Regular physical activity is associated with decreased mortality and a lower risk of developing diabetes, hypertension, and colon cancer, compared with inactivity.1 It also increases muscle and bone strength, improves psychological well-being, and decreases body fat.1 The latter is particularly pertinent, given the current “obesity epidemic.” According to the most recent data from the National Health and Nutrition Survey,2 nearly two thirds of adult American women are overweight or obese (body mass index [BMI] ≥ 25) and one third are obese (BMI ≥ 30). Despite the impact exercise might have on reducing obesity and gaining other benefits, few Americans exercise regularly. One of the aims of Healthy People 2010 is to increase to 30% the proportion of adults who engage regularly in moderate physical activity. It is of increasing importance to understand what women perceive as barriers to exercise. One potential barrier to exercise is urinary incontinence. Numerous epidemiologic studies have found that half of community-dwelling women leak urine on occasion and 5–10% have daily leakage.3
Over one quarter of college varsity athletes and physical education students leak urine on occasion.4–7 In a study of 291 young elite female athletes and dancers, 43% experienced leakage while participating in their sport; of those, 60% occasionally wore pads because of urine loss and 6.6% restricted fluids to decrease leakage.8
While studies of young healthy athletes are valuable in demonstrating how common leakage during sports is, this population is not ideal to study the degree to which leakage is a barrier to exercise for women in the community. Elite athletes are more likely to perform extremely strenuous, high-impact activities than are recreational athletes. In addition, women participating in organized sports are highly motivated to continue, despite physical injuries or ailments. Those women for whom a medical condition is a barrier may have already stopped exercise and thus would not be represented in populations of elite athletes.
We conducted this cross-sectional study to improve our understanding of urinary incontinence and exercise in a broad population of young and middle-aged women. Our aims were to (1) describe the prevalence of urinary incontinence during exercise in women, (2) estimate whether an association exists between exercise intensity and leakage severity, and (3) estimate to what degree urinary incontinence is a barrier to exercise in women.
MATERIALS AND METHODS
We conducted a national, mailed survey of women between the ages of 18 and 60 years drawn from National Family Opinion survey research panels. We chose our sample size of approximately 5,000 women based on the assumptions that (1) approximately one third of the sample would report incontinence in the last 30 days, and (2) approximately two thirds of the sample would respond to the survey (based on prior similar surveys).
Approximately, 500,000 U.S. households volunteer to participate in National Family Opinion survey research studies on a wide range of topics. Of these 500,000 households, 200,000 actively receive surveys and are divided into mini-panels of 5,000 households. Each mini-panel is balanced to match U.S. census distributions in 5 areas: geographic region, household income, household size, age of male and female heads of households, and market size. From 20 of the mini-panels, a list was developed of all women between the ages of 18 and 60 years. A sample of 5,130 women was sent questionnaires. The questionnaire was not sent to more than 1 woman in a household. All women who received the survey were given $1. The survey research was submitted to Western Institutional Review Board, and met the requirements for a waiver of documentation of consent.
We made no attempt to send more than one survey to nonrespondents. Because the survey was anonymous, we were unable to directly compare respondents and nonrespondents. However, for the original sample, we knew the aggregate characteristics for education, age, geographic division, household income, and race. By subtracting the respondents’ characteristics from the overall characteristics of the sample, we estimated the characteristics of the nonrespondents. Proportions of respondents and associated confidence intervals were adjusted using an iterative proportional fitting methodology (rim weighting) to account for potential bias due to nonresponse.
To assess physical activity, we used the International Physical Activity Questionnaire short form,9 a validated instrument that includes physical activity related to a comprehensive set of domains including leisure activity, domestic and yard work, and work-related activity. The International Physical Activity Questionnaire assesses walking, moderate-intensity activities, and vigorous intensity activities, and summarizes frequency per week and duration per day. Each level of activity is assigned a metabolic energy requirement score (in MET-minutes) by multiplying the duration, frequency, and MET score (walking = 3.3 METs, moderate activity = 4.0 METs and vigorous activity = 8.0 METs).
Three levels of physical activity are used to classify populations: insufficiently, sufficiently, and highly active. A woman is considered sufficiently active if, in a week, she meets any 1 of the following 3 criteria: ≥ 3 days of vigorous activity for ≥ 20 minutes/day; ≥ 5 days of moderate-intensity activity or walking for ≥ 30 minutes per day; or ≥ 5 days of any combination of activity achieving a minimum of ≥ 600 MET-minutes per week. Criteria for classification as “highly active” are vigorous-intensity activity on at least 3 days achieving a minimum of at least 1,500 MET-minutes per week, or 7 or more days of any combination of walking, moderate-intensity, or vigorous-intensity activities achieving a minimum of at least 1,500 MET-minutes per week. Data from the International Physical Activity Questionnaire were scored in accordance with established guidelines.10
We asked a series of questions about exercise patterns and self-reported barriers to exercise. We considered women to be consistent exercisers if they reported regular exercise within the last 5 years or over their lifetime. Women who reported never or sporadically exercising, or who reported regular past exercise but not in the last 5 years, were categorized as inconsistent exercisers.
To assess perceived barriers to exercise, we asked about 11 factors, including urinary incontinence. For each factor, women were asked to what degree that factor prevented participation in physical activities on a scale of 1 (not at all) to 5 (completely). For the purposes of reporting, these categories were compressed into 3: minimal (not at all and very little), moderate (somewhat), and substantial (great deal and completely). The questions about barriers to exercise were asked before any questions about the presence of urinary incontinence symptoms.
Comorbid medical conditions were assessed by asking women which of 16 health problems (including a category of “other”) they currently experience. For purposes of analysis, we categorized women as having 0, 1, 2, or 3 or more comorbidities.
Women were considered to have urinary incontinence if they reported involuntary leakage or loss of urine in the previous 30 days. We classified women as having stress incontinence symptoms if they answered yes to at least one of four items on involuntary leakage of urine due to (1) laughing, (2) coughing or sneezing, (3) exercising, or (4) lifting or physical activity. Women were classified as having urge incontinence symptoms if they answered yes to losing urine with (1) the urge to urinate but unable to get to the toilet before losing urine, or (2) a strong sudden urge to go to the toilet to urinate with no advanced warning. Respondents reporting both stress and urge symptoms were classified as having mixed incontinence symptoms. Women unable to categorize leakage were classified as having “other” incontinence symptoms.
To assess incontinence severity, we used the Sandvik severity index.11 In this index, frequency (0 = never, 1 = less than once a month, 2 = a few times a month, 3 = a few times a week, 4 = every day/night) and amount of leakage (0 = no leakage, 1 = drops, 2 = small splashes, 3 = more) are multiplied to form a four level severity index (1–2 = slight, 3–6 = moderate, 8–9 = severe, 12 = very severe). The severity index has been compared with 48-hour pad testing and urodynamic testing, and was reproducible and valid.12 We assessed global bother of incontinence symptoms on a five-point Likert scale ranging from not at all bothersome to extremely bothersome. Previous use of this scale has shown increased bother to be associated with increased frequency of incontinence episodes.13
Among incontinent women, we asked about the potential impact of incontinence on physical activity. Binary outcomes were compared between incontinent and continent groups using the Pearson χ2 statistic. Differences in ordinal column scores among rows of contingency tables were assessed using the Cochran-Mantel-Haenszel statistic. Odds ratios for dichotomous outcomes were calculated with their corresponding 95% Wald confidence intervals. All analyses were performed using SAS 8.2 for Windows.
Logistic regression models were constructed on the binary outcome insufficient activity versus sufficient activity or highly active. Predictors in these models included incontinence severity (treated as nominal categories relative to continent women), perceived bother of urinary incontinence (nominal categories relative to continent women), age in years (continuous), number of comorbidities (0, 1, 2, or 3 or more), education (less than high school, high school graduate, some college, graduated college, post graduate) and race (white, yes/no). To explore perceived barriers to exercise, logistic regression models were constructed for each barrier with the binary outcome of moderate or substantial barrier versus no or minimal barrier and predictors of exercise habit (consistent, inconsistent), urinary incontinence (dichotomous), education (five levels), race (white, yes/no), number of comorbidities and age (continuous in decades).
Of the 5,130 questionnaires mailed, 3,562 were completed. Of these, 72 respondents were pregnant and were excluded from this analysis. Although our a priori intention was to restrict questionnaire completion to women aged 18–60 years, 126 women older than 60 years responded, and these women were also excluded from our analyses. Thus, our study population consisted of 3,364 women (response rate 68%).
Respondents had a mean age of 44.4 years (standard deviation [SD] 9.8 years) with 6.1% of the women aged 18–30 years, 36.4% aged 31–45 years, and 57.5% aged 46–60 years. The mean weight was 171.0 lb (SD 48.7 lb; median 160.0 lb); and mean BMI was 28.6 (SD 7.9). A total of 41.5% of respondents were college graduates or higher, whereas 30.2% had an annual household income less than $35,000. A total of 68.4% were married; 87.5% were white, 6.9% were black, and the rest identified themselves as Asian or Pacific Islander, Native American, other, or a combination of the above. Of all respondents, 4.6% considered themselves to be of Spanish, Latin, or Hispanic origin. Demographic characteristics of the sample by International Physical Activity Questionnaire activity level are shown in Table 1.
A comparison of respondents and nonrespondents revealed that nonrespondents were more likely to have graduated college (40% versus 29%), less likely to have attended postgraduate school (13% versus 7%) and less likely to be in the 45–64-year age group (37% versus 27%). Characteristics of respondents and nonrespondents regarding geographic division, income, and race were very similar.
Eighty-one percent of respondents reported at least one comorbidity, whereas 56.7% had 2 or more and 35.8% had 3 or more. Comorbidities included allergies/hay fever (47.8%), arthritis (22.8%), hypertension (20.7%), hypercholesterolemia (17.6%), depression (16.2%), anxiety (13.0%), asthma (11.6%), chronic pain (11.5%), and osteoporosis (5.1%).
Four hundred thirty-one (13.6%) respondents did no exercise (including walking) for at least 10 minutes a week, and were classified as part of the insufficiently active group. Based on the International Physical Activity Questionnaire criteria, 34.3% of subjects were insufficiently active, 29.9% were sufficiently active, and 35.8% were highly active. Post-stratification adjustment yielded estimates of 34.6% insufficiently active (95% CI 32.7–36.5%), 29.7% sufficiently active (95% CI 27.9–31.5%) and 35.7% highly active (95% CI 33.8–37.6%); 38.2% of respondents engaged in vigorous activity at least once a week. Activity level was not associated with age group. Of respondents aged 18–30 years, 33.7% were insufficiently active; of women aged 31–45 years, 34.0% were insufficiently active; and of women aged 46–60 years, 34.8% were insufficiently active (P = .87).
Respondents summarized the consistency of their life-long exercise habits as follows: 19.1% exercised regularly for all their lives, 14.6% began exercising regularly within the last 5 years, 21.2% used to exercise regularly but decreased exercising in the past 5 years, 40.0% exercised only sporadically throughout their lives, and 5.2% never exercised. On the basis of these responses, we categorized 33.7% of the women as consistent exercisers and 66.3% as inconsistent exercisers.
Overall, 1297 respondents met our definition of incontinence (ie, involuntary urine leakage in the past 30 days). Of these, 74% reported leakage in the pre-vious week. Severity was slight in 46.5% of women, moderate in 41.4%, severe in 9.1%, and very severe in 2.9%. A total of 48.9% reported no bother or slight bother, and 52.1% reported moderate to extreme bother of urinary incontinence. A total of 46.7% had stress incontinence, 9.7% urge, 41.9% mixed, and 1.7% other. Post-stratification adjustment yielded an estimate of 34.3% (95% CI 32.5–36.1%) of women with urinary incontinence.
We estimate that 8.3% (95% CI 5.1–7.4%) experienced leakage during exercise, 9.1% (95% CI 8.1–10.0%) experienced leakage during lifting or physical activity, and 15.7% (95% CI 14.2–17.2%) experienced leakage during exercise, lifting, or physical activity. Of respondents with urinary incontinence in the last 30 days, 27.1% experienced leakage during exercising and 31.4% during lifting or physical activity.
More active women were more likely to report urinary incontinence while exercising than less active women. Among all respondents, 11.8% of those with insufficient activity experienced urinary incontinence during exercising or physical activity compared to 15.1% of sufficiently active and 15.9% of highly active women (P = .01). Type of leakage (stress, urge, or mixed) was not associated with International Physical Activity Questionnaire activity level (P = .27).
There was no difference in the overall prevalence of urinary incontinence among respondents in the three activity level (International Physical Activity Questionnaire) groups: 38.8% for those insufficiently active, 37.4% for women sufficiently active, and 38.2% for highly active women (P = .81).
We hypothesized that women with more severe and more bothersome urinary incontinence might avoid physical activity to decrease their chances of experiencing leakage. Women with very severe urinary incontinence were more likely to be insufficiently active (60%) than women with severe (38.3%), moderate (38.2%), or slight (29.5%) incontinence (P < .01). The adjusted (for age, number of comorbidities, education, and race) odds ratios of insufficient activity (versus sufficient or highly active levels) for various levels of incontinence severity were 0.80 (95% CI 0.65–0.99) for slight, 1.19 (95% CI 0.9–1.47) for moderate, 1.16 (95% CI 0.77–1.75) for severe, and 2.64 (95% CI 1.25–5.55) for very severe urinary incontinence (each relative to continent women). None of the other factors adjusted for achieved significance in the model.
Similarly, the adjusted odds ratios (adjusted for same predictors as above) of insufficient activity for levels of bother associated with urinary incontinence were 0.83 (95% CI 0.68–1.02) for no/slight bother and 1.22 (95% CI 1.00–1.48) for moderate to extreme bother relative to continent women. None of the other predictors achieved significance in the model.
Barriers to exercise were compared for women with consistent and inconsistent exercise habits, given the likelihood that lifetime habits are a major predictor of regular exercise. On univariate analysis, inconsistent exercisers were more likely than consistent ones to endorse each of the barriers as being a substantial barrier to exercise. We estimate that 9.8% (95% CI 8.8–10.9%) of women perceive leakage of urine as at least a moderate barrier to exercise. Of the women who reported urinary incontinence in the previous 30 days, 27.9% of women cited leakage of urine as at least a moderate barrier to exercise, compared with 1.7% of the women who did not report leakage within the last 30 days.
As incontinence frequency increased, women were more likely to report leakage to be at least a moderate barrier to exercise. Five percent of women with leakage less than once a month, 19.4% with monthly leakage, 37.7% with leakage a few times a week, and 65.9% of women with leakage every day and/or night considered leakage to be at least a moderate barrier to exercise (P < .01). Of incontinent women, the proportion for whom urinary incontinence was at least a moderate barrier to exercise also increased with each Sandvik severity category: 9.2%, slight; 37.8%, moderate; 64.6%, severe; and 85.3%, very severe (P < .01). Women with mixed urinary incontinence were more likely (42.5%) to report leakage as at least a moderate barrier to exercise than women with urge urinary incontinence alone (12.6%) or stress urinary incontinence alone (21.5%) (P < .01).
After adjusting for age, exercise habit, urinary incontinence, education, race, and comorbidities, the barrier models (Table 2) confirm a number of expected associations; for example, a higher number of comorbidities is associated with a higher probability of identifying health problems as a barrier to exercise, more education is associated with a lower probability of identifying money as a barrier to exercise, etc. Experiencing leakage in the last 30 days was a highly significant predictor (odds ratio 19.0; 95% CI 13.1–27.5) of reporting urinary incontinence as a barrier to exercise, after adjusting for other potential confounders.
Women reporting urinary incontinence in the previous 30 days were asked more specific questions about the impact of incontinence on their decisions about exercise. Most (64.6%) stated that the amount of urinary incontinence was so small that there was no effect on exercise. However, over 1 in 10 (11.6%) stated that they did not exercise because of urinary incontinence, whereas 11.3% exercised less because of urinary incontinence. A total of 12.4% changed the types of exercise performed, and 5.0% no longer exercised in a gym or with other people, because of urinary incontinence. Women with severe and very severe urinary incontinence were more likely to stop or change exercise than were women with slight or moderate urinary incontinence (Table 3). A separate analysis stratified for age group (18–30, 31–45, 46–60) revealed no differences in impact of incontinence on exercise by age.
Similar to other population-based studies, we found that over one third of women between the ages of 18 and 60 years reported at least monthly urinary leakage. Despite the high prevalence of this disorder, little information exists about urinary incontinence and recreational exercise.
In our study population, 1 in 7 women overall experienced urinary leakage during exercise and/or physical activity and 1 in 8 perceived leakage to be at least a moderate barrier to exercise. Of women with urinary incontinence (at least monthly), over one quarter experienced leakage during exercise/physical activity and a similar proportion considered leakage to be at least a moderate barrier to exercise. A small number of women reported that urinary incontinence was a barrier to exercise despite the fact that they reported no leakage in the last month. It is possible that these were women who experienced leakage in the past mainly during exercise and subsequently limited exercise to avoid leakage.
Of interest, women with mixed urinary incontinence were more likely to perceive leakage as a barrier to exercise than women with pure stress or urge incontinence. Compared with women with pure symptoms, women with mixed incontinence reported greater severity of leakage and greater bother from symptoms. Leakage may also be more of a barrier to exercise in the mixed group because they not only have leakage with exercise to contend with, but also may have the inconvenience and interruption associated with having to stop exercising frequently to find a bathroom facility.
Clinicians may be surprised that three fourths of women with incontinence did not consider leakage a barrier to exercise. This likely reflects the difference between women seeking care for this disorder versus those with less severe incontinence who may not choose to seek care. Indeed, the proportion of women considering urinary incontinence a barrier to exercise increased with increasing severity and frequency of urinary incontinence such that two thirds of women with daily urinary incontinence and 85% of women with very severe urinary incontinence considered it a barrier to exercise. Similarly, a woman was more likely to be insufficiently active as urinary incontinence severity increased. One in 10 women with urinary incontinence, and 1 in 4 women with severe or very severe urinary incontinence, stopped exercising specifically because of this condition.
After adjusting for confounders, we found that experiencing leakage in the last 30 days was a substantial predictor of urinary incontinence as a barrier to exercise, increasing this risk nearly 20-fold compared with women with no leakage in the last 30 days. That most women had mild or moderate incontinence speaks to the respondents’ perception of the impact of this condition, regardless of severity, on exercise.
Other researchers have also found that urinary incontinence affects exercise. In a cross-sectional study14 of 326 women (mean age 38.5 years) attending gynecology clinics at one center, 30% of exercisers noted incontinence during at least one type of exercise. Twenty percent of these incontinent exercisers stopped an exercise because of incontinence whereas 18% changed the way an exercise was done and 55% wore a pad.
More women leak while doing activities involving repetitive bouncing and high-impact landings than low-impact activities.14,6,4 Swedish researchers15 found that 80% of 35 elite trampolinists (mean age 15 years, range 12–22 years) reported leakage during trampoline training. Most subjects underwent further objective testing and their leakage averaged 28 cc during one jump session. As recreational exercisers, women in our study were generally not subjected to similar high-impact forces. However, women in our study were also more likely to experience urinary incontinence during exercise as their activity level increased.
In a large population-based Australian study,16 1,050 women that previously reported incontinence on a questionnaire (defined as “leaking urine often” if middle or older aged and as “sometimes or more often and more than just drops” if younger) were queried later about leakage and physical activity. Women with more frequent leakage were more likely to report low levels of physical activity. A total of 43.7% of middle-aged (48–53), 13.9% of young (21–26), and 15.3% of older (72–79) women reported leakage during sport or exercise. A total of 6.7% of young, 38% of middle-aged, and 27.5% of older women “sometimes or often” avoided sporting activities because of leakage, whereas 4.5%, 25.6%, and 19.1%, respectively, avoided recreational activities that they used to do. Women in this study were selected for inclusion because they had moderate to severe leakage and the results are thus not reflective of a population of women. In our study, women with varying degrees of continence are included; therefore, the proportion of women that stopped exercising because of leakage is smaller.
Strengths of our study include the large sample size, use of validated instruments to assess incontinence severity and activity level, and 68% response rate. However, it is possible that even with adjustments for nonresponse, the sample may not fully represent the population of U.S. women in the 18–60-year age group. In particular, the proportion of women in our survey who reported a sufficient or high level of physical activity is considerably higher (29.9% and 35.8%, respectively) than the proportion observed in other national surveys. For instance, the National Health Interview Survey reported that only 30.8% of women aged 25–64 years of age engaged in regular physical activity (Centers for Disease Control and Prevention, early release of selected estimates based on data from the January to March 2004 National Health Interview Survey17). Even the 2001 Behavioral Risk Factor Surveillance Survey,18 which expanded the definition of physical activity to include such domains as housework and transportation, found that fewer than half of all adults (45.4%) met current recommendations for physical activity. To some extent, this discrepancy may reflect the variability in physical activity prevalence estimates obtained with different types of measures and scoring protocols,19 and a tendency for the International Physical Activity Questionnaire short form to overestimate activity levels.20 Our study would have been strengthened had we been able to use more objective measures of assessing activity (such as a pedometer or prospective activity record). However, the discrepancy may suggest a selection bias in our sample that could limit the generalizability, although not internal validity, of our findings. Decisions about exercise are multifactorial, as demonstrated by the multiple barriers women in our study reported to exercise. Our study is the first large-scale U.S. study to assess the degree to which urinary incontinence is perceived as a barrier to exercise. For some women, urinary incontinence is an important factor. The fact that 9.8% of women overall, and 1 in 4 women with at least monthly urinary incontinence, perceive leakage as a significant barrier to exercise is concerning.
Improving our knowledge about barriers to exercise in women is important, given the increase in both inactivity and obesity among American women. There are few data about the impact of successful incontinence treatment on exercise patterns. In a prospective study of 69 fairly active (only 13% did little leisure exercise) Finnish women seeking care for urinary incontinence, physical activity did not increase after successful incontinence treatment.21 However, in our population-based study, urinary incontinence was considered a barrier to exercise, particularly in women with more severe leakage. Further research is needed in various populations to determine whether healthier exercise patterns can be established if incontinent women are successfully treated.