Abstract

INTRODUCTION

UI affects over 15 million people in the US and about 30–40% of the population aged 75 years and older. It can have adverse psychological, physical, and social effects including skin breakdown, recurrent urinary tract infections, impaired sleep, falls and fractures, social withdrawal, anxiety, depression, and institutionalization.¹ The cost of UI is expected to increase dramatically as the overall US population continues to age.

Annual incident UI in community-dwelling adults aged 60 years or older ranges from 8% to 18%.^2,3 Incidence of UI in older women has been associated with poor general health, lower income, higher BMI, limited mobility, and depression.^4,5 Although recent studies suggest that the risk factors for incident UI may vary according to racial or ethnic group, few studies report on the incidence in different ethnic groups.^6–8

Latinos currently represent 15% of the U.S. population and are the fastest growing group of Americans aged 60 years and older.⁷ Studies of UI in middle-aged and older Latinos indicate a similar or slightly higher prevalence of UI compared to non-Latino white women.^8,9 Though Latinos are a heterogeneous group, it might be expected that Latinos would have higher rates of UI: Latinos are more likely to state that age-related changes (such as urinary incontinence) are an expected part of normal aging than non-Latinos¹⁰ and also experience lower rates of high quality care.¹¹ It is important to identify correlates of UI among older Latinos to design and implement appropriate interventions to prevent UI and its sequelae in this rapidly growing population.

The purpose of this study was to improve our understanding of UI and its predisposing characteristics among older Latinos. The prevalence of UI and its cross-sectional correlates among urban Latinos seniors enrolled in a randomized trial of a physical activity intervention, “Caminemos” was described previously.² This study took advantage of the longitudinal data from the Caminemos trial to accomplish the following specific aims: 1) measure 1-year incidence of UI; and 2) identify longitudinal correlates of 1-year UI incidence among urban Latino seniors.

METHODS

RESULTS

Mean age of the 328 participants at baseline was 72.5 years (range 60–93). Seventy-five percent of participants were female and 14.3% had been hospitalized in the previous 6 months (Table 1). The rate of incident UI at 1-year was 17.4%, with 18.5% of women and 13.8% of men reporting 1-year incident UI (P=.33 for difference between groups). Severity of incident UI was less than one time per week for 4.6% of entire sample, two to three times per week for 2.1%, daily for 3.4%, several times per day for 5.5% and all the time for 1.8% of the entire sample. Table 1 shows the characteristics of the study participants according to UI-incidence.

DISCUSSION

In this sample of urban older Latinos participating in a senior-center based clinical trial of an exercise intervention in the greater Los Angeles region, overall 1-year incidence of UI was 17.4%, with 18.5% of women and 13.8% of men reporting new UI. This incidence rate is consistent with other studies that have shown annual incident UI rates of 8% to 18% among community-dwelling adults aged 60 years or older,² but higher than the rate of 11.5% found in Hispanic Established Populations for Epidemiologic Studies of the Elderly (EPESE) using a very different measure of UI.³ In the multivariate models adjusting for sociodemographic, behavioral, medical, physical, and psychosocial characteristics, we found that improvement in physical performance score at 1-year and higher baseline HRQoL were independently associated with lower rates of incident UI. In addition, an increase in depressive symptoms at 1-year was independently associated with a higher rate of incident UI.

Decreased general health, obesity, and poor baseline physical performance measures have been previously observed to be associated with incident UI.^2,3,5 However, this is the first empiric study to show that an improvement in physical performance over time is associated with lower incidence of UI. Physiologically this might occur by seniors being more easily able to get to the toilet in a timely manner or through a pathway in which overall physical fitness improves bladder stability as well. Though not conclusive, this new finding suggests that it might be possible to prevent or forestall UI among older Latinos by improving physical performance. Improvement in physical performance was experienced by 62% of the sample, with 26% of the sample increasing their scores on the summary scale by 2 points, a level likely to be associated with lower rates of disability.²² While one might argue that physical performance is a marker of overall health status rather than a modifiable “outcome,” an increasing body of data suggests that older adults can successfully improve their physical performance with focused interventions.^23,24 In the landmark “PREHAB” intervention of 188 frail adults aged 75 and older, Gill et al.²³ found in their randomized controlled trial that balance and conditioning exercises performed over 6 months led to sustained increases in physical function and an improvement of 15.6% over the control group in the modified Physical Performance Test score at 12 months. Another randomized control trial evaluated the benefit of power and strength training in 39 community-dwelling adults aged 65–90 years over 4 months.²⁴ Physical Functional Performance Test total score was significantly greater for the power training group than in the strength training (P=.03) and control (P=.02) groups. The study concluded that power training was more effective than strength training for improving physical function in community-dwelling older adults. In light of our finding that increasing physical performance is associated with lower UI incidence, further investigations should examine whether interventions similar to the Gill and Miszko interventions can also decrease UI incidence.

A robust association was found between an increase in depressive symptoms at 1-year and incident UI. Worsening of depressive symptoms was experienced by 94/328 (29%) of the sample. Among these, 29 (31%) scored >2 on the 5-item GDS, indicating high likelihood of clinical depression.²¹ However, baseline higher depression symptom score was not significantly associated with incident UI. Depression is a well-recognized risk factor for UI in nursing home residents, the elderly, and those admitted for hip fracture.²⁵ Depression has also previously been described in 60% of those with idiopathic urge incontinence,²⁶ however, the causal mechanism is not clear. To date there is contradicting literature on whether depressive symptoms occur as a cause of incontinence, or vice versa.²⁷ Although UI may lead to depression and social withdrawal, there is some evidence that bladder overactivity may be related to a decline in serotonin function,²⁶ (commonly seen in depression) or an alteration in dopamine neurotransmitter function.²⁸ It is possible that after the start of the Caminemos trial, some participants were subsequently diagnosed with depression and treated with antidepressants, altering the neurotransmitter levels in the bladder; unfortunately it is a limitation of this study that we do not have data on whether subjects were taking antidepressant medication. This could be why an increase in depression symptom score at 1-year was found to be a significant correlate of incident UI, but baseline depression symptom score was not. It is also possible that common medical conditions, such as stroke, could predispose older adults to both depression and UI, and that the depression and UI were part of a “symptom cluster.” There is already a known association between physical performance and depression in the elderly.²⁹ Exercise and strength training may therefore have an enhanced potential to positively impact both depression and UI concomitantly.

The association between UI and HRQoL has been previously recognized.^2,9,27 In this study, the fully adjusted model (model 3) showed a significant independent association between higher baseline physical and mental HRQoL and a lower incidence of UI at 1-year. Poor QoL is generally considered to represent multiple contributing physical and psychological factors. Even after controlling for medical comorbidity, ADL function and depressive symptoms, HRQoL scores at baseline remained significantly correlated with incident UI, suggesting that people with new UI have lower QoL for other reasons beyond UI’s effect on function and mood.

Obesity is associated with multiple medical comorbidities and weight loss of 5–10% is known to improve physical health and hypertension, hyperglycemia and hyperlipidemia, reduce the risk of type 2 diabetes, and result in improve QoL.³⁰ Although it is well established that obesity can also cause UI or contribute to the condition’s severity,^4,5 our study did not show BMI or gain of >5% body weight to be a significant correlate of incident UI. However, the trend towards incident UI with >5% gain in body weight was strong and it is possible that had we had greater power with a larger sample we would have found an association. This trend should be further evaluated in future studies since obesity is so prevalent and weight reduction may be an effective treatment for overweight and obese women with UI.

Some important bivariate correlates did not maintain statistical significance in the multivariate models, including older age, baseline ADL impairment, and mean steps per day. Most likely other constructs, most notably the physical performance score, captured the impact of these factors. In contrast to cross-sectional associations with prevalence of UI reported earlier,² neither the Charlson Comorbidity Index, the presence of cerebrovascular disease, nor the presence of diabetes correlated with 1-year incident UI. The cross-sectional association may demonstrate the cumulative effect of years of chronic diseases, such as diabetes mellitus and vascular disease, and their effect on the innervation of the lower urinary tract leading to UI. It is possible that those who were susceptible to these effects had already reported UI at baseline and thus were not in our sample. Alternatively, because the Charlson Comorbidity Index was only assessed at baseline it is possible that many participants experiencing incident UI had new comorbidity such as stroke that we did not measure and this could account for the lack of an observed association. It is also interesting to note that unlike in other studies, neither cerebrovascular disease nor diabetes were associated with incident UI, possibly

Major limitations of this study include selection bias, attrition, and possible measurement error. Our sample was made up of ambulatory Latinos in a single geographic area recruited from community centers for a physical activity trial. Senior Latinos who were less likely to go to a senior center or in poorer health in nursing care or at home may have led to an underestimation of true UI incidence. Missing data (18% missing in Model 4) may have biased our findings either towards or away from the null hypotheses depending on whether the subjects missing data experienced the variables of interest in the direction we observed to be associated with UI. Also, there was no physical exam conducted; these data are based on self-reported instruments and as such are subject to recall and social desirability bias. Another limitation of the study is the absence of questions to separate urge and stress UI, which may have different risk factors and etiologies. This study’s use of a single-item measure of UI modified from the 4-item ICIQ with previously unmeasured validity may have resulted in a different outcome than those from another UI measure. It is also important to point out that this study identifies correlates of any UI, but there may be different correlates of having UI less than once per week versus all the time. Future studies should also examine whether a change in living situation, for example moving from independent living to relying on others for access to a toilet contributes to incident UI.

CONCLUSION

The 1-year incidence of UI among community-dwelling urban older Latinos was high with 18.5% of women and 13.8% of men reporting incident UI at 1-year follow-up. The robust associations between incident UI and HRQoL, worsened depressive symptoms, and improved physical performance highlights the potential impact of interventions to improve physical performance and decrease depressive symptoms upon rates of UI.

Future studies should examine whether focusing on improving physical performance measures in the elderly may be a powerful way to prevent and treat UI among sedentary urban older Latinos.

ACKNOWLEDGMENTS

Footnotes

REFERENCES