URINARY INCONTINENCE AND DIABETES IN WOMEN AGES 50-90 IN THE UNITED STATES

C. Lewis, R. Schrader, A. Many, R. Rogers

University of New Mexico, Albuquerque, NM  

OBJECTIVE:  To examine the association between urinary incontinence and diabetes in a large community-based population of women.
MATERIALS AND METHODS:  The Health and Retirement Study (HRS) is a large multistage area probability sample of households in the United States, yielding high power for all tests.  In 2000, trained interviewers collected self reported data from a population of 10,678 women aged 50 to 90 years.  Survey based ordered logistic regression was used to simultaneously analyze associations between no, mild and severe incontinence and survey based logistic regression was performed to evaluate individual odds ratios using Stata (StatCorp, Texas).  These methods account for the non-proportional sampling of the HRS by incorporating sampling weights.  Urinary incontinence, defined as losing urine on one or more days within the last month, was the dependent variable.  Incontinence severity was defined as mild if urine loss occurred on 15 or fewer days in the last month and severe if urine loss occurred on greater than 15 days.  Independent variables consisted of demographic and health data including age, ethnicity, parity, body mass index (BMI), medical disorders, and the ability to perform activities of daily living (ADL). Diabetes was dichotomized into insulin requiring (IRDM) and non-insulin requiring disease (NIRDM).  Age was analyzed in five-year intervals and BMI was calculated as kg/m2.  Pearson’s Chi-square was used to determine independence between paired variables.  Significance was defined by P < .05.
RESULTS:  Urinary incontinence was reported by 22% (2,319/10,678) of women; 57% had mild and 43% severe incontinence.  Mild incontinence was associated with increasing BMI, hypertension, stroke and arthritis but not diabetes as independent risk factors.  Severe incontinence was associated with increasing age, increasing BMI and all medical disorders including both NIRDM and IRDM.  Using ordered logistic regression, NIRDM was no longer associated with urinary incontinence, however, IRDM was the second strongest independent risk factor for severe incontinence (OR 1.63; 95% CI 1.28, 2.09).  Because ADL were strongly associated with both incontinence as well as other independent variables, this variable was excluded from the analysis.  Parity was not associated with incontinence in this age group (P = NS).
CONCLUSION:  Diabetes is independently associated with severe urinary incontinence in women ages 50 to 90 years, independent of patient BMI, medical comorbidities, age, or parity.  Mild incontinence is not associated with diabetes.

Key Words:  diabetes, urinary incontinence

Disclosure - Grant/Research: R. Rogers, Pfizer, Wyeth-Ayerst; Educational Support: R. Rogers, Yamanouchi Pharma America, Proctor & Gamble.