RE: "PREVALENCE OF AND RISK FACTORS FOR URINE LEAKAGE IN A RACIALLY AND ETHNICALLY DIVERSE POPULATION OF ADULTS: THE BOSTON AREA COMMUNITY HEALTH (BACH) SURVEY"

Wewelcome the recent paper from the Boston Area Community Health (BACH) Survey (1) on the prevalence and risk factors for urinary incontinence in an ethnically diverse population of adults. In his invited commentary, Thom (2) stated that the association with coronary heart disease had not been reported previously as a risk factor for urinary incontinence. However, we would like to draw your attention to our study of comorbid factors in relation to the urinary symptom syndromes, overactive bladder and stress urinary incontinence. This cohort study investigated the associations of a range of comorbid indicators for hypothetical mechanisms, including diabetic, ischemic, neurogenic, myogenic, and immunologic, using a large (>12,500) representative sample of women in the United Kingdom aged 40 years or more (3). We demonstrated a cross-sectional relation for coronary heart disease (myocardial infarction and angina) with the outcomes of stress urinary incontinence and overactive bladder, as well as a prospective association with overactive bladder incident at 1-year follow-up. This is consistent with the findings of Tennstedt et al. (1), suggesting that evidence is building for a relation between heart disease and incontinence. We also note the lack of association found with lifestyle factors. This is inconsistent with findings in our prospective studies in men (4) and women (5) and other studies that have shown associations with, for example, physical activity (6–8). One likely reason for this inconsistency is that, within the causal chain, lifestyle factors act before comorbidities. Therefore, inclusion of both within the current analysis is likely to underestimate any genuine lifestyle associations and to bias the results in favor of comorbid associations. For the purposes of comparing consistency, it would be preferable to see the results of a univariate age-adjusted analysis from the BACH Study, as well as the multivariate analysis, which depends on the particular confounding factors included in the model. Interactions between lifestyle factors may also distort the factors identified by using logistic regression and could usefully be checked by including interaction terms. We look forward to seeing the prospective results from this interesting study.