Asymptomatic Bacteriuria and Incontinence

To the Editor: — Abrutyn and his colleagues1 have contributed greatly to our understanding of asymptomatic bacteriuria in the elderly through a series of studies reviewed in their recent editorial. I would like to comment on their conclusions regarding the relationship of bacteriuria to urinary incontinence. Citing data from one of their studies in which symptoms of 72 subjects were assessed when they were bacteriuric versus nonbacteriuric,2 the authors conclude that “… incontinence or frequency of change in incontinence in patients with bacteriuria is unrelated to otherwise asymptomatic urinary tract infection” and that “an increase in incontinence should not be used as an indication to treat nondysuric subjects with bacteriuria.”1 They support this conclusion by claiming a power of .995 to detect “clinically relevant differences in symptoms.” While these conclusions may be justified in the population studied, they cannot be readily extrapolated to the population of patients with urinary incontinence (as opposed to bacteriuria) a geriatrician is likely to encounter. Among their 72 subjects, the mean symptom scores for incontinence and associated symptoms such as frequency and urgency were, as they note in their article, skewed toward the lower end of the scale (range of means 1.08-1.57 on a scale ranging from 1-6).2 Thus, these subjects had these symptoms on average between never (1) and less than once per week (2). In sharp contrast, most of the patients we and others have evaluated for incontinence in outpatients and nursing home settings have episodes of incontinence several times a week to several times a day.3, 4 Although data are lacking, it is entirely possible that, among elderly patients with more frequent incontinence and/or very bothersome frequency and urgency, eradicating otherwise asymptomatic bacteriuria may have a clinically relevant impact on their symptoms. In a study of 15 incontinent nursing home patients, we were unable to document any dramatic reduction in incontinence frequency after eradication of bacteriuria, but symptoms were not quantitated precisely and our statistical power was low.5 It is also possible that there is a subgroup of patients who have frequent or persistent bacteriuria over a period of years which, in combination with detrusor instability, contributes to a small capacity, highly unstable bladder with intractable urge incontinence. Thus, while I think the work of Abrutyn and his colleagues is excellent and their conclusions valid for patients with minimal incontinence symptoms, I suggest they are premature for the more typical incontinent geriatric patient. Until further data are available, I still think it prudent to recommend that in the initial assessment and management of an incontinent elderly patient, and in the management of an incontinent patient whose incontinence suddenly worsens without other obvious causes, that bacteriuria, if present, be eradicated to determine the impact of sterilizing the urine on incontinence and associated symptoms. Dr Ouslander's comment was referred to the authors and their reply follows.