BMI, waist circumferences and urinary incontinence in older women compared with older men: findings from three prospective longitudinal cohort studies

Background
Obesity and urinary incontinence (UI) among older people, particularly older men, are yet to be fully explored.
Utilising multiple nationwide prospective longitudinal cohorts representative of the US, UK, and European
samples, we examined the association of body mass index (BMI) and waist circumference (WC) with UI among
both older women and men.
Methods
We derived the data from the Health and Retirement Study (HRS, 2010-2018), the English Longitudinal Study of
Aging (ELSA, 2011-2019), and the Survey of Health, Ageing and Retirement in Europe (SHARE, 2004-2010) that
surveyed UI. Participants were asked if they had experienced urine leakage within the past 12 months (HRS and
ELSA) or within the past six months (SHARE). The measure of obesity was based on BMI and WC. We
employed a random-effect logistic model to associate BMI and WC with UI, adjusting for covariates including
age, race, education, residence area, marital status, number of children, smoking, drinking, hypertension,
diabetes, cancer, stroke, functional ability, and cognitive impairment. We visualised the associations by using
restricted cubic spline curves.
Findings
A total of 200,717 participants with 718,822 observations (207,805 in HRS; 98,158 in ELSA; 412,859 in SHARE)
were included in the baseline analysis. The 12-months prevalence of UI among female and male participants
were 15.6% and 6.6% in the HRS, 10.6% and 4.4% in the ELSA. The 6-months prevalence of UI were 2.8% and
1.4% in the SHARE’s female and male participants. Compared to those without UI, both female and male
participants with UI demonstrated a higher BMI and WC. Among females, the fully adjusted models showed
linear associations between BMI, WC, and UI (Ps<0.001) in three cohorts. However, we observed U-shaped
associations of BMI, WC with UI among males. The lowest likelihood of having UI was found among male
participants with a BMI between 24 and 35 kg/m2
.
Interpretation
Findings from our study revealed that the associations of obesity indices with UI varied among older men
compared to older women. As a result, weight loss interventions could be applied to older women rather than
older men as a means of treating UI. Interventions aimed at preventing UI among older adults must take sex
into account.