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

doi:10.21203/rs.3.rs-2441866/v1

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.

Health sciences/Diseases

Health sciences/Health care/Public health/Epidemiology

Evidence before this study

We searched PubMed for studies related to obesity and urinary incontinence (UI) in the older population without any time restriction. We used the following keywords "Body Mass Index", "Waist Circumference", "Obesity", "Urinary Incontinence", "Aged", "Female", and "Male". In PubMed, 302 articles were collected using all of these keywords, in which 47 articles studied the association of obesity indices and UI. However, only seven studies considered sex differences regarding the relationship between UI and obesity indices. Specifically, there are four cross-sectional studies and three longitudinal studies analysed the UI-obesity relationships by sexes. There has been only one study, based on the Boston Area Community Health (BACH) Survey, that reported that central obesity increases the odds of weekly urinary leakage for women, but not for men. However, there has been little prospective longitudinal research conducted to investigate the sex differences in associations between body mass index (BMI), waist circumference (WC) and UI, limiting the effectiveness of interventions for UI in older women and men.

Added value of this study

By utilising multiple nationwide prospective longitudinal cohorts representative of the US, UK, and European samples (HRS, ELSA, and SHARE), we provide a unique insight and comprehensive assessment of the association of BMI and WC with UI among older women compared to men. As far as we know, this is among the first longitudinal study with large sample size to investigate the association of BMI and WC with UI among older adults by taking sex into consideration. According to our study, after controlling for a variety of confounders, BMI is linearly associated with higher prevalence of UI among older women, while it has a U-shaped association among older men.

Implications of all the available evidence

Previous studies suggested weight reduction for UI interventions among older people. However, our study revealed that the association of obesity indices with UI varied among older men compared to older women. We observed linear associations between BMI, WC, and UI in women, while we observed U-shaped associations among men. The lowest likelihood of having UI was found among male participants with a BMI between 24 and 35 kg/m². Weight loss intervention could therefore only be applied to older women rather than to older men as a way to treat UI. Developing interventions to address UI among older adults should take sex into account.

A rapidly ageing population, coupled with age-related illnesses, is a pressing concern around the globe. Urinary incontinence (UI) is a common geriatric syndrome that has been associated with adverse health outcomes and poor quality of life. It is estimated that 17 to 55% of older women and 11 to 34% of older males had UI, which places a substantial financial and care burden on society and families.¹ The direct costs of UI in 1995 were estimated at $26.3 billion for those older than 65 years in the USA.² Furthermore, the annual cost-of-illness estimate for urgency urinary incontinence (UUI) in Canada, Germany, Italy, Spain, Sweden, and the United Kingdom was €7 billion in a 2005 multinational study.³ UI is also associated with a reduced quality of life (QoL) and self-esteem for those who have a stressful, incapacitating condition and a social isolation element, with sufferers reporting embarrassment, distress, and depression⁴. The caregivers of older adults with UI are also more likely to feel a strain both physically and mentally than those without UI.⁴ Thus, identifying and managing of modifiable risk factors of UI are of paramount importance.

In spite of the fact that obesity is well documented as a modifiable risk factor for the development of UI, most epidemiological, biological, and clinical research focuses on older women rather than older men.^5–18 Meanwhile, despite some studies focused on both sexes,^19–26 few longitudinal research targeting the older population investigated the sex heterogeneity in patterns of associations between UI and obesity.^21,22 Moreover, there are varied obesity indices including Body mass index (BMI) and waist circumference (WC). BMI estimates general obesity established by the World Health Organization (WHO), while WC better describes central obesity, although it cannot discriminate visceral fat from subcutaneous fat. However, the relationship between obesity indices, i.e., BMI and WC, and UI in older men and women has not been thoroughly examined in prior longitudinal studies. Furthermore, although the EpiLUTS study comprised research subjects from three countries,²¹ samples in most prior findings were based on samples from one country or region, which led to poor generalization of the study's findings.

Extant evidence has reported an “obesity paradox” in which BMI is negatively associated with mortality, which indicates that being normal weight and overweight could be more beneficial to health than being underweight.²⁷ This paradoxical association of obesity with mortality has been inconsistent due to a discrepancy between BMI and central obesity.²⁸ However, whether this paradox applies to older women and men in terms of UI was largely unknow. Therefore, study with a representative sample of older men and women from a global perspective should be conducted to reexamine the associations between obesity indices, including BMI and WC, and UI.

By utilising multiple nationwide prospective longitudinal cohorts representative of the US, UK, and European samples, we contribute to the literature by analysing the association between BMI, WC, and UI among older women compared to men. This study aimed to illustrate the sex difference in impact of obesity on UI and provide indication of evidence for future intervention.

Study design and participants

Data were accessed from three international cohorts of aging: Health and Retirement Study (HRS), English Longitudinal Study of Ageing (ELSA), Survey of Health, Ageing and Retirement in Europe (SHARE), which were used to provide the representative sample and comparable measures on BMI, WC, UI, and other covariates, covering 21 countries including both developed and developing ones on two continents. Given the availability of UI measurements and similar time ranges, we used data from the following time period in this analysis: 2010–2018 for HRS, 2010–2018 for ELSA, and 2004–2010 for SHARE. Participants who were younger than 50 years old were excluded. Following the exclusion criteria, 121,450 SHARE participants with 360,800 observations, 19,791 ELSA participants with 98,158 observations, and 42,132 HRS participants with 207,805 observations were analyzed.

Data collection

Data on urinary incontinence were collected through questionnaires. In HRS and ELSA, UI was assessed by asking: “During the last 12 months, have you lost any amount of urine beyond your control?” In SHARE, the dependent variable was constructed based on the question: “For the past six months at least, have you been bothered by incontinence or involuntary loss of urine?” The responses ranged from 0 = ‘‘No’’, which meant the respondent hadn’t been bothered by any urinary incontinence in the time frame corresponding to the question, to 1 = ‘‘Yes’’, which was considered as the respondent had been bothered by any urinary incontinence in the time frame corresponding to the question. The measures of obesity indices were based on BMI and WC. The BMI (kg/m²) was derived by dividing weight by the square of height. The measurement of WC was reported in centimeters. The obesity indices (BMI and WC) were then equally divided into six quantiles when analyzing the association of obesity with UI prevalence.

In addition, covariates for sociodemographic status (age, race, residence area, marital status, number of children), lifestyles (smoking, drinking), the history of the disease (hypertension, diabetes, cancer, stroke) as well as educational attainments, functional ability, and cognitive impairment were added. The race was divided into white or not. Residence area was attributed to whether living in rural or urban areas. Marital status was imputed to married/partnered or not. We used the harmonized education attainments classification, namely lower than the secondary/upper secondary and vocational training/tertiary. Physical active participants were defined as those without difficulty using the toilet. Cognitive impairment was defined as a score of 8 or below on a 35-point scale for HRS^29,30, while a score between 0 and 11 on a 27-point scale for the ELSA³¹. In SHARE, a summary cognitive function score of averaged z-scores of the verbal fluency, immediate and delayed recall, orientation, and numeracy for individuals was constructed using the mean and standard deviation of the first wave, where respondents were labeled as having cognitive impairment if their score fell within the lowest decile³².

Statistical analysis

Separate descriptive characteristics of the observations in the HRS, ELSA, and SHARE were provided, in which the mean (SD) or the median (IQR) was used for continuous variables, and number (percentage) was used for categorical variables. The observations of BMI, WC, and other covariates were then displayed by whether with UI in both females and males for HRS, ELSA, and SHARE, respectively. Additionally, the statistically significant difference between those with UI and without UI in BMI, WC, and other covariates was reported to identify the driving factors of UI by sex among older adults in the HRS, ELSA, and SHARE.

All statistical analyses were conducted by females and males, seperately. Chi-squared tests were used to evaluate differences among those with UI and without UI for categorical variables, whereas t-tests were used for continuous variables. We employed random-effect logistic models to control the individual-specific characteristics that might change in different waves and to construct odds ratios (ORs) and 95% confidence intervals (CIs) to analyze the relationships between obesity indices and the prevalence of UI. The dependent variable is the prevalence of UI. The sociodemographic status, lifestyle factors, history of diseases, educational attainments, functional ability, and cognitive impairment were incorporated as the co-variates. Three models were conducted in the analyses by sex. Model 1 was a crude model. In Model 2, we controlled for age, race (except for the SHARE), educational attainments, residence area (except for the ELSA), marital status, and the number of children. Model 3 was further adjusted for current smoking, ever-smoked, alcohol consumption, physical activity, hypertension, diabetes, stroke, cancer, and cognitive impairments, which was regarded as the fully adjusted model. Statistical significance was considered to be P < 0.05. Accordingly, the restricted cubic spline (RCS) curves were plotted to visualize the associations of UI with BMI by sex in HRS, ELSA, and SHARE, as well as the associations of UI with WC by sex in HRS rather than in ELSA and SHARE due to only one wave for WC statistics in ELSA and data unavailability of WC in SHARE.

Additionally, subgroup analyses were performed for each age group (50–59/60–69/70–79/80 over) within each gender (Female/Male). The RCS curves were shown for relationships between UI and BMI for each age group in HRS and SHARE, as well as the associations between UI and WC for each age group in HRS. The subgroup analyses in ELSA, however, were not conducted because there was no significantly discernible gender difference in the associations. All statistical analyses were performed by Stata 17.0.

The descriptive characteristics of observations by sex in the HRS, ELSA, and SHARE are presented in Table 1. The 666,763 participant observations (HRS 207,805; ELSA 98,158; SHARE 360,800) were collected for the baseline survey. The median ages included in this analysis for HRS, ELSA, and SHARE studies were 66, 67, and 64 for females, and 65, 67, and 64 for males, respectively. The white was 72% in HRS and 95.4% in ELSA for females, as well as 73.6% in HRS and 95.4% in ELSA for males. More importantly, the females and males with UI were responsible for 15.6% & 6.6% in the HRS study, 10.6% & 4.4% in the ELSA study, and 2.8% & 1.4% in the SHARE study, respectively. The average BMI in the surveys of HRS, ELSA, and SHARE was 28.6 (± 6.5), 28.20 (± 5.6), and 26.5 (± 4.8) for females, and 28.5 (± 5.3), 28.2 (± 4.4), and 27.0 (± 4.0) for males, respectively. Meanwhile, the average WC for females and males were 99.3 (± 14.5) and 105.2 (± 12.4) in HRS, and 92.0 (± 13.6) and 102.2 (± 11.9) in ELSA.

The BMI, WC, and covariates of observations by sex in analyses were displayed in Table 2. Across three cohort studies, there are 29,187 female observations with UI and 10,241 male observations with UI being incorporated. Notably, males and females with and without UI exhibited significant differences in baseline characteristics. Specifically, females with UI showed significantly higher median age, the proportion of white, married and partnered, and physically inactive, the prevalence of chronic diseases and cognitive impairments (except HRS study), and higher BMI and WC, compared to females without UI. Comparably, most baseline characteristics for males with and without UI indicated similar trends with females but with less statistical significance in the race, residence area, marital status, the number of children, variables for lifestyle, and obesity indices.

The association between BMI, WC and the prevalence of UI was outlined in Table 3. The fully adjusted model for females shows a linear trend of monotonically increasing prevalence of UI as BMI and WC rise. Specifically, the first quantile had the lowest prevalence of UI and the sixth quantile had the significantly highest prevalence of UI compared to participants in the second quantile of BMI. Similarly, compared to participants in the second quantile of WC, those in the first quantile had the lowest prevalence of UI, whereas the significant highest prevalence of UI was observed in the sixth quantile.

However, a U-shaped trajectory in the UI prevalence can be found with rising BMI and WC in the fully adjusted model for males, even though the linear associations between obesity indices and UI were less pronounced than in women. In comparison to those in the second quantile of BMI, the lowest prevalence of UI was found in the fourth quantile for HRS, the third quantile for ELSA and SHARE study, and those in the sixth quantile of BMI had the significant highest prevalence of UI. In terms of WC, we discovered that those in the fourth quantile of WC for both HRS and ELSA experienced the lowest prevalence of UI. However, those statistics were not statistically significant. Additionally, although the statistic in the ELSA study did not reach statistical significance, the statistic for the HRS in the sixth quantile of WC had the significant highest UI prevalence, compared to those in the second quantile.

Additionally, we draw the restricted cubic splines (RCS) curve to visualize the association between UI and BMI by sex in HRS, ELSA, and SHARE, as shown in Fig. 1. Besides, the association between UI and WC by sex in the HRS study was also plotted in Fig. 2. The shapes of RCS curves showed heterogeneity across gender in HRS and SHARE studies. Specifically, among three longitudinal cohort studies, the RCS curves for females concerning the association of UI with both BMI and WC were almost monotonically increasing, and the slopes of these curves were relatively larger at higher BMI and higher WC. However, for males in HRS and SHARE, it was obvious that a “U-” shaped curve was traced with increasing OR of the prevalence of UI on the Y axis and increasing BMI or increasing WC on the X axis, even though this pattern was not apparent for the figure in ELSA study.

To assess the gender and age heterogeneity of UI on BMI and WC, RCS curves were also drawn in the Supplementary information to visualize the UI-BMI association and UI-WC association in different age groups within each gender in HRS and SHARE. As can be seen, we found that both the UI-BMI association and the UI-WC association were more pronounced in the 60–69 age group for males in HRS and SHARE.

By utilising three longitudinal cohorts representative of the US, UK, and European samples, this study offered unique insight and extensive evidence on the relationship between obesity and UI in older men and women. Specifically, the higher prevalence of UI in older women is linearly correlated with higher BMI and WC, while there was a U-shaped association between BMI and UI in older men. Our study indicated a revisiting of association between obesity and UI among older men compared to women.

Prior epidemiological studies reported that obesity is a modifiable risk factor for UI development in older women or men^{5–8, 10,11,14,15,17,18}, and some clinical trials provided evidence that weight reduction can reduce the incidence of UI^9,12,13,16. However, only seven studies—three longitudinal and four cross-sectional—focused on men and women seperately regarding the relationship between UI and obesity indices^19,21−26. In particular, the Subak et al.’s cohort study based on the longitudinal assessment of bariatric surgery 2 revealed that obese participants who lose weight via surgical interventions experience reduced UI²⁴, which was in conflict with the male analysis of our study since this study focused on severely obese people with BMI (kg/m^2) varied from 41.6 to 52.8, whereas our study subjects concentrated on the general elderly population.

In addition, the cohort study by Tsui et al. from a British birth cohort study demonstrated that increased BMI at age 60–64 years was an independent risk factor for urgent urinary incontinence (UUI) in men and women²⁶. Another article by Tennstedt et al. examining the population-based adults aged 30–79 years based on the Boston Area Community Health (BACH) Survey investigated that WC was a risk factor for leakage in women rather than men, with the odds of weekly leakage increasing by 15 percent with each 10-cm increase²⁵. In general, our findings concurred with these investigations. However, the participants in these two research specifically targeted the elderly in Britain and Boston, respectively, rather than the elderly worldwide. Second, only UUI, rather than general UI, was used as the outcome assessment in the Tsui et al. study. Furthermore, we could only be enlightened if obesity was related to UUI or UI rather than the movement of UI as WC grew and the tendency of UUI with growing BMI. Also, Tsui et al.’s study did not take central obesity into consideration.

Moreover, even though there were four cross-sectional studies also considering both genders in regard to the UI-obesity associations^19,21−23, one of them from the French 3C study that concentrated on the elderly subjects aged 65–101 suggested that the relationship tended to be linear for UI and obesity in females, whereas the pattern that a higher risk of UI for both underweight and obese subjects was found in males²³. This study was in line with our findings but concentrated primarily on French and did not take into account the obesity indices for central adiposity.

In terms of mechanisms, several biological studies have pointed out that obesity is associated with low-grade systemic inflammation and the release of pro-inflammatory cytokines, generating reactive oxygen species and oxidative stress, which alters collagen metabolism. At the same time, intra-abdominal pressure increases with obesity that bears down on pelvic tissues, causing chronic strain, stretching, and weakening of the muscles, nerves, and other structures of the pelvic floor, which thus leads to negatively affecting pelvic organ function. These physical and biochemical stresses predispose obese geriatric patients to develop UI⁸. Comparably, being underweight also resulted in a higher prevalence of UI for males, according to our study. This could be as a result of the several mechanisms through which being underweight may cause disability, especially a lack of physical activity and a higher risk of falls, increasing the chance of UI incidence^33–35. The disability would, in turn, lead to underweight due to partial loss of life skills, thus less accessibility to get nutrition through cooking and etc. Nonetheless, the association between being underweight and UI is still debatable and appears to be weaker than the relationship with obesity. The discovery of permanent rather than reversible alterations to the prostate caused by weight loss once men approach old age¹¹, as well as postmenopausal hormonal changes in women, may be viable explanations in light of the mechanism underlying the gender difference. The type of incontinence, however, was not taken into account in this study's outcome measures, which was likely the cause of the less substantial linear association in men compared to women.

This study delivers distinctive and substantial policy implications. On the one hand, insufficient attention has been paid to the relationship between UI and obesity indices in men. For instance, earlier research has focused less on the patterns between UI and obesity in males than in women. Male participants were not even asked in the Korean Longitudinal Study on Aging (KLoSA) questionnaire if they had ever experienced UI within a certain time period. Although men are less prevalent than women to report UI, it is nevertheless essential to revisit the UI-obesity association in men to better corroborate our findings and thus further improve our present UI prevention and treatment practices. On the other hand, the majority of developed countries examine UI and fecal incontinence separately due to crucial diversity in pathogenic mechanisms, risk factors, and treatment approaches, which HRS, ELSA, KLoSA, and SHARE could support. Contrarily, some nationwide cohort studies of aging, including the China Health and Retirement Longitudinal Study (CHARLS), did not differentiate these two disorders clearly and instead created a single question to target them explicitly. Therefore, it was imperative to distinguish UI and fecal incontinence when developing survey questionnaires to advance pertinent intervention research.

There are several strengths. Firstly, our study's excellent generalizability can be attributed to the utilization of three sizable, diverse, well-characterized, cross-cultural longitudinal studies that included older adults from 22 developed and developing countries on two continents. Second, the validity of our study was strengthened by the use of nationally representative samples for participant recruitment, standardization of the three surveys for database comparisons, and a more extended follow-up period of almost ten years. Third, random-effect logistic models regulate individual-level traits that could alter over longitudinal data waves and thus reducing the likelihood of misestimation. Finally, this study provides a basis for demonstrating the obesity paradox in the relationship between obesity and urinary incontinence in men, showing that those older men with both lower and higher BMI and WC tend to report higher UI prevalence compared to those with BMI or WC in the normal range, although this relationship is not particularly significant.

However, our study has limitations. Firstly, due to data availability for UI measurements, the present study has an early but narrower time period for SHARE compared to HRS and ELSA, which could cause an overestimation or underestimation of the underlying relationships between obesity indices and outcomes. Secondly, this study's UI, BMI, and WC measurements rely on self-report rather than clinician diagnosis, which might be a limitation as responses could be influenced by recall and social desirability biases. Additionally, even though we considered numerous potential confounders that had a significant impact on our findings, there may still be confounding variables that are unaccounted for due to limited information, such as ethnicity in SHARE and residential areas in ELSA. A previous study provided insights that indicators of central obesity, including WC and waist-to-hip ratio, appear to be more sensitive than BMI (a proxy for body fat) in explaining the association between obesity and urine leakage^22,25. However, our study did not fully consider the indicators of central obesity due to a lack of comparable data. Finally, the type of incontinence was not specified in the longitudinal cohort, which poses an unavoidable obstacle to further research on its mechanisms.

In summary, the associations between obesity indices and UI is different in older women compared to older men. Therefore, weight loss as a treatment for UI can only be applied to older women but not necessarily to men. With this in mind, developing interventions to address UI among older adults should take sex into account.

Data availability

Original survey datasets from HRS, ELSA, and SHARE are freely available to all bona fide researchers. The data that support the findings of this study are available from the GATEWAY TO GLOBAL AGING DATA (https://g2aging.org/).

Competing interests

The authors declare that they have no competing interests.

Funding

This study received no external funding.

Acknowledgement

We thank the GATEWAY TO GLOBAL AGING DATA for providing the harmonised data. We thank the Dr. Rui Yang and Mingzhi Yu from Peking University for gaving us technical support in data analyses and interpretation. We also thank the support from the Healthy Aging Group of the China Cohort Consortium (see http://chinacohort.bjmu.edu.cn/).

Contributors

Yao Yao designed the study. Data collection and analysis was led by Xiyin Chen. The manuscript was written by Xiyin Chen and verified by Shaoxiang Jiang. All authors have edited and reviewed the manuscript. All authors had full access to the data and accept the responsibility to submit the manuscript

Thom D. Variation in estimates of urinary incontinence prevalence in the community: effects of differences in definition, population characteristics, and study type. J Am Geriatr Soc 1998; 46(4): 473-80.
Wagner TH, Hu TW. Economic costs of urinary incontinence in 1995. Urology 1998; 51(3): 355-61.
Milsom I, Coyne KS, Nicholson S, Kvasz M, Chen CI, Wein AJ. Global prevalence and economic burden of urgency urinary incontinence: a systematic review. Eur Urol 2014; 65(1): 79-95.
Tamanini JT, Santos JL, Lebrao ML, Duarte YA, Laurenti R. Association between urinary incontinence in elderly patients and caregiver burden in the city of Sao Paulo/Brazil: Health, Wellbeing, and Ageing Study. Neurourol Urodyn 2011; 30(7): 1281-5.
Hunskaar S. A systematic review of overweight and obesity as risk factors and targets for clinical intervention for urinary incontinence in women. Neurourol Urodyn 2008; 27(8): 749-57.
Subak LL, Richter HE, Hunskaar S. Obesity and urinary incontinence: epidemiology and clinical research update. J Urol 2009; 182(6 Suppl): S2-7.
Brown JS, Seeley DG, Fong J, Black DM, Ensrud KE, Grady D. Urinary incontinence in older women: who is at risk? Study of Osteoporotic Fractures Research Group. Obstet Gynecol 1996; 87(5 Pt 1): 715-21.
Doumouchtsis SK, Loganathan J, Pergialiotis V. The role of obesity on urinary incontinence and anal incontinence in women: a review. BJOG 2022; 129(1): 162-70.
Subak LL, Wing R, West DS, et al. Weight loss to treat urinary incontinence in overweight and obese women. N Engl J Med 2009; 360(5): 481-90.
Wing RR, Creasman JM, West DS, et al. Improving urinary incontinence in overweight and obese women through modest weight loss. Obstet Gynecol 2010; 116(2 Pt 1): 284-92.
Bauer SR, Grimes B, Suskind AM, et al. Urinary Incontinence and Nocturia in Older Men: Associations with Body Mass, Composition and Strength in the Health ABC Study. J Urol 2019; 202(5): 1015-21.
Breyer BN, Phelan S, Hogan PE, et al. Intensive lifestyle intervention reduces urinary incontinence in overweight/obese men with type 2 diabetes: results from the Look AHEAD trial. J Urol 2014; 192(1): 144-9.
Brown JS, Wing R, Barrett-Connor E, et al. Lifestyle intervention is associated with lower prevalence of urinary incontinence: the Diabetes Prevention Program. Diabetes Care 2006; 29(2): 385-90.
Donnelly DW, Donnelly C, Kearney T, et al. Urinary, bowel and sexual health in older men from Northern Ireland. BJU Int 2018; 122(5): 845-57.
Subak LL, Johnson C, Whitcomb E, Boban D, Saxton J, Brown JS. Does weight loss improve incontinence in moderately obese women? Int Urogynecol J Pelvic Floor Dysfunct 2002; 13(1): 40-3.
Subak LL, Whitcomb E, Shen H, Saxton J, Vittinghoff E, Brown JS. Weight loss: a novel and effective treatment for urinary incontinence. J Urol 2005; 174(1): 190-5.
Townsend MK, Curhan GC, Resnick NM, Grodstein F. BMI, waist circumference, and incident urinary incontinence in older women. Obesity (Silver Spring) 2008; 16(4): 881-6.
Townsend MK, Danforth KN, Rosner B, Curhan GC, Resnick NM, Grodstein F. Body mass index, weight gain, and incident urinary incontinence in middle-aged women. Obstet Gynecol 2007; 110(2 Pt 1): 346-53.
Bortolotti A, Bernardini B, Colli E, et al. Prevalence and risk factors for urinary incontinence in Italy. Eur Urol 2000; 37(1): 30-5.
Goode PS, Burgio KL, Redden DT, et al. Population based study of incidence and predictors of urinary incontinence in black and white older adults. J Urol 2008; 179(4): 1449-53; discussion 53-4.
Khullar V, Sexton CC, Thompson CL, Milsom I, Bitoun CE, Coyne KS. The relationship between BMI and urinary incontinence subgroups: results from EpiLUTS. Neurourol Urodyn 2014; 33(4): 392-9.
Kim IH, Chun H, Kwon JW. Gender differences in the effect of obesity on chronic diseases among the elderly Koreans. J Korean Med Sci 2011; 26(2): 250-7.
Larrieu S, Peres K, Letenneur L, et al. Relationship between body mass index and different domains of disability in older persons: the 3C study. Int J Obes Relat Metab Disord 2004; 28(12): 1555-60.
Subak LL, King WC, Belle SH, et al. Urinary Incontinence Before and After Bariatric Surgery. JAMA Intern Med 2015; 175(8): 1378-87.
Tennstedt SL, Link CL, Steers WD, McKinlay JB. Prevalence of and risk factors for urine leakage in a racially and ethnically diverse population of adults: the Boston Area Community Health (BACH) Survey. Am J Epidemiol 2008; 167(4): 390-9.
Tsui A, Kuh D, Cardozo L, Davis D. Vascular risk factors for male and female urgency urinary incontinence at age 68 years from a British birth cohort study. BJU Int 2018; 122(1): 118-25.
Ades PA, Savage PD. The obesity paradox: perception vs knowledge. Mayo Clin Proc 2010; 85(2): 112-4.
Donini LM, Pinto A, Giusti AM, Lenzi A, Poggiogalle E. Obesity or BMI Paradox? Beneath the Tip of the Iceberg. Front Nutr 2020; 7: 53.
Suthers K, Kim JK, Crimmins E. Life expectancy with cognitive impairment in the older population of the United States. J Gerontol B Psychol Sci Soc Sci 2003; 58(3): S179-86.
Herzog AR, Wallace RB. Measures of cognitive functioning in the AHEAD Study. J Gerontol B Psychol Sci Soc Sci 1997; 52 Spec No: 37-48.
Williams BD, Pendleton N, Chandola T. Cognitively stimulating activities and risk of probable dementia or cognitive impairment in the English Longitudinal Study of Ageing. SSM Popul Health 2020; 12: 100656.
Leist AK, Glymour MM, Mackenbach JP, van Lenthe FJ, Avendano M. Time away from work predicts later cognitive function: differences by activity during leave. Ann Epidemiol 2013; 23(8): 455-62.
Davison KK, Ford ES, Cogswell ME, Dietz WH. Percentage of body fat and body mass index are associated with mobility limitations in people aged 70 and older from NHANES III. J Am Geriatr Soc 2002; 50(11): 1802-9.
Galanos AN, Pieper CF, Cornoni-Huntley JC, Bales CW, Fillenbaum GG. Nutrition and function: is there a relationship between body mass index and the functional capabilities of community-dwelling elderly? J Am Geriatr Soc 1994; 42(4): 368-73.
Manandhar MC. Functional ability and nutritional status of free-living elderly people. Proc Nutr Soc 1995; 54(3): 677-91.

Table 1. Descriptive statistics by sex in HRS, ELSA, and SHARE

Characteristics	HRS(N = 207,805)		ELSA(N = 98,158)		SHARE(N = 360,800)
	Female(N = 115,849)	Male(N = 91,956)	Female(N = 53,325)	Male(N = 44,833)	Female(N = 198,788)	Male(N = 162,012)
Age (years)	66(58,76)	65(58,75)	67 (60, 75)	67 (61, 74)	64 (57, 73)	64 (57, 72)
White (%)	83,372(72.0%)	67,635(73.6%)	50,857 (95.4%)	42,776 (95.4%)	-	-
Educational attainments (%)
Less than secondary	17,590 (15.2%)	15,178 (16.5%)	19,469 (36.5%)	13,181 (29.4%)	91,628 (46.1%)	62,338 (38.5%)
Upper secondary and vocational training	36,638 (31.6%)	44,463 (48.4%)	21,542 (40.4%)	20,322 (45.3%)	69,998 (35.2%)	62,676 (38.7%)
Tertiary	16,088 (13.9%)	14,082 (15.3%)	6,613 (12.4%)	8,015 (17.9%)	37,162 (18.7%)	36,998 (22.8%)
Residence area (%)
Rural	28,101 (24.3%)	23,160 (25.2%)	-	-	45,775 (23.0%)	39,663 (24.5%)
Urban	78,431 (67.7%)	61,214 (66.6%)	-	-	102,982 (51.8%)	83,292 (51.4%)
Married and partnered (%)	61,393 (53.0%)	27,527 (29.9%)	20,607 (38.6%)	11,283 (25.2%)	58,966 (29.7%)	27,708 (17.1%)
Number of Children (n)	2(1,4)	2(1,4)	2(1,3)	2(1,3)	2(1,3)	2(1,3)
Urinary incontinence (%)	18,070 (15.6%)	6,065(6.6%)	5,629(10.6%)	1,973 (4.4%)	5,488 (2.8%)	2,203 (1.4%)
Current smoking (%)	7,148(6.2%)	6,175(6.7%)	2,817 (5.3%)	2,387 (5.3%)	10,286 (5.2%)	12,422 (7.7%)
Ever smoked (%)	27,041 (23.3%)	25,978 (28.3%)	14,419 (27.0%)	14,141 (31.5%)	22,454 (11.3%)	34,081 (21.0%)
Alcohol consumption (%)	27,408 (23.7%)	25,528 (27.8%)	18,412 (34.5%)	15,956 (35.6%)	37,284 (18.8%)	36,532 (22.5%)
Physically active (%)	50,811 (43.9%)	38,153 (41.5%)	24,689 (46.3%)	20,305 (45.3%)	65,096 (32.7%)	53,100 (32.8%)
History of diseases
Hypertension (%)	33,872 (29.2%)	24,563 (26.7%)	10,802 (20.3%)	9,490 (21.2%)	27,516 (13.8%)	20,544 (12.7%)
Diabetes (%)	13,547 (11.7%)	10,716 (11.7%)	2,607 (4.9%)	2,988 (6.7%)	7,873 (4.0%)	7,163 (4.4%)
Cancer (%)	8,127(7.0%)	6,253(6.8%)	3,203 (6.0%)	2,272 (5.1%)	4,662 (2.3%)	3,300 (2.0%)
Stroke (%)	4,981(4.3%)	3,850(4.2%)	1,162 (2.2%)	1,159 (2.6%)	3,017 (1.5%)	3,032 (1.9%)
Cognitive impairment (%)	5,110(4.4%)	4,677(5.1%)	1,736 (3.3%)	1,264 (2.8%)	7,738 (3.9%)	5,532 (3.4%)
BMI (kg/m^2)	28.6±6.5	28.5±5.3	28.2±5.6	28.2±4.4	26.5±4.8	27.0±4.0
Waist circumferences (cm)	99.3±14.5	105.2±12.4	92.0±13.6	102.2±11.9	-	-
Data are presented as mean ± SD or median (IQR), n (%)
HRS, the Health and Retirement Study; ELSA, the English Longitudinal Study of Ageing; SHARE, Survey of Health, Ageing and Retirement in Europe
BMI, body mass index

Table 2 and 3 are available in the Supplementary Files section.

There is NO Competing Interest.

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

Status:

Journal Publication

Version 1

Abstract

Figures

Research In Context

Introduction

Methods

Study design and participants

Data collection

Statistical analysis

Results

Discussion

Declarations

References

Tables

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1