, 1988; Smart, 1991; Stanisic, Jennings, et al., 1991; Politano, 1992; Santarosa and Blavias, 1994). Sample sizes ranged from a minimum of 3 to a maximum of 720. The mean age was 69 years, and mean followup time was 2.0 years. The "cure" rate was reported in eight studies and ranged from 0 to 66 percent, with a mean of 20 percent. The "cure"-improvement rate was reported in nine studies and ranged from 0 to 81 percent, with a mean of 42 percent.

Complications reported with PTFE included infection, urinary retention, fever, temporary erectile dysfunction, periurethral inflammatory reaction, extrusion of the material into the urine or perineal area, and burning sensation or perineal discomfort. Particles of PTFE have been found in patients' lungs after periurethral injection of PTFE, but the exact incidence and the clinical significance of this migration are not known.

Placement of an Artificial Sphincter

Before periurethral injection therapy became available, placement of an artificial urinary sphincter was the most commonly used surgical procedure for the treatment of underactive outlet in men. Data on the current rate of comparative utilization of these two techniques are not available. Before implantation, urodynamic evaluation to confirm a stable, compliant, low pressure bladder is critical.

The analysis included 10 studies that presented data on 346 men, with sample sizes ranging from 11 to 96 (Brito, Mulcahy, Mitchell, et al., 1993; Gundian, Barrett, and Parulkar, 1989; Marks and Light, 1989; Nordling, Holm-Bentzen, and Hald, 1986; Lowe, Schertz, and Parsons, 1988; Malloy, Wein, and Carpiniello VL, 1989; Motley and Barrett, 1990; Schreiter, 1985; Wang and Hadley, 1991; Warwick and Abrams, 1990). The average age of the patients was 61.4 years. The "cure" rate was presented in nine studies and ranged from 33 to 88 percent, with a mean of 66 percent. The "cure" or improvement rate in the 10 studies ranged from 75 to 94.5 percent, with a mean of 85.3 percent.

Initial preoperative complications are mainly associated with urethral or bladder injury during implantation. Delayed complications included mechanical problems such as pump malfunction, fluid leak, or tubing kink; infection; or cuff-related site atrophy, incomplete compression, or erosion. In addition, urethral injury, pump erosion, and herniated reservoir were reported in fewer patients.

The utilization of the artificial urinary sphincter in patients after radiation therapy, cryotherapy, or pelvic fracture with urethroplasty is controversial because of concern about compromising the urethral blood supply. Also controversial is the use of intermittent catheterization after sphincter implantation. The experience of the implanting surgeon may be related to the incidence of complications.

Augmentation Intestinocystoplasty or Urinary Diversion

Various surgical procedures have been proposed for treating intractable, severe bladder instability and poor compliance. Augmentation cystoplasty with a patch of detubularized intestine is usually considered the procedure of choice. Urinary diversion with a urostomy or continent urinary diversion may be utilized as a last resort.

The risks of augmentation cystoplasty, in addition to those of any bowel surgery, include voiding difficulties that may require catheterization, mucus or stone formation, metabolic decompensation, and the rare long-range possibility of tumor formation. Contraindications for augmentation cystoplasty include renal insufficiency, bowel disease, intractable urethral disease, and inability to perform self-catheterization.

Twelve articles were reviewed (Bramble, 1982; Fenn, Conn, German, et al., 1992; George and Russel, 1991; Kockelbergh, 1991; Linder, Leach, and Rax, 1983; Lockhart, Ellis, Helal, et al., 1990; Mundy and Stephenson, 1985; Raz, Ehrlich, Zeidman, et al., 1988; Robertson, Davies, et al., 1991; Sethia, Webb, and Neal, 1991; Sidi, Becher, Reddy, et al., 1990; Strawbridge, Kramer, Castillo, et al., 1989). The studies included 403 subjects, with sample sizes ranging from 11 to 112. The mean age was 35.8 years and the known mean followup was 2.2 years (four studies). All studies included both men and women but did not separate out results. Thirty-eight percent of patients were rendered continent with spontaneous voiding (range 0-87 percent). If patients continent with CICS are included, then the mean "cure" rate is 82 percent (range 56-100 percent) and the improvement rate is 90 percent (71-100 percent).

Complications included recurrent UTI (29 patients), persistent mucus formation (11 patients), hourglass stricture at the vesicocecal anastomosis (1 patient), and complications from the artificial sphincter (3 mechanical problems, 1 pump erosion, 1 cuff erosion). The total complication rates were 47 of 87 (54 percent) for the augmentation and 5 of 22 (44 percent) for the artificial sphincter.

Bladder Denervation Procedures

Subtrigonal phenol injections (Blackford, Murray, Stephenson, et al., 1984; Ewing, Bultitude, and Shuttleworth, 1982; Nordling, Steven, Meyhoff, et al., 1986; Rosenbaum, Shaw, and Worth, 1990; Wall and Stanton, 1989) and bladder denervation (Alloussi, Loew, Mast, et al., 1984; Diokno, Hollander, and Alderson, 1987; Hodgkinson and Drukker, 1977; Lucas, Thomas, Clarke, et al., 1988; McGuire and Savastano, 1984; Opsomer, Klarskov, Holm-Bentzen, et al., 1984; Rockswald, Chou, and Bradley, 1978; Torring, Petersen, Kelmar, et al., 1988) are not presently recommended because the "cure" rates are low, and therefore the risk-to-benefit ratio is too great.

The Female Patient

In women, anatomic obstruction can result from prior anti-incontinence surgery or severe pelvic prolapse. For women who have anatomic obstruction after anti-incontinence surgery, two procedures can relieve the obstruction. Obstruction from an endoscopic needle bladder neck suspension can be relieved by cutting one of the suspending sutures. To evaluate the outcome of this procedure, five studies were reviewed (Araki, Takamoto, Hara, et al., 1990; Fowler, 1986; Huland and Bucher, 1984; Mundy, 1983; Vordermark, Brannen, Wettlauffer, et al., 1979). The combined series of 4 studies included 182 patients who had had a needle bladder neck suspension. Ten patients (5 percent) developed obstructive voiding symptoms, which were resolved after one suture was out; only one patient had a recurrence of incontinence.

The other procedure -- urethrolysis (remobilization of the periurethral adhesions) with or without resuspension -- can be used after almost any type of anti-incontinence surgery. Three studies included 41 patients who were classified by symptoms and urodynamic studies and underwent urethrolysis through a transvaginal approach in combination with a repeat endoscopic suspension regardless of the presence or absence of preoperative incontinence McGuire, Letson, and Wang, 1989; Nitti and Raz, 1994; Zimmern, Hadley, Leach, et al., 1987). Seventy-one percent of the patients had improved voiding patterns by urodynamic evaluation, and 80 percent had some improvement in symptoms, with 79 percent of those who were incontinent preoperatively exhibiting resolution of their stress incontinence.

The Male Patient

In elderly men, the most common cause of anatomic obstruction is benign prostatic hyperplasia (BPH). The treatment of BPH is multifaceted and is addressed inBenign Prostatic Hyperplasia: Diagnosis and Treatment. Clinical Practice Guideline.

Toileting Assistance

Prompted voiding and regularly scheduled toileting are effective for preventing urinary accidents in some long-term care or homebound persons (Burgio, Engel, McCormick, et al., 1988; Schnelle, Newman, and Fogarty, 1990; Schnelle, Newman, Fogarty, et al., 1991).

These toileting procedures require the assistance of staff, following a predetermined protocol. Prompted voiding programs require the discernment of wet or dry status and the ability to request toileting, whereas neither of these criteria is necessary for regularly scheduled toileting programs.

Several multisite clinical trials have studied a prompted voiding or scheduled toileting intervention strategy to decrease UI in nursing homes. One study using a 2-week baseline assessment and a 5-week intervention phase examined the effects of a prompted voiding intervention in 30 women whose average age was 85 Creason, Grybowski, Burgener, et al., 1989). Objective evaluation using the Folstein MMSE and the Katz Activities of Daily Living (ADL) measure indicated that 65 percent had severe mental impairment and 78 percent had severe physical impairment. During the intervention, staff approached subjects hourly during waking hours and prompted them to void. Results showed that the experimental group's UI (i.e., percentage of time the patient was wet when checked) was reduced from 40 to 30 percent by the end of the study. The control group's UI increased from slightly more than four episodes to over five per day. Furthermore, all but one of those in the experimental group demonstrated definite patterns of UI that were incorporated into nursing care plans for toileting, and these patients showed continued improvement in UI status. Variations in staff adherence to the intervention may decrease the magnitude of the effect of this treatment.

In a 13-week behavioral therapy program in seven nursing homes carried out by project-trained research assistants 14 hours per day 7 days per week, 133 elderly women were checked hourly and prompted to toilet (Hu, Igou, Kaltreider, et al., 1989 ). Subjects' age averaged 85 years, and most were highly physically and mentally impaired as determined by the Katz ADL instrument and Folstein's MMSE. Subjects in the treatment group decreased the number of their wet episodes from 2 to 1.4 per day. Those who had severe incontinence, more physical dependence, and less mental impairment, as well as normal CMG and bladder capacity, responded better to the training program.

Schnelle, Newman, and Fogarty (1990) enrolled 126 subjects in a multiphase prompted treatment program. Subjects' average age was 82, the majority had severe mental impairment as determined on Folstein's MMSE, and only 11 percent could ambulate independently. During phase 1, which served as baseline, research staff checked each subject hourly for 12 hours per day over a 5-day period and changed them when they were wet. During phase 2, subjects were randomly assigned to immediate- and delayed-treatment groups for 5 days. Those in the immediate-treatment group were prompted to void every hour, asked if they were wet or dry, and then apprised of the accuracy of their response. Finally, they were asked to try to toilet, but were not toileted unless they requested assistance. Subjects were praised both for staying dry and for toileting successfully. Subjects in the delayed-treatment group continued to be checked hourly. Both the immediate and delayed groups participated in the prompted toileting treatment in phase 3. Subjects responded to prompted voiding as follows: 42 percent had less than one incontinent episode per 12 hours; 35 percent decreased their UI by two episodes per 12 hours but were still incontinent one or more times over the 12-hour period; and the rest did not reduce their UI by two or more episodes per 12 hours. The most successful subjects were less than 50-percent incontinent, could recognize when they needed to void, and had higher voided volumes and lower frequency of voiding.

Colling, Ouslander, Hadley, et al. 1992) conducted a clinical trial study among 88 subjects in four nursing homes over a 37-week period. Only subjects with urge or mixed UI were included in the study. Subjects' average age was 85, and all subjects were moderately to severely impaired, physically and mentally, as measured by the Katz ADL and the Short Portable Mental Status Questionnaire (SPMSQ). The study design consisted of a 12-week baseline period, a week during which subjects' individual patterns of voiding were established using an ambulatory electronic monitoring device over a 72-hour period, and a 12-week treatment period with habit training intervention carried out 24 hours per day 7 days per week, followed by a 12-week maintenance period. Data were gathered by research staff for one 24-hour period every 3 weeks throughout the duration of the study by checking subjects every hour. The intervention had three parts: nursing staff received a 4-hour inservice on how to accomplish the treatment, an individual toileting schedule was constructed for each subject based on the electronic monitoring data, and staff and patients received positive reinforcement for adhering to the toileting program and for successful toileting behavior, respectively (Harke and Richgels, 1992). Overall, the intervention significantly decreased UI for 86 percent of the experimental group, with one-third of the group improving more than 20 percent. Incontinence decreased an average of one episode per 24 hours for the entire group. Subjects with greater bladder capacity and less mental impairment responded best to treatment. There was resistance to changing nursing staff routines because of staff shortages, and nursing staff complied with the treatment regimen only 70 percent of the time.

Schnelle, Newman, Fogarty, et al. 1991) noted that 35 percent of nursing home residents were poor candidates for prompted voiding, specifically those who had frequent incontinent episodes, did not respond to prompts, and could not cooperate with toileting. Up to 40 percent of those in the study responded well. Responders can be identified by assessing their responsiveness to a 3-day trial of prompted voiding (Ouslander, Schnelle, Uman, et al., 1995). This may be true for homebound patients as well. Toileting persons according to their individual voiding habits (Colling, Ouslander, Hadley, et al., 1992) is an excellent means of avoiding accidents. If staff is untrained or undersupervised in nursing homes, this intervention is difficult; however, it provides caregivers of homebound patients with a viable management option.

Although research indicates that toileting interventions are promising, they may be compromised, depending on compliance of nursing staff (Campbell, Knight, Benson, et al., 1991; Schnelle, 1990). Nursing staff often expect total continence when these procedures are implemented and become disappointed when this does not occur (Schnelle, Newman, and Fogarty, 1990; Colling, Ouslander, Hadley, et al., 1992; Harke and Richgels, 1992). Several studies point to the importance of monitoring staff implementation of these procedures and providing feedback regarding performance (Burgio and Burgio, 1990; Schnelle, 1990). More research is needed to determine if using predictability of the resident's need to void for an individualized or fixed toileting schedule is more cost-efficient. Future studies should address formal mechanisms of performance appraisal and incentives to increase and maintain staff behaviors that affect patient outcomes. All studies, however, report a modest saving to institutions in laundry and supply costs with the decrease in incontinent episodes.

Homebound Patients

Although management of UI for homebound populations has not been as thoroughly studied as that for those receiving care in a nursing home or ambulatory care setting, two reports focused on the use of PME programs in homebound individuals. Rose, Baigis-Smith, Smith, et al. (1990) reported the successful use of PME training augmented by biofeedback therapy for cognitively intact home care subjects. McDowell, Engberg, Weber, et al. (1994) reported two case studies of severely disabled but cognitively intact homebound individuals who were successfully treated with a biofeedback-assisted PME program. A study of homebound elderly women found that women with UI had significantly less social interaction than continent women (Breakwell and Walker, 1988). Several studies are ongoing using a variety of treatment modalities for both cognitively impaired and cognitively intact homebound persons living in both urban and rural areas.

Protective Pads and Garments

Absorbent products, classified as disposable or washable, include underpads, pant liners (shields and guards), adult diapers (briefs), and a variety of washable pants and disposable pad systems or combinations (Brink, 1990). In a survey questionnaire mailed to 36,500 Help for Incontinent People (HIP) members throughout the United States, more than 50 percent of the 10,427 respondents reported that they had used some form of protective garment (Jeter and Wagner, 1990). In addition, a survey of 512 community-dwelling elderly men and women with UI showed that 47 percent used absorbent products (Herzog, Fultz, Normalle, et al., 1989). Although there are no studies describing the prevalence of the use of protective garments among incontinent nursing home residents, results of most scheduled voiding programs in nursing homes directly indicate that disposable diapers or reusable pad and pant systems are used by their study subjects (Colling, Ouslander, Hadley, et al., 1992; Schnelle, Newman, Fogarty, et al., 1991).

Very little research has focused on these products despite their frequent use. Brink (1990; 1994) reported that of 30 studies performed between 1965 and 1987 and 14 studies between 1988 and 1993, all but 3 of the latter were performed outside of the United States. One study reported that continence care was labor intensive, requiring 25 minutes per day to change residents (Cella, 1988). In a cost study of UI in nursing homes, Hu, Kaltreider, and Igou 1990) randomly assigned 42 matched pairs of residents to treatment with either disposable or reusable cloth diapers. The total number of pads per day was 5.9 for the disposable pad users and 6.6 for those in the reusable pad treatment group. The average cost per day per person was $2.48 for disposable products and $2.61 for cloth products. Daily cost of disposable products when related to laundry cost revealed a lower daily cost for disposable products ranging from $0.44 to $0.68 or annual savings of $161-$248 per person. In addition, the skin of those using disposable products showed improvement after 1-1/2 months; the skin of subjects using cloth pads and diapers deteriorated.

At present, there are no studies regarding the advantages or disadvantages of the various products available on the market. Data also are lacking comparing reusable with disposable products. Most evaluations are carried out by product manufacturers rather than independent researchers in the health care field and thus carry a bias. This limits their usefulness in assisting nurses, long-term care administrators, and consumers in deciding what products to use. Until objective information is available regarding patient comfort, ease of application and removal, containment of urine, and control of odor, product selection will be made by trial and error or will depend on product availability in the practice setting.

External Collection Devices

External collection devices such as condom catheters can be useful for management of UI in men because they contain urine and keep the skin dry. Although these devices are preferable to indwelling catheters, which have many side effects, adverse reactions do occur. Abrasion, dermatitis, ischemia, necrosis, edema, and maceration of the penis, as well as UTI, can be problems, especially as a result of improper or prolonged use (Ouslander, Greengold, and Chen, 1987a). External catheters need to be applied carefully and changed according to product recommendation and institutional or agency policy. Careful observation of the penile skin and the urine for adverse changes is mandatory.

Although a variety of external catheters is available, no comparative studies have been conducted. There is a need for studies to address the optimal length of time external catheters should be worn before changing and to compare the incidence of UTI and other complications between catheters and external and suprapubic catheters.

Acceptable external collecting devices for women are not widely available. Pieper and Cleland (1993) report that most products developed to date have never been marketed or sold. Still others have been removed from the market because of major problems of leakage and skin abrasions. The literature regarding these products consists of small descriptive studies performed outside of the United States (Pieper and Cleland, 1993). The results of two small studies performed in the United States are encouraging (Johnson, Muncie, O'Reilly, et al., 1990; Johnson, O'Reilly, and Warren, 1989); however, comparative or randomized studies have not been conducted. Thus, conclusions as to the efficiency and safety of these devices cannot be made.

Intermittent Catheterization

Persons who have overflow incontinence can benefit from IC. Several studies indicate that IC is preferable to indwelling catheters for both men and women (Kuhan, Rist, and Zaech, 1991; Webb, Lawson, and Neal, 1990; Perkash and Giroux, 1993; Warren, 1990). However, the use of IC in the homebound patient may present considerable difficulty. Most studies have been conducted in young patients with spinal cord injuries. The section on Other Measures and Supportive Devices in Chapter 3 describes the various benefits and problems associated with IC.

Suprapubic Catheters

Suprapubic catheterization may be an acceptable alternative to indwelling catheters or IC in some patients (Barnes, Shaw, Timoney, et al., 1993; Feneley, 1983; Hebel and Warren, 1990; Stower, Massey, and Feneley, 1989). Small descriptive studies indicate that suprapubic catheterization is a promising method of emptying the chronically full bladder. Large, randomized studies need to be conducted to compare the benefits and risks of this intervention with those of IC and indwelling catheters. In addition, studies should be performed on techniques that prevent the occurrence of complications.

Indwelling Catheters

The use of indwelling catheters should be avoided with few exceptions. They may be useful as comfort measures for the terminally ill, to avoid contamination of decubitus ulcers, and for management of UI in patients with inoperable obstruction that prevents bladder emptying. In Chapter 3, the section on Other Measures and Supportive Devices describes the serious problems associated with long-term use of indwelling catheters.

Skin Care

Individuals with severe intractable incontinence are often immobile and at major risk for skin breakdown. No one would dispute the importance of good skin care and preventing skin breakdown, yet no randomized, controlled clinical trials have examined the methods most likely to achieve this goal. Recommendations about skin care are generally based on experience or made through marketing materials provided by product manufacturers (DeWitt, 1988; Fowler and Goupil, 1987; Klein, 1988). The most common directions to nurses are to keep the skin dry and keep the patient off any areas of skin breakdown.

Incontinence was noted as one of the risk factors for pressure ulcers in the AHCPR Pressure Ulcers in Adults: Prediction and Prevention. Clinical Practice Guideline ( 1992). The AHCPR guideline panels for Prediction and Prevention of Pressure Ulcers in Adults (1992) and Treatment of Pressure Ulcers (1994) made several recommendations for the prevention of skin chafing and, ultimately, ulcer formation. Because of the lack of scientific evidence, these suggestions were based on expert opinion. Suggested preventive measures included regular inspection of the skin, gentle cleansing with a mild cleansing agent immediately after soiling, avoidance of force and friction during cleansing, use of absorptive briefs or pads, use of topical barriers to protect the skin from moisture, and minimizing skin injury caused by friction and shear through proper positioning, turning, and transfer techniques.

For a full discussion of skin breakdown, with recommendations for prevention and treatment of pressure ulcers, refer to the Clinical Practice Guidelines on pressure ulcers (Bergstrom, Bennett, Carlson, et al., 1994; Panel for the Prediction and Prevention of Pressure Ulcers in Adults, 1992).