Long-term effects of lifestyle and metformin interventions in DPP on bone density.

In the Diabetes Prevention Program Outcome Study (DPPOS), a cohort at high risk of diabetes, randomization to intensive lifestyle intervention or metformin, both associated with weight loss, did not have long-term negative effects on BMD compared with the placebo group. Potential positive effects of metformin on bone warrant further investigation. INTRODUCTION: Randomization to lifestyle intervention (ILS) or metformin in the Diabetes Prevention Program (DPP) resulted in weight loss and reduced progression to diabetes. Weight loss is associated with reduced bone mineral density (BMD), but the long-term effects of these interventions on BMD are unknown. In the DPP Outcome Study (DPPOS), we determined if randomization to ILS or metformin, compared with placebo, was associated with differences in BMD approximately 16 years later. METHODS: Of 3234 DPP participants, 2779 continued in DPPOS and were offered ILS in group format. Those randomized to metformin were offered unmasked metformin. At DPPOS year 12, 1367 participants had dual-energy X-ray absorptiometry scans. BMD in metformin and ILS groups was compared to placebo using sex-specific linear regression models, adjusted for age, race/ethnicity, and weight and weight-bearing activity at DPP baseline. RESULTS: At DPPOS year 12, mean age was 66.5 (±9.5) years. Femoral neck BMD was similar in the ILS and placebo groups in men (difference = -0.021 g/cm2, 95%CI (-0.063, 0.021)) and in women (+0.014 g/cm2, 95%CI (-0.014, 0.042)). Femoral neck BMD was higher in the metformin compared to placebo group although not statistically different in men (+0.017 g/cm2, 95% CI (-0.023, 0.058)) and in women (+0.019 g/cm2, 95% CI (-0.009, 0.047)). Prevalence of osteoporosis was low and similar across treatment groups in men (0.9%; p=0.745) and women (2.4%; p=0.466). CONCLUSION: In a cohort at high risk of diabetes, lifestyle intervention or metformin did not appear to have long-term negative effects on BMD. Potential positive effects of metformin on bone warrant further research.

Long-term changes in dietary and food intake behaviour in the Diabetes Prevention Program Outcomes Study.

AIMS: To compare change in dietary intake, with an emphasis on food groups and food intake behaviour, over time across treatment arms in a diabetes prevention trial and to assess the differences in dietary intake among demographic groups within treatment arms. METHODS: Data are from the Diabetes Prevention Program and Diabetes Prevention Program Outcomes Study. Participants were randomized to a lifestyle intervention (n = 1079), metformin (n = 1073) or placebo (n = 1082) for an average of 3 years, after which the initial results regarding the benefits of the lifestyle intervention were released and all participants were offered a modified lifestyle intervention. Dietary intake was assessed using a food frequency questionnaire at baseline and at 1, 5, 6 and 9 years after randomization. RESULTS: Compared with the metformin and placebo arms, participants in the lifestyle arm maintained a lower total fat and saturated fat and a higher fibre intake up to 9 years after randomization and lower intakes of red meat and sweets were maintained for up to 5 years. Younger participants had higher intakes of poultry and lower intakes of fruits compared with their older counterparts, particularly in the lifestyle arm. Black participants tended to have lower dairy and higher poultry intakes compared with white and Hispanic participants. In the lifestyle arm, men tended to have higher grain, fruit and fish intakes than women. CONCLUSIONS: Changes in nutrient intake among participants in the lifestyle intervention were maintained for up to 9 years. Younger participants reported more unhealthy diets over time and thus may benefit from additional support to achieve and maintain dietary goals.

Change in Objectively Measured Activity Levels Resulting from the EMPOWER Study Lifestyle Intervention.

Purpose: To examine changes in physical activity (PA) during a behavioral weight-loss intervention and determine baseline factors associated with PA goal achievement. Methods: Overweight/obese community-dwelling adults with valid PA accelerometer data (N=116; mean age 51.7 years; 89% female; 83% non-Hispanic White) were recruited into a single-arm prospective cohort study examining the effects of a 12-month intervention that included 24 in-person group sessions, weight-loss, calorie, fat gram, and PA goals, self-monitoring, and feedback. Minutes of moderate-to-vigorous (MV) PA and steps were measured using a waist-worn accelerometer (ActiGraph GT3x) at baseline, 6 months, and 12 months. Achievement of the 150 minute/week MVPA goal was examined using total minutes and bout minutes (i.e., counting only PA occurring in bouts ≥10 minutes in length). Change in PA was analyzed using non-parametric tests for multiple comparisons. Associations of factors with meeting the PA goal were modeled using binary logistic regression. Results: At 6 months, there were increases from baseline in MVPA (median [p25, p75]: 5.3 [-0.9, 17.6] minutes/day) and steps (863 [-145, 2790] steps/day), both p<0.001. At 12 months, improvements were attenuated (MVPA: 2.4 [-2.0, 11.4] minutes/day, p=0.047; steps: 374[-570, 1804] p=0.14). At 6 months, 33.6% of individuals met the PA goal (using total or bout minutes). At 12 months, the percent meeting the goal using total MVPA [31%] differed from bout MVPA [22.4%]. Male gender (OR=4.14, p=0.027) and an autumn program start (versus winter; OR=3.39, p=0.011) were associated with greater odds of goal achievement at 6 months. Conclusions: The intervention increased PA goal achievement at 6 and 12 months with many making clinically meaningful improvements. Our results suggest female participants may require extra support toward improving PA levels.

A randomized walking trial in postmenopausal women: effects on physical activity and health 10 years later.

BACKGROUND: It is important to determine if permanent lifestyle changes may result from physical activity interventions and whether health may be affected by these changes. OBJECTIVE: To conduct a 10-year follow-up of physical activity and self-reported health status in participants of a randomized clinical trial of walking intervention. METHODS: Of the original 229 volunteer postmenopausal women who participated in the original clinical trial, 196 (N = 96 intervention and 100 controls) completed the 10-year follow-up telephone interview. The interview protocol included questions on self-reported walking for exercise and purposes other than exercise, the Paffenbarger sport and exercise index, functional status, and various chronic diseases and conditions. RESULTS: The median values for both usual walking for exercise and total walking were significantly higher for walkers compared with controls (for both, P = .01), with median differences of 706 and 420 kcal/wk, respectively. After excluding women who reported heart disease during the original trial, 2 women in the walking group (2%) and 11 women in the control group (12%) reported physician-diagnosed heart disease over the last 10 years (P = .07). There were also fewer hospitalizations, surgeries, and falls among women in the walking group, although these differences were not statistically significant (P>.05). CONCLUSIONS: Although limited by self-report, this study may be the first to demonstrate long-term exercise compliance to a randomized control trial in older women and to suggest that health benefits may have ensued as a result of these increased activity levels.

Prevention of type II diabetes by physical training. Epidemiological considerations and study methods.

Disease prevention may be considered at three levels: primary (avoiding disease occurrence), secondary (early detection and reversal), and tertiary (prevention or delay of complications). Physical exercise could potentially contribute to all of these. Metabolic studies suggest the major effect of exercise is at the level of insulin sensitivity/resistance. Therefore, it may have the greatest benefit in primary prevention and in the early stages of the disease. Studies of migrants and of active and inactive professions support this notion. There is also provisional support for the benefit of exercise on metabolic control and prevention or delay of chronic complications in non-insulin-dependent (type II) diabetic patients. In designing a trial of exercise, aspects such as single/multifactorial intervention, the age range of subjects, and choice of outcome measures must be considered. The most widely used methodological tool in assessing physical activity levels in population studies is the activity questionnaire, which is nonreactive, practical, applicable, and accurate relative to other methods. A positive approach to exercise training for both the patient with type II diabetes and the general community appears to be warranted. However, trials examining the efficacy of activity intervention, with independent evaluation of both short- and long-term outcomes, are still needed.

Physical activity, physical fitness, and insulin and glucose concentrations in an isolated Native Canadian population experiencing rapid lifestyle change.

OBJECTIVE: Little is known about the relation of physical activity and physical fitness to insulin resistance and glucose intolerance in isolated subarctic Native Canadian populations. The purpose of this effort was to examine the relation between activity and fitness and obesity and glucose concentrations in such a unique population. RESEARCH DESIGN AND METHODS: This study describes 530 men and women from the community of Sandy Lake, Ontario, located in the boreal forest region of central Canada. Fasting blood glucose and insulin concentrations were determined after an overnight fast. Past year physical activity levels were assessed using a modified version of an interviewer-administered questionnaire. Maximal oxygen uptake, a measure of cardiovascular fitness, was estimated using a submaximal step test. RESULTS: Total (leisure and occupational) physical activity and physical fitness were significantly associated with fasting insulin concentrations after adjusting for age, BMI or percent body fat, waist circumference, and fasting glucose concentration in men but not in women. The relations between physical activity, fitness, and fasting glucose concentrations were not as strong or as consistent as they were when fasting insulin concentration was the dependent variable. CONCLUSIONS: In this isolated Native Canadian community, both physical activity and fitness were independently associated with fasting insulin concentrations, suggesting a beneficial role of physical activity/fitness on insulin sensitivity that is separate from any influence of activity on body composition. The fact that this relation was found in men but not in women is most likely explained by issues related to the measurement of activity and fitness in this study and the fact that the women in this population appear to be less active than the men.

Relationship of locus of control to physical activity among people with and without diabetes.

OBJECTIVE: To examine the relationship between locus of control (LOC) (internal and external) and physical activity in Pima Indians and to determine whether this relationship is affected by the presence of diabetes. RESEARCH DESIGN AND METHODS: A population-based sample of 580 Pima Indians was recruited from an ongoing research study. LOC was measured on a 1-40 modified Rotter scale, and past year total physical activity (leisure and work physical activity levels combined) was measured by interviewer-administered questionnaire. RESULTS: Among both men and women without diabetes, individuals with an internal LOC (score 1-16) were significantly (P < 0.01) more active than those with an external (score 17-40) LOC (70 vs. 30 median metabolic equivalent [MET] hours per week for men: 12 vs. 5 median MET hours per week for women). Controlled for age and BMI, an internal LOC was significantly associated with a higher level of physical activity among men (P = 0.04) and women (P = 0.001) without diabetes, but not among those with diabetes. CONCLUSIONS: Nondiabetic Pima Indians with an internal LOC are more physically active than those with an external LOC. Enhancing perceptions of internal control may influence physical activity and thus have implications for diabetes prevention.

Development of questionnaire to examine relationship of physical activity and diabetes in Pima Indians.

There was a need to design a questionnaire that could accurately assess the activity patterns of Native Americans to evaluate the relationship between physical activity and diabetes. Such a questionnaire was developed and implemented into the data collection scheme of the prospective Pima Indian Study of Arizona. The questionnaire, which assesses historical, past-year, and past-week leisure and occupational activity, was examined in 29 Pima individuals aged 21-36 yr and was shown to be reliable with test-retest correlations (rank-order correlations ranged from 0.62 to 0.96 for leisure and occupational activity). Reproducibility of the past-year leisure physical-activity estimate was determined in 69 participants aged 10-59 yr and was found to be reliable in all age-groups with the exception of the 10- to 14-yr-old age-group (rank-order correlations were 0.31 in the 10- to 14-yr-old age-group compared to 0.88 to 0.92 in those greater than 20 yr of age). Validity of the current-activity section of the questionnaire was demonstrated indirectly through comparisons with activity monitors. The past-week leisure-activity estimate was related to the Caltrac activity monitor counts per hour (rho = 0.62, P less than 0.05, n = 17). In summary, a physical-activity questionnaire has been developed that is both reliable and feasible to use in the Pima Indian population to evaluate the relationship of physical activity to non-insulin-dependent diabetes mellitus.

The effect of Indian or Anglo dietary preference on the incidence of diabetes in Pima Indians.

OBJECTIVE: In short-term studies, adoption of a traditional diet is associated with reduction in metabolic abnormalities often found in populations experiencing rapid lifestyle changes. We examined the long-term effects of a self-assessed traditional or nontraditional dietary pattern on the development of type 2 diabetes in 165 nondiabetic Pima Indians. RESEARCH DESIGN AND METHODS: Dietary intake was assessed in 1988 by a quantitative food frequency method, and subjects were asked to classify their diet as "Indian," "Anglo," or "mixed." The Indian diet reflects a preference for Sonoran-style and traditional desert foods. The Anglo diet reflects a preference for non-Sonoran-style foods typical of the remaining regions of the U.S. RESULTS: In women, the intake of complex carbohydrates, dietary fiber, insoluble fiber, vegetable proteins, and the proportion of total calories from complex carbohydrate and vegetable proteins were significantly higher (P < 0.05) in the Indian than in the Anglo diet. The mixed diet was intermediate in of all these constituents. In men, the intake for these nutrients was also higher in the Indian than in the Anglo group, but not significantly. Diabetes developed in 36 subjects (8 men and 28 women) during 6.2 years of follow-up (range 0.9-10.9). The crude incidence rates of diabetes were 23. 35, and 63 cases per 1,000 person-years in the Indian. mixed, and Anglo groups, respectively. After adjustment for age, sex, BMI, and total energy intake in a proportional hazards model, the risk of developing diabetes in the Anglo-diet group was 2.5 times as high (95%) CI 0.9-7.2) and the rate in the mixed-diet group was 1.3 times as high (0.6-3.3) as in the Indian-diet group. CONCLUSIONS: This study suggests that the adoption of an Anglo diet may increase the risk of developing diabetes in Pima Indians, but it does not provide unequivocal evidence for or against this hypothesis.

Association of physical activity and serum insulin concentrations in two populations at high risk for type 2 diabetes but differing by BMI.

OBJECTIVE: Physical activity and insulin sensitivity are related in epidemiological studies, but the consistency of this finding among populations that greatly differ in body size is uncertain. The present multiethnic epidemiological study examined whether physical activity was related to insulin concentrations in two populations at high risk for diabetes that greatly differ by location, ethnic group, and BMI. RESEARCH DESIGN AND METHODS: The study populations consisted of 2,321 nondiabetic Pima Indian men and women aged 15-59 years from Arizona and 2,716 nondiabetic men and women aged 35-54 years from Mauritius. Insulin sensitivity was estimated by mean insulin concentration (average of the fasting and postload insulin), and total (i.e., leisure and occupational) physical activity was assessed by questionnaire. RESULTS: Pima men and women who were more active had significantly (P < 0.05) lower mean insulin concentrations than those less active (BMI and age-adjusted means were 179 vs. 200 and 237 vs. 268 pmol/l). Similar findings were noted in Mauritian men and women (94 vs. 122 and 127 vs. 148 pmol/l). In both populations, activity remained significantly associated with mean insulin concentration controlled for age, BMI, waist-to-thigh or waist-to-hip ratio, and mean glucose concentrations. CONCLUSIONS: Physical activity was negatively associated with insulin concentrations both in the Pima Indians, who tend to be overweight, and in Mauritians, who are leaner. These findings suggest a beneficial role of activity on insulin sensitivity that is separate from any influence of activity on body composition.

Weight threshold and blood pressure in a lean black population.

Hypertension is virtually absent in very lean rural African populations but is becoming more common in higher-weight urban African populations and is very common in predominantly obese Westernized black populations. This implies that there is a threshold above which weight is related to blood pressure. We studied urban Nigerian civil servants, a lean population in transition toward a more Westernized lifestyle. Blood pressure, fat-related measurements, fasting insulin, physical activity, alcohol intake, macronutrient intake, and electrolyte excretion were measured in 500 male and 299 female civil servants in Benin City, Nigeria, in 1992. Median body mass index (BMI) was 21.5 kg/m2 in men and 24.0 kg/m2 in women. Examination of age-adjusted mean blood pressure across quantiles of BMI in men and women suggested a threshold of 21.5 kg/m2 below which blood pressure was not correlated with BMI. Above this threshold blood pressure was correlated with BMI. Comparison of groups above and below the lower BMI threshold found that differences in blood pressure-BMI covariation were not explained by differences in alcohol intake, caloric or macronutrient intake, or electrolyte excretion. Physical activity was higher in men below the threshold. Fasting insulin and waist-hip ratio were strongly correlated with BMI even in this very lean population but neither was independently related to blood pressure. We conclude that there is a threshold below which little relationship between blood pressure and weight is observed. Above this threshold even at levels considered lean in US blacks, weight is a major determinant of blood pressure in this population of African blacks, which shares ancestry with US blacks.

Comparison of methods to measure physical activity in postmenopausal women.

Five methods for assessing physical activity were evaluated in a population of 255 white, postmenopausal women participating in a randomized trial on the effects of walking on bone loss. Methods were the Paffenbarger survey, a modified Paffenbarger, the large-scale integrated activity monitor (LSI), caloric intake, and the Baecke survey. Significant increase in physical activity was observed in the intervention group. Activity measures were weakly related to each other. Results of factor analyses suggest that factor I reflected voluntary leisure-time pursuits (such as walking) and factor II, activities of daily living. Participation in sports was not a significant contributor to overall activity of the women. Research on physical activity must define the particular dimension of activity measured. Several types of instruments should be used because different instruments reflect different activity patterns which, in turn, may be differentially related to disease.

Postmenopausal bone density and milk consumption in childhood and adolescence.

Skeletal mass is a major determinant of susceptibility to osteoporotic fracture in menopause. At menopause, the skeletal mass is the resultant of the Peak Skeletal Mass (PSM) reached early adulthood minus the bone mass lost through the process of Adult Bone Loss (ABL). Current interventions for the maintenance of skeletal resilience in advanced age address the ABL peri- or postmenopausally. This study indicates that the effects of milk consumption in childhood and adolescence on bone density may manifest as higher bone density decades later in menopause. The assumed mechanism of the reported effect is through augmentation of the PSM, and acquisition of favorable nutritional habits which may influence the extent of ABL.

A two year randomized exercise trial in older women: effects on HDL-cholesterol.

Most of the research on the level of high density lipoprotein cholesterol (HDL-C) and physical activity (PA) has been cross-sectional and thus self-selection of the exercisers may occur. In the current research, 229 white postmenopausal women, mean age 57.7 years, were randomized into either a walking or a control group. Of these 229 women, 204 women had blood samples available for lipid determinations. PA was measured subjectively by the Paffenbarger Survey and objectively with activity monitors. At baseline, there were no differences in PA, total HDL-C (HDL-TC), HDL-2C or HDL-3C between the two randomized groups. After two years, the PA of the walking group was significantly higher than the PA of the control group. This increase in PA was not accompanied by changes in any of the lipids or lipoproteins. Examination of the lipid changes in the walking group by compliance status and actual activity changes revealed little difference between groups. These results suggest that it is possible to increase physical activity in older women. However, the long-term effects of the increased activity on HDL-C were not apparent despite an observed strong cross-sectional relationship between PA and HDL-C.

Correlates of serum lipids in a lean black population.

Exposure to an urban, sedentary work environment and higher socioeconomic status (SES) may stimulate adoption of Westernized lifestyles by populations in developing countries reversing the historically low risk for coronary heart disease. In a study of serum lipids in 1407 Nigerian civil servants, aged 25-54 years, we found a more atherogenic lipid profile among higher SES males and females compared with lower SES (LDL-cholesterol, 113 vs. 97 mg/dl, males, 125 vs. 114 mg/dl, females). Mean body mass index (BMI, kg/m2) in higher and lower SES was 22.6 and 21.3, respectively, males, and 24.7 and 24.4, respectively females. A strong relationship was observed between BMI and lipids although this relationship was absent among the leanest half of the population (BMI < 21.8). In multiple regression, SES and BMI were both strong and independent predictors of cholesterol. Both high and low SES consumed a typical Nigerian low fat, high carbohydrate diet, but somewhat higher meat, milk and egg intake suggested that some Westernization of the diet had occurred among the higher SES. Physical activity was lower among the higher SES. We conclude that SES related changes in lifestyle contribute to substantially higher total and LDL-cholesterol even in a generally lean population consuming a low fat diet.

The association of physical activity and diabetic complications in individuals with insulin-dependent diabetes mellitus: the Epidemiology of Diabetes Complications Study--VII.

The relationship between leisure-time physical activity and diabetes complications was examined in 628 individuals with insulin-dependent diabetes mellitus. Neuropathy was based upon clinical exam; retinopathy by fundus photographs; nephropathy by urine samples and macrovascular disease by clinician diagnosis. Past week physical activity was found to be inversely related to complication status but since current inactivity could be the result of the complications, historical activity (estimated leisure activity during ages 14-17) was also examined in relation to complication status. Males reporting higher levels of historical physical activity had a significantly lower prevalence of nephropathy and neuropathy but not retinopathy as demonstrated by multivariate analysis (controlling for duration of disease, age, and current activity levels). The lack of similar findings in women may be due to their low levels of reported physical activity. The consistent relationship between historical leisure physical activity and development of complications in males suggests that activity may be protective or at least not detrimental to the individual with insulin-dependent diabetes.

Reproducibility and validity of a self-administered physical activity questionnaire.

BACKGROUND: The reproducibility and validity of self-administered questionnaires on physical activity and inactivity were examined in a random (representative) sample of the Nurses' Health Study II cohort and a random sample of African-American women in that cohort. METHODS: Repeat questionnaires were administered 2 years apart. Past-week activity recalls and 7-day activity diaries were the referent methods; these instruments were sent to participants four times over a 1-year period. RESULTS: The 2-year test-retest correlation for activity was 0.59 for the representative sample (n = 147) and 0.39 for the African-American sample (n = 84). Correlations between activity reported on recalls and that reported on questionnaire were 0.79 and 0.83 for the representative and African-American samples, respectively. Correlations between activity reported in diaries and that reported on questionnaire were 0.62 and 0.59, respectively. Test-retest coefficients for inactivity were 0.52 and 0.55, respectively. Correlations between inactivity reported in diaries and that reported on questionnaire were 0.41 and 0.44, respectively. CONCLUSION: The simple, short questionnaires on activity and inactivity used in the Nurses' Health Study II are reasonably valid measures for epidemiological research.

A new approach of measuring muscle impairment during a functional task: quadriceps muscle activity recorded during chair stand.

BACKGROUND: Biologic changes are expected to occur prior to disability. Compared with physical disability measures, measures of muscle impairment may be an earlier indicator of functional decline. The purpose of this study was to describe a new approach of measuring muscle impairment during a functional task. METHODS: Right quadriceps muscle activity was recorded using surface electromyography (sEMG) from 160 older women (age 73.9 +/- 3.9 years, mean +/- SD). Specific patterns of muscle activity during the chair stand task were determined using an exploratory principal components factor analysis (PCFA). Muscle activity parameters were validated by comparison to the Physical Performance Test, gait speed, and the Functional Status Questionnaire. RESULTS: The PCFA indicated two factors (magnitude and timing) that represented important components of quadriceps muscle activity during chair stand, explaining 68.6% of the variance in performance. The slope of the rise of muscle activity represents a combination of the magnitude and timing components of muscle activity. Compared with women with a slope <1, women with a slope > or = 1 walked faster (1.17 m/s vs 1.09 m/s; p = .02) and reported less difficulty with activities of daily living (ADL) (98.6 vs 95.8; p = .003) and instrumental ADL (97.3 vs 92.2; p = .001). CONCLUSIONS: Quadriceps muscle activity recorded during chair stand is a valid and reliable measure of muscle performance during a functional task. As a biologic measure of muscle activation, sEMG may identify muscle impairment, which could indicate functional decline earlier than measures of functional status.

The role of socioeconomic status and injury morbidity risk in adolescents.

OBJECTIVE: To examine the patterns of socioeconomic status and injury morbidity in adolescents. DESIGN: Prospective cohort study. SETTING: Metropolitan school district in Allegheny County, Pennsylvania. PARTICIPANTS: There were 1245 students (89%) recruited from 1400 aged 12 to 16 years. INTERVENTIONS: None. MEASURES AND RESULTS: Baseline variables of socioeconomic status used for this analysis included township of residence (from the 1990 census data ranked by the percentage of households below the poverty level) and employment status of the parents (student self-report). Life-table analysis revealed no differential risk of injury by socioeconomic status. Similar results were found when stratified by gender, race, injury type (sport vs nonsport), and injury place (school vs home). CONCLUSIONS: The results indicate that in this cohort of adolescents, socioeconomic status does not seem to be a contributing risk factor for injury.

Physical activity and reduced risk of ovarian cancer.

OBJECTIVE: To examine the relation between leisure-time physical activity and ovarian cancer. METHODS: We used data from a population based case-control study. Cases (n = 767) were women 20-69 years of age in whom epithelial ovarian cancer was diagnosed during 1994-1998 and who resided in a defined region of Pennsylvania, New Jersey, or Delaware. Controls (n = 1367) were ascertained by using random-digit dialing and Health Care Financing Administration files and were frequency-matched to cases for age and county of residence. Information on lifetime leisure-time physical activity was obtained during in-person interviews. RESULTS: Leisure-time physical activity was significantly associated with reduced occurrence of ovarian cancer (P =.01). After adjustment for age, parity, oral contraceptive use, tubal ligation, family history of ovarian cancer, race, and body mass index, women with the highest level of activity had an odds ratio of 0.73 (95% CI 0.56, 0.94) for ovarian cancer compared with women with the lowest level of activity. When the relation was analyzed by various recalled time periods during life, the odds ratios for the highest versus the lowest category of activity at ages 14-17, 18-21, 22-29, 30-39, 40-49, and >50 years ranged from 0.64-0.78. CONCLUSION: Leisure-time physical activity is associated with reduced occurrence of epithelial ovarian cancer.