Improving Treatment of Urinary Incontinence

More than 15 million persons in the United States have urinary incontinence,¹ a condition more prevalent than diabetes mellitus. Urinary incontinence not only causes considerable medical and psychosocial morbidity but it also engenders enormous costs. Conservative projections for costs related to diagnosis and treatment of urinary incontinence exceed $26 billion annually,² more than is expended by Medicare on dialysis (O. Cohen, oral communication, November 1998) and coronary bypass grafting combined.³ Moreover, these costs apply only to individuals older than 65 years, who constitute less than half of those with urinary incontinence. Unfortunately, most physicians have received little education about incontinence, fail to screen for it, and view the likelihood of successful treatment as low.¹ Thus, it is fitting that the study by Burgio et al, which examines the efficacy of biofeedback for urge incontinence, should appear in JAMA.⁴

Urinary incontinence has a broad differential diagnosis, including causes within and beyond the lower urinary tract.^{1 ,5} Within the urinary tract, overactive bladder (detrusor overactivity) is the most common cause of urinary incontinence in the elderly and also contributes to urine leakage in many women with stress incontinence. Although behavioral interventions are recommended by most authoritative groups,^{1 ,6 - 7} pharmacological therapy remains the most frequently used treatment for patients with overactive bladders. Several factors may contribute to this practice. There are multiple behavioral techniques and protocols, but their comparative efficacy is unknown. Because technical aspects cannot be detailed sufficiently in reports of clinical trials, behavioral interventions also are difficult to replicate in practice. One behavioral technique, biofeedback, has been even less widely used for urge incontinence because it often has required repeated instrumentation of the bladder and urinary sphincter. Moreover, despite the expertise and time entailed, behavioral techniques are poorly reimbursed. By contrast, pharmacotherapy works more quickly and also requires no behavioral expertise, less physician time, and less patient participation. Nonetheless, although drugs help most patients, no drug restores continence to the majority. Furthermore, all of the agents currently used engender adverse effects, expense, and inconvenience,¹ and most must be taken several times daily and indefinitely. Thus, an equally or more effective one-time intervention would be welcome.

The study by Burgio et al⁴ demonstrates that a less-invasive biofeedback approach can achieve these goals. The study also serves as a rich source of information and provides valuable lessons for clinicians. The investigators' decision to begin therapy with oxybutynin at 2.5 mg 3 times daily was wise. The effect of oxybutynin was equivalent to that achieved in trials using higher dosages, but it caused far fewer adverse effects and a lower rate of subject attrition.¹ Equally important, efficacy continued to increase beyond 2 weeks, longer than previously reported, but consistent with other recent data.^{8 - 9} Thus, clinicians should avoid escalating dosages of oxybutynin too quickly or abandoning pharmacotherapy too soon.

The decision to use oxybutynin at this dosage, frequency, and duration should have enhanced its benefit, thereby making the biofeedback findings even more impressive. However, it is premature to conclude that biofeedback is superior for patients with urge incontinence. Half the subjects had a component of stress incontinence, which responds well to biofeedback but not to a bladder relaxant medication and which may even worsen as bladder capacity and residual urine volume increase. In addition, the decision to include patients with residual urine volume as high as 200 mL may have affected the results, since such patients may be less responsive to a drug that further impairs detrusor muscle contractility. This approach also may explain the higher proportion of subjects who reported "inability to void" while receiving therapy. Thus, oxybutynin may work as well as or even better than biofeedback for subjects who have pure urge incontinence and low residual urine volume.

Nonetheless, the efficacy of biofeedback was impressive, and it was equal or superior to oxybutynin on every measure of outcome in this group of patients. This finding highlights another strength of the trial: inclusion of patients' perspectives although they may not have been gathered blindly. Patients taking oxybutynin correctly perceived a 68% reduction in leakage and, in 80% of cases, described themselves as better or much better. Yet, only half of these subjects were willing to continue oxybutynin indefinitely and 75% wished to receive another form of therapy (proportions that are not much better than those for individuals receiving placebo). These data contrast strikingly with results from individuals randomized to biofeedback, nearly all of whom were comfortable continuing it and only 14% of whom desired another therapy. Furthermore, since biofeedback is effective for subjects with mixed (ie, urge plus stress) incontinence, and for those with pure urge incontinence, it may be more widely applicable. The feasibility of biofeedback is further enhanced by the fact that a much less invasive, staged approach was as effective as the traditional method.

The traditional approach to urge incontinence is bladder training, which involves increasing the voiding interval and teaching strategies to cope with urgency.¹ By avoiding the expensive equipment and expertise required for biofeedback, bladder training is easier to implement and less expensive, but its comparative efficacy is unknown. Equally important is whether adding biofeedback for those who fail bladder training would be as efficacious as using biofeedback for all patients with incontinence. Similarly, it would be useful to know whether combining biofeedback with pharmacotherapy would provide additive or even synergistic effects. Also, it is important to determine whether biofeedback works as well for subjects with severe incontinence as for those with milder forms, how durable the effects are, and how often "refresher courses" will be required.

Because the mechanisms mediating the effects of biofeedback are unknown, whether or how it could be further optimized or streamlined cannot be surmised. Because all the therapists were nurse practitioners who were trained and supervised by Dr Burgio, it is difficult to determine the generalizability of biofeedback to nurses without such training, to different types of therapists, and to therapists outside a clinical trial setting. Additionally, because participants were all volunteers and comprised fewer than half of those recruited over several years, the technique's efficacy in routinely encountered patients is less certain. And, as the authors note, the study subjects were relatively young, cognitively intact, and relatively free of comorbidity, all factors that limit applicability of this technique to other groups of patients.

Nonetheless, with other research on incontinence, this well-conceived trial underscores the fact that therapeutic nihilism is no longer tenable. Given the hidden nature of the condition, primary care clinicians should redouble their efforts toward identifying the majority of patients in their practice who, unbeknownst to the physician, are incontinent and toward addressing the causes related to medications (eg, anticholinergic, sedative, loop diuretic, and adrenergic agents and angiotensin-converting enzyme inhibitors) and diseases outside the urinary tract (eg, heart failure) that they already know how to treat.^{1 ,6 ,10}

The increasing array of simple algorithms and guidelines^{1 ,5 - 7 ,10} also makes empiric management of many patients feasible for primary care clinicians. Even simple interventions can be effective. Moreover, all studies document a high placebo response. Although the mechanisms have not been well studied, they may include patients' learning that incontinence is not untreatable, drinking a little less, voiding a little more often, and becoming more aware of bladder fullness. Importantly, the role of a concerned caregiver's encouragement should not be minimized. Despite the need for continued research to advance and refine treatment of urinary incontinence, at present physicians probably should be less concerned about the relative benefits of various types of behavioral interventions or drugs (all of which appear to have equivalent efficacy)¹ and more concerned that every patient receives some individualized intervention.