SEARCH

SEARCH BY CITATION

Keywords:

  • weight gain;
  • physical activity;
  • sedentary;
  • postmenopausal;
  • women

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective: To assess the relationship among recreational physical activity (PA), non-occupational sedentary behavior, and 7-year weight gain among postmenopausal U.S. women 40 to 69 years old.

Research Methods and Procedures: In 1992 and 1999, 18,583 healthy female participants from the Cancer Prevention Study II Nutrition Cohort completed questionnaires on anthropometric characteristics and lifestyle factors. The associations between recreational PA [in metabolic equivalent (MET) hours per week] and non-occupational sedentary behavior (in hours per day) at baseline and risk for 7-year weight gain (5 to 9 or ≥10 vs. ±4 pounds) were assessed using multivariate logistic regression analysis.

Results: Neither PA nor sedentary behavior was associated with a 5- to 9-pound weight gain. Among women who were not overweight at baseline (BMI <25.0), the odds of ≥10-pound weight gain were 12% lower (odds ratio, 0.88; 95% confidence interval, 0.77 to 0.99) for those in the highest category of recreational PA (≥18 MET h/wk) compared with >0 to <4 MET h/wk; odds were 47% higher (odds ratio, 1.47; 95% confidence interval, 1.21 to 1.79) for non-overweight women who reported ≥6 h/d of non-occupational sedentary behavior compared with <3 h/d. Neither PA nor sedentary behavior were associated with risk of ≥10-pound weight gain weight among women who were overweight at baseline (BMI ≥25.0).

Discussion: Both recreational PA and non-occupational sedentary behavior independently predicted risk of ≥10-pound weight gain among postmenopausal women who were not overweight at baseline. Public health messages to prevent weight gain among normal-weight postmenopausal women may need to focus on decreasing time spent in sedentary behaviors and increasing the amount of time spent on PA.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The prevalence of overweight and obesity is increasing among all age groups in the U.S. (1), and the increase among middle-age and older women has been striking. In an analysis of U.S. national survey data, researchers found that, on average, women 40 to 49 years old were 25.4 pounds heavier, women 50 to 59 years old were 22.7 pounds heavier, and women 60 to 74 years old were 17.4 pounds heavier in 2002 than in 1960 (2). Although weight gain among white U.S. women had tended to occur before middle age (3)(4), results of an analysis of data from the Nurses’ Health Study showed that 7.5% of non-obese women 46 to 71 years old at baseline became obese over the 6-year follow-up period between 1992 and 1998 (5).

Prospective studies of women have shown that weight gains of 10 to 15 pounds were associated with an increased risk for coronary heart disease (6) and diabetes (7)(8), and weight gains of 20 to 45 pounds were associated with an increased risk for ischemic stroke (9) and postmenopausal breast cancer (10)(11)(12). To prevent such weight gain and the health risks associated with them, we need to identify modifiable lifestyle behaviors that contribute to weight gain.

In recent decades, there has been renewed interest in how patterns of physical activity (PA)1 and sedentary behavior are related to weight changes and to various health consequences (13)(14)(15). However, the majority of research has been devoted to identifying associations between health outcomes and various types of PA (e.g., exercise, physical fitness, sports), whereas associations with sedentary behavior [e.g., television (TV) viewing] have been less explored and are usually inferred from the absence of activity (16). In addition, most studies have had relatively small sample sizes, and many have included both younger and older adults. Two large longitudinal studies specifically of middle-age American women have assessed weight change and both PA and sedentary behavior. Results of the Study of Women's Health Across the Nation, an observational study conducted in seven U.S. sites, in which 3064 women 42 to 54 years old were followed for 3 years, showed that women's baseline PA and daily routine (more transportation activity and less TV) were both inversely related to changes in body weight (17). In another study of 50,277 American women 46 to 71 years old (the Nurses’ Health Study), researchers found that time spent daily in brisk walking was inversely associated and time engaged in sedentary behavior, such as time spent watching TV, was positively associated with the development of obesity over a 6-year period (5). To add to this evidence base of observational studies in middle-age women, we used data from the nutrition cohort of the American Cancer Society (ACS) Cancer Prevention Study (CPS) II, a large observational study of lifestyle and cancer risk, to examine the independent associations of recreational PA and non-occupational sedentary behavior with odds of weight gain over 7 years among a prospective cohort of healthy postmenopausal women 40 to 69 years old.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Study Sample

As previously noted, women in this study were members of the Nutrition Cohort of the ACS CPS II, a prospective study of cancer incidence and mortality among 86,404 men and 97,786 women (18). The Nutrition Cohort was begun by the ACS in 1992 and is a subgroup of the ∼1.2 million participants in CPS II, a prospective study of cancer mortality established in 1982 (19). In brief, Nutrition Cohort participants were recruited from members of the CPS II cohort who resided in 21 states with population-based cancer registries; their recruitment has been described in detail elsewhere (18). Participants were 40 to 74 years of age at enrollment in 1992 or 1993, when they completed a 10-page confidential, self-administered mailed questionnaire that included questions on demographic, medical, lifestyle, and dietary factors.

Participants were mailed follow-up surveys in 1997 and 1999 to update information on their behavior (e.g., smoking, PA) and disease status. Non-responders to the full survey were mailed shorter questionnaires on disease outcome and limited exposure information or called by telephone. The overall follow-up was 91%. The Emory University School of Medicine Institutional Review Board approved all aspects of the CPS II Nutrition Cohort research.

For this analysis, we excluded those women who were lost to follow-up from 1992 to 1999 (n = 9177) or who were missing the key variables due to not completing the long follow-up survey questionnaire (n = 7573). We also excluded women with a prior diagnosis of a serious or chronic disease including diabetes, heart disease, stroke, respiratory disease, and a cancer history other than non-melanoma skin cancer (n = 24,373) and women who developed these conditions during the follow-up period (n = 20,636) because these diseases may have been associated with changes in weight and/or usual behavior (4)(20). To avoid influence of recent cessation of smoking on weight gain (21), we also excluded women who had stopped smoking cigarettes during the interval and those for whom smoking cessation data were missing (n = 2841). We excluded women over age 70 (n = 2860) because the age-related loss of lean body mass could attenuate detection of an association between weight change and behaviors (20). We further excluded 2106 women who were perimenopausal or premenopausal at baseline and 78 whose baseline menopausal status was missing. In addition, we excluded women with missing baseline data on weight (n = 2625), height (n = 280), PA (n = 973), sedentary behavior (n = 618), education (n = 328), or hormone replacement therapy (HRT) (n = 285), and those who had extreme self-reported weight, height, and/or BMI values (i.e., below the 0.1 or above the 99th percentile of measured values in National Health and Nutrition Examination Study III, defined as weight <55 or >473 pounds, height <49 or >74.8 inches, BMI <10 or >80; n = 136). We excluded women who lost ≥5 pounds during the follow-up (n = 4314) because our primary objective was to identify factors associated with weight gain (not weight loss), and we were not able to determine whether weight loss was intentional or unintentional (for example, due to the effect of illness). The remaining 18,583 postmenopausal women 40 to 69 years old who were free of chronic disease constituted the analytic sample.

Outcome Variable: Weight Gain

We defined weight gain as the difference between survey participants’ self-reported weight (in pounds) at follow-up (1999) and baseline (1992).

Exposure Variables

The major independent variables we examined were recreational PA and non-occupational sedentary behavior. We ascertained participants’ baseline recreational PA from their response to the question, “During the past year, what was the average time per week you spent at the following kinds of activities: walking; jogging/running; lap swimming; tennis or racquetball; bicycling/stationary bike; aerobics/calisthenics; and dancing?” For each of the seven activities, response options were none, 1 to 3 h/wk, 4 to 6 h/wk, or 7+ h/wk. Summary metabolic equivalent (MET) hours per week were calculated for each participant. A MET is the ratio of metabolic rate during a specific activity to resting metabolic rate (22). The summary MET hours per week for each participant were calculated by multiplying the hours spent engaged in each activity times the MET score (22). Due to the older age of this population, MET hours per week were calculated using the lowest value from the range of hours provided for each category (i.e., 0 for none, 1 for 1 to 3 h/wk, 4 for 4 to 6 h/wk, and 7 for 7+ h per week). The following MET scores were used: 3.5 for walking, 7.0 for jogging/running, 7.0 for lap swimming, 6.0 for tennis or racquetball, 4.0 for bicycling/stationary bike, 4.5 for aerobics/calisthenics, and 3.5 for dancing, such that summary measures would be estimated conservatively (22)(23). We estimated quartiles of MET hours per week for those who reported some recreational activity: >0 to <4.0 (referent group), 4.0 to <10.0, 10.0 to <18.0, and ≥18.0. Women who reported 0 hours for all activities were placed in a fifth (zero) MET category.

Participants were also asked about their non-recreational activities: “During the past year, what was the average time per week you spent at the following kinds of activities: gardening, mowing, planting, etc.; heavy housework, vacuuming, etc.; heavy home repair, painting, etc.; and shopping?” For each of the four activity groupings, response options were none, 1 to 3 h/wk, 4 to 6 h/wk, and 7+ h/wk. The following MET scores were used: 3.0 for gardening, mowing, planting, etc.; 2.5 for heavy housework, vacuuming, etc.; 3.0 for heavy home repair, painting, etc.; and 2.5 for shopping. Again, we used the lowest number of hours within each response option to calculate summary MET hours per week for each participant. We divided participants into three categories of non-recreational activity based on distribution of responses: 0 to 5.0, >5.0 to 17.0 (referent group), and >17.0 MET h/wk.

To assess the amount of time participants engaged in non-occupational sedentary behavior, the survey asked “During the past year, on an average day (not counting time spent at your job), how many hours per day did you spend sitting (watching TV, reading, etc.)?” The response options in hours per day were: none, <3, 3 to 5, 6 to 8, and >8. We collapsed the two highest categories to create a three-level variable (<3 hours, ≥3 to 5 hours, ≥6 h/d).

Covariates and Modifiers

Baseline age was calculated as the difference between the date of baseline questionnaire and the year of birth. Because the majority of participants were white (98.0%), race was not included as a covariate. Height and education status (≤high school, some college or technical school, college graduate, graduate school) were self-reported in 1982. Baseline BMI was calculated as self-reported weight in 1992 (kilograms) divided by self-reported height in 1982 (meters)2 and categorized using standard overweight cut-off points (<25 kg/m2, non-overweight; ≥25 kg/m2, overweight) (24). A small proportion (n = 354, 1.9%) of women were underweight (BMI < 18.5) and were placed in the non-overweight category. Other 1992 baseline covariates included HRT (current, former/never) and smoking status (never, former, current). Quartiles of baseline total daily energy intake (kilocalories) were determined from participants’ responses to a modified version (n = 68 items) of the 60-item Block food frequency questionnaire (25). The food frequency questionnaire was validated using four 24-hour recalls over a 1-year period as the comparison measure (r = 0.40 for energy, women) and was shown to be highly reproducible over a 1-year period (r = 0.68 for energy, women) (26).

Data Analyses

Analyses were conducted using SAS statistical software version 9.1 (SAS, Cary NC). Statistical significance was set at p < 0.05 for all comparisons (two-sided p values). We used logistic regression to analyze the relation between the independent variables and odds of weight gain (5 to 9 or ≥10 pounds), using stable weight (±4 pounds) as the reference category for each weight gain subgroup.

Potential two-way interactions between recreational PA or sedentary behavior and selected covariates [i.e., age (above and below age 60); HRT use (current, former/never); and BMI (<25 and ≥25) were assessed by entering appropriate cross-product terms into the models. We observed a significant interaction between baseline BMI and recreational PA (5- to 9-pound gain model, p = 0.04; ≥10-pound gain model, p = 0.09) and between baseline BMI and sedentary behavior (5- to 9-pound gain model, p = 0.46; ≥10-pound gain model, p = 0.01); thus, all analyses were stratified on baseline overweight status. We also assessed the interaction between PA and non-occupational sedentary behavior, PA and non-recreational activity, and non-occupational sedentary behavior and non-recreational activity on outcomes. These interactions terms were not statistically significant.

Multivariable logistic regression models were constructed as follows: age-adjusted model included age (continuous) and both baseline PA and sedentary behavior (simultaneous predictors), and multiadjusted model further included education, smoking status, HRT, and total energy. The category of >0 to <4.0 MET h/wk of PA was used as the referent category to increase stability of the reference group and to reduce potential confounding by underlying medical conditions that were not assessed that may prevent PA, as has been observed in other analyses of this cohort (23). In secondary analyses, non-recreational activity (0 to 5.0, >5.0 to 17.0, >17.0 MET h/wk) was included in regression modeling to determine its role on risk of weight gain. However, it was not an independent predictor of weight gain, nor did it meaningfully confound or modify associations of recreational PA or sedentary behavior with weight gain; therefore, it is not included in the primary regression table results.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The average age of our analytic sample of women 40 to 69 years old (n = 18,583) was 60.1 years; 36.7% of the women reported having a college degree or attendance at graduate school, and 3.8% reported being current smokers. Among non-overweight women, 53.2% had stable weight, 25.1% gained 5 to 9 pounds, and 27.1% gained ≥10 pounds over the follow-up period; among women who were already overweight (or obese) at baseline, 39.8% had stable weight, 23.0% gained 5 to 9 pounds, and 37.2% gained ≥10 pounds over the follow-up period. One-tenth (12.2%) and 1.3% of non-overweight women progressed to overweight and obese, respectively. Almost one-quarter (23.3%) of overweight women progressed to obese during the follow-up period.

Table 1 shows the baseline characteristics by level of recreational PA and sedentary behavior for non-overweight women, and Table 2 shows the same for overweight women. In both BMI groups, patterns of sedentary behavior and recreational PA appeared to be independent. The proportion of women in the highest quartile of daily energy intake increased with more sedentary behavior, but there were no patterns of daily energy intake observed with increasing recreational PA. Physically active women in both BMI groups tended to have a higher education, and among non-overweight women, there were more former smokers in the highest recreational PA category than in the lowest recreational PA category.

Table 1.  Baseline characteristics and weight gain patterns according to MET hours per week of recreational PA and hours per day of sedentary behavior among non-overweight women* in the ACS CPS II Nutrition Cohort (1992–1999), n = 11,540
  Recreational leisure time PA (MET h/wk)Non-occupational sedentary behavior (h/d)
  • MET, metabolic equivalent; PA, physical activity; ACS, American Cancer Society; CPS, Cancer Prevention Study.

  • *

    BMI < 25.0.

 TotalNone>0 to <44 to <1010 to <18≥18<33 to <6≥6
% total 5.926.015.822.130.258.136.05.9
Age in years (mean)60.159.860.059.960.360.259.461.061.1
  Percent within each PA categoryPercent within each sedentary category
Weight change category         
 Stable (±4 lbs)53.250.252.953.652.354.653.553.250.4
 Gained 5 to 9 lbs25.126.524.125.125.725.225.325.023.4
 Gained ≥10 lbs21.723.423.021.322.120.221.221.826.2
Education         
 ≤High school29.439.731.927.530.425.628.231.230.9
 Some college or technical30.228.431.429.529.930.329.831.328.2
 College graduate25.020.922.327.924.427.025.724.123.4
 Graduate school15.411.014.515.015.417.216.413.517.5
Smoking status         
 Current smoker4.310.24.33.44.13.63.25.47.7
 Former smoker38.734.434.838.638.543.038.039.242.2
 Never smoked57.155.461.058.057.453.458.855.450.2
Hormone therapy use         
 Current use46.043.546.548.844.645.646.944.744.9
 Former use16.816.217.115.817.716.515.817.820.0
 Never used37.240.336.435.437.737.937.337.535.2
Daily total energy (kcalories)         
 ≤110030.732.230.831.031.829.431.429.829.7
 >1100 to 140026.525.426.527.626.626.126.427.123.2
 >1400 to 170019.919.920.919.519.219.819.819.821.0
 >170017.716.916.717.416.819.416.818.521.0
 Missing5.25.65.14.55.75.35.64.75.0
Recreational PA (MET h/wk)         
 None5.9100.0    5.56.08.8
 >0 to <4.026.0 100.0   25.526.627.4
 4.0 to <10.015.8  100.0  15.816.311.6
 10.0 to <18.022.1   100.0 22.322.220.3
 ≥18.030.2    100.030.928.931.9
Non-recreational activity (MET h/wk)         
 0 to 525.329.630.129.424.318.824.525.928.8
 >5 to 1750.042.949.552.551.249.751.149.442.5
 >1724.127.119.917.723.730.923.724.128.2
 Missing0.60.40.50.40.80.60.60.60.4
Non-occupational sedentary behavior         
 <3 h/d58.154.657.058.458.559.4100.0  
 3 to 5 h/d36.036.636.837.336.134.4 100.0 
 ≥6 h/d5.98.86.24.45.46.2  100.0
Table 2.  Baseline characteristics and weight gain patterns according to MET hours per week of recreational PA and hours per day of sedentary behavior among overweight women* in the ACS CPS II Nutrition Cohort (1992–1999), n = 7,043
  Recreational leisure time PA (MET h/wk)Non-occupational sedentary behavior (h/d)
  • MET, metabolic equivalent; PA, physical activity; ACS, American Cancer Society; CPS, Cancer Prevention Study.

  • *

    BMI ≥ 25.0.

 TotalNone>0 to <44 to <1010 to <18≥18<33 to <6≥6
% total 9.532.117.220.021.148.143.68.3
Age in years (mean)60.059.559.860.060.460.459.460.661.0
  Percent within each PA categoryPercent within each sedentary category
Weight change category         
 Stable (±4 lbs)39.842.738.041.140.539.438.840.541.8
 Gained 5–9 lbs23.019.224.422.124.422.223.622.323.9
 Gained ≥10 lbs37.238.137.636.835.138.437.737.234.4
BMI (kg/m2)         
 Overweight (25.0 to <30.0)75.365.672.374.978.881.378.872.072.5
 Obese (≥30.0)24.734.427.825.121.218.721.328.027.5
Education         
 ≤High school38.043.038.436.836.937.036.939.238.0
 Some college or technical31.232.031.332.529.531.231.830.630.8
 College graduate18.916.218.618.720.019.518.518.920.6
 Graduate school12.08.811.612.113.612.412.811.410.7
Smoking status         
 Current smoker3.15.13.13.32.82.62.43.74.8
 Former smoker36.234.134.437.337.637.734.138.238.3
 Never smoked60.660.962.559.459.759.663.658.156.9
Hormone therapy use         
 Current use36.133.937.936.136.034.537.534.735.4
 Former use19.422.217.920.619.219.718.719.921.1
 Never used44.543.944.243.444.845.943.945.443.5
Daily total energy (kcalories)         
 ≤110026.627.127.726.625.426.029.024.921.8
 >1100 to 140024.923.523.826.025.625.724.125.526.5
 >1400 to 170021.219.921.420.921.920.920.522.120.3
 >170022.225.022.320.822.421.621.322.526.0
 Missing5.24.54.95.74.85.85.25.15.5
Recreational PA (MET h/wk)         
 None9.5100.0    8.610.112.0
 >0 to <4.032.1 100.0   33.032.028.0
 4.0 to <10.017.2  100.0  16.817.418.2
 10.0 to <18.020.0   100.0 19.320.918.9
 ≥18.021.1    100.022.319.522.9
Non-recreational activity (MET h/wk)         
 0–525.234.528.628.322.016.623.626.329.6
 >5 to 1748.442.349.148.152.346.750.047.643.5
 >1725.722.321.922.624.836.526.025.326.3
 Missing0.60.90.41.00.90.30.40.90.7
Non-occupational sedentary behavior (h/d)         
 <348.143.349.447.146.550.8100.0  
 3 to 543.646.343.444.245.740.3 100.0 
 ≥68.310.47.28.87.88.9  100.0

Among non-overweight women, amount of time engaged in recreational PA was not related to odds of 5- to 9-pound weight gain vs. stable weight over the 7-year follow-up period (Table 3). However, overweight or obese women who reported no recreational PA had 29% lower odds of 5- to 9-pound weight gain compared with those who reported >0 to 4.0 MET h/wk of PA [odds ratio (OR), 0.71; 95% confidence interval (CI), 0.56, 0.90]. There was no association between sedentary behavior and 5- to 9-pound weight gain in either weight group.

Table 3.  Association among PA, sedentary behavior, and 5- to 9-lb weight gain (vs. stable weight, ±4 lbs) stratified by overweight status among eligible women (40 to 69 years of age) in the CPS II Nutrition Cohort
 Weight status (n)Age-adjusted model*Multi-adjusted model
  • PA, physical activity; CPS, Cancer Prevention Study; OR, odds ratio; CI, confidence interval; MET, metabolic equivalent.

  • *

    Model adjusted for age, recreational PA, and non-occupational sedentary behavior.

  • Model adjusted for age, recreational PA, non-occupational sedentary behavior, education, smoking status, hormone therapy use, and total energy.

PA and sedentary behavior categoriesStableGainOR(95% CI)OR(95% CI)
Non-overweight, BMI <25.0 kg/m2      
 Recreational leisure time PA (MET h/wk)      
  None3411801.15(0.94, 1.40)1.14(0.93, 1.40)
  >0 to <4.015887231.00 1.00 
  4.0 to <10.09744561.02(0.88, 1.17)1.03(0.89, 1.19)
  10.0 to <18.013346561.10(0.96, 1.25)1.10(0.97, 1.25)
  ≥18.019048791.02(0.91, 1.15)1.04(0.92, 1.17)
 Non-occupational sedentary behavior (h/d)      
  <3358916971.00 1.00 
  3 to 5220910381.07(0.97, 1.17)1.07(0.97, 1.18)
  ≥63431591.05(0.86, 1.29)1.06(0.87, 1.30)
Overweight, BMI ≥25.0 kg/m2      
 Recreational leisure time PA (MET h/wk)      
  None2871290.69(0.53, 0.91)0.71(0.56, 0.90)
  >0 to <4.08595531.00 1.00 
  4.0 to <10.04982670.87(0.71, 1.07)0.84(0.70, 1.01)
  10.0 to <18.05703440.89(0.74, 1.09)0.96(0.81, 1.15)
  ≥18.05873300.83(0.68, 1.00)0.89(0.75, 1.06)
 Non-occupational sedentary behavior (h/d)      
  <313147981.00 1.00 
  3 to 512446860.96(0.82, 1.11)0.95(0.83, 1.08)
  ≥62431390.99(0.76, 1.30)1.00(0.80, 1.26)

In contrast, among women who were not overweight at baseline, those in the highest category of recreational PA had lower odds of ≥10-pound weight gain compared with those with stable weight (OR, 0.88; 95% CI, 0.77, 0.99 for ≥18 MET h/wk compared with >0 to 4 MET h/wk) (Table 4). Sedentary behavior was also related to ≥10-pound weight gain among women who were not overweight at the start of follow-up. The odds of ≥10-pound weight gain were 16% higher among women who reported 3 to 5 h/d of sedentary behavior (OR, 1.16; 95% CI, 1.04, 1.28) and 47% higher among women who reported ≥6 h/d of sedentary behavior (OR, 1.47; 95% CI, 1.21, 1.79) than among those who reported <3 h/d; however, sedentary behavior was not significantly associated with weight gain among women who were already overweight or obese at baseline.

Table 4.  Association between recreational PA, sedentary behavior, and ≥10 pound weight gain (vs. stable weight, ±4 lbs) stratified by overweight status among eligible women (40 to 69 years or age) in the CPS II Nutrition Cohort
 Weight status (n)Age-adjusted model*Multi-adjusted model
  • PA, physical activity; CPS, Cancer Prevention Study; OR, odds ratio; CI, confidence interval; MET, metabolic equivalent.

  • *

    Model adjusted for age, recreational PA, and non-occupational sedentary behavior.

  • Model adjusted for age, recreational PA, non-occupational sedentary behavior, education, smoking status, hormone therapy use, and total energy

PA and sedentary behavior categoriesStableGainOR(95% CI)OR(95% CI)
Non-overweight, BMI <25.0      
 Recreational leisure time PA (MET h/wk)      
  None3411591.04(0.84, 1.29)1.01(0.82, 1.25)
  >0 to <4.015886901.00 1.00 
  4.0 to <10.09743880.92(0.79, 1.07)0.93(0.80, 1.08)
  10.0 to <18.013345630.99(0.87, 1.13)0.99(0.87, 1.14)
  ≥18.019047050.87(0.76, 0.98)0.88(0.77, 0.99)
 Non-occupational sedentary behavior (h/d)      
  <3358914211.00 1.00 
  3 to 522099061.16(1.05, 1.29)1.16(1.04, 1.28)
  ≥63431781.47(1.21, 1.80)1.47(1.21, 1.79)
Overweight, BMI ≥25.0 kg/m2      
 Recreational leisure time PA (MET h/wk)      
  None2872560.72(0.57, 0.92)0.88(0.72, 1.07)
  >0 to <4.08598511.00 1.00 
  4.0 to <10.04984460.88(0.73, 1.07)0.91(0.78, 1.08)
  10.0 to <18.05704940.90(0.75, 1.08)0.91(0.78, 1.06)
  ≥18.05875721.03(0.87, 1.23)1.03(0.88, 1.20)
 Non-occupational sedentary behavior (h/d)      
  <3131412771.00 1.00 
  3 to 5124411420.99(0.87, 1.13)1.05(0.94, 1.18)
  ≥62432000.96(0.75, 1.22)0.98(0.80, 1.21)

Results of sensitivity analyses that excluded current smokers (n = 711) were similar to those among the full sample. Inclusion of women who lost weight as the referent group did not meaningfully change our findings. Further subgroup analyses of women with baseline BMI of 25.0 to 29.9 and women with baseline BMI ≥ 30.0 (obese women) found similar results to those presented for women with baseline BMI ≥ 25.0; however, we do not present these results due to small cell sizes among the obese subgroup. Modeling weight gain (>0 pounds during follow-up) in a continuous manner using linear regression with models that stratified on baseline weight status (BMI < 25, ≥25) found that among women who were not overweight at baseline, recreational PA (β, 0.95; lower bound, 0.92; upper bound, 0.99; for ≥18 MET h/wk) and non-recreational activity (β, 0.95; lower bound, 0.92; upper bound, 0.99; for >17 MET h/wk) were inversely related to weight gain and the highest level of sedentary behavior (β, 1.11; lower bound, 1.05; upper bound, 1.09; for ≥6 h/d) was related to weight gain. With the exception of non-recreational activity (β, 0.94; lower bound, 0.89; upper bound, 0.98; for >17 MET h/wk), we did not observe these same relationships with weight gain among women who were already overweight at baseline.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Among normal-weight postmenopausal women, those who reported spending the most time engaged in sedentary behaviors, such as watching TV and reading, had almost 50% greater odds of gaining at least 10 pounds over the 7 years compared with those who reported little sedentary activity. Non-overweight postmenopausal women who were the most physically active had 12% lower odds of gaining at least 10 pounds. These associations remained after adjustment for other lifestyle factors including education, smoking status, HRT use, and energy intake.

Because the associations between non-occupational sedentary behavior and recreational PA on ≥10-pound weight gain were independent of one another, the increased risk we observed among women who spent more time on sedentary behaviors cannot be explained by lower levels of PA alone, and the small reduction in risk of weight gain with highest levels of PA is not due simply to avoidance of sedentary behavior. This finding of independence is similar to that observed for development of obesity in the Nurses’ Health Study (5) and in an analysis of 4-year weight change among men 45 to 54 years old (27).

Clinical trial results have shown that lifestyle intervention, both dietary and PA, can prevent weight gain during menopause (28). Several observational studies of free-living adults have also found inverse relationships between greater PA and weight gain (20)(27)(29)(30)(31)(32). Brown et al. (30) studied 8071 women 45 to 55 years of age over a 5-year time period (baseline survey, 1996) and found independent relationships between the odds of gaining >5 kg and lower levels of habitual PA and more time spent sitting. Lahmann et al. (31) studied 5464 Swedish women 45 to 73 years old and found a relationship between weight gain (from recalled weight at age 20 to measured weight between 1994–1996) and sedentary leisure activity. Haapanen et al. (32) used a single-item self-assessment of leisure-time PA and found that compared with women who reported vigorous PA twice or more per week in 1980, women who reported vigorous PA once per week and some light activity or no regular weekly PA had increased odds of a 5-kg weight gain and an overweight BMI at the end of a 10-year follow-up in 1990 (32). However, not all observational studies have found positive associations (33). Guthrie (33) assessed 5-year weight change among 233 volunteers from a population-based cohort of Australian-born women 46 to 57 years old and found no association between weight change and baseline weight, exercise, alcohol intake, or smoking.

The 2005 Dietary Guidelines for Americans state that individuals need to engage in at least 30 minutes of PA on most days of the week for prevention of chronic disease and to help manage body weight (34). To prevent gradual, unhealthy body weight gain in adulthood, individuals are encouraged to engage in ∼60 minutes of moderate- to vigorous-intensity activity on most days of the week, and, further, to sustain weight loss in adulthood, they are encouraged to participate in at least 60 to 90 minutes of daily moderate-intensity PA while not exceeding caloric intake requirements (34). In our analysis, the highest quartile of recreational PA (≥18 MET h/wk) corresponded to ∼45 minutes of daily walking, a moderate activity. Thus, our findings may indicate that in women with excess weight, 45 minutes of recreational PA is not sufficient to prevent weight gain and suggest that additional activity, dietary change, and/or behavior modification may be necessary for prevention of further unhealthy weight gain. The latter often involves strategies that reinforce healthy diet and PA, including assessing weight loss readiness, self-monitoring, and health education (24)(35)(36). An evaluation of the impact of PA on weight in a work site setting found that employees who took part in a 3-month health-related fitness curriculum and also engaged in habitual PA (≥2 years) had reduced weight after 5 years (37). Informational, behavioral and social, and environmental and policy approaches are recommended by The Guide to Preventive Community Services to increase physical activity (38). Thus, work sites and communities interested in increasing PA among members/employees are encouraged to combine education with building exercise facilities, provide access to nearby exercise facilities, and/or build walking trails (38).

The most common activities among physically active women in our sample, regardless of age, included walking followed by biking, with the addition of modest amounts of the other five reported activities. These findings of low amounts of vigorous activity are similar to those found during the same period in the Women's Health Initiative Observational Cohort Study (1993 to 1998), which assessed vigorous activity retrospectively for ages 18, 35, and 50 years and currently at enrollment into the study (median age, 65 years) (39). Current participation in vigorous activity (>3 days/wk) was low and consistent across racial/ethnic groups (13% to 16%) in the Women's Health Initiative study group. The prevalence of vigorous activity declined with age, with the largest decrease in vigorous activity occurring after age 50 years for all racial/ethnic groups (39). A recent prospective study of vigorous PA among 4871 female runners found that age-related weight gain occurred even among the most active individuals when exercise was constant (40). Using regression slopes, the authors determined that vigorous exercise may need to increase annually in women to compensate for the expected gain in weight associated with aging. The levels of activity assessed among most middle-age women in our sample do not seem to be of sufficient amount or vigor to compensate for the expected gain in weight associated with aging.

In the area of PA and weight, there are at least three distinct hypotheses tested by researchers, including the amount of PA needed for prevention of weight gain, for weight loss, and for weight maintenance. In our paper, we sought to address only the prevention of weight gain. Our stratification by baseline weight also allowed us to determine the relation between varying PA levels (in MET hours per week) and sedentary behavior among two groups of postmenopausal women: those not yet overweight and those with excess weight. Our results suggest that recreational PA, such as daily walking or biking, may prevent ≥10-pound weight gain among women who are not overweight or obese. However, we failed to find this relation between either PA or sedentary behavior and ≥10-pound weight gain among overweight postmenopausal women. The Study of Women's Health Across the Nation authors (17) and Wier et al. (37) also examined the potential interaction between weight change and baseline BMI but found no meaningful effect modification (17). Our larger sample size may have allowed us to detect this interaction. At least one other study found that initial BMI was a factor that explained a large part of the variation in weight change over time (31).

We also found lower odds of weight gain with physical inactivity among overweight women, which was an unexpected finding. It is possible that these overweight women had more error in reporting activities compared with non-overweight women (41). It is possible that with their already elevated risk of chronic disease due to obesity, non-physically active overweight women may have experienced less weight gain due to underlying morbidity or illness (23) (which may cause a lack of weight gain in the absence of exercise) that we did not capture in our questionnaires.

Our survey also asked subjects about non-recreational activities including time spent engaged in gardening, vacuuming, painting, and shopping. These activities may contribute largely to a woman's total energy expenditure (42). However, there was neither an independent relationship between these activities and weight gain, nor did non-recreational PA modify the association between recreational PA or sedentary behavior and weight gain in logistic regression models. This is in contrast to earlier reported findings from this cohort that found lower odds of 10-year BMI change from 1982 to 1992 among women who reported gardening, mowing, and/or planting 4 h/wk at both age 40 (i.e., ≤1982) and in 1992 compared with women who reported no such activities at the two time-points (29). An explanation may be that subjects performed household activities with varying levels of intensity. The lack of information on the intensity of individual behavior increases possible misclassification of summary MET scores. In addition, the lack of association with non-recreational activities might also be due to the lack of precision with which these activities are reported.

The observational nature and reliance on self-reported activity and other lifestyle factors is a limitation of the study. It is known that under-reporting of weight among survey participants can occur, especially among overweight individuals (43). The PA questions we used have not been validated in this cohort and are subject to misreporting; however, they are very similar to those used and validated by the Nurses’ Health Study, which found strong correlations between activity reported on past week activity recalls and 7-day diaries and activity reported on the questionnaire (0.79 and 0.62, respectively) (44). Specifically, the Nurses’ Health Study asked about the average time spent per week walking, jogging, running, bicycling, swimming laps, playing tennis or squash, and participating in calisthenics, a similar list of activities to that of the ACS Nutrition Cohort, with the exception that nutrition cohort members were also asked about dancing.

Our findings in normal-weight postmenopausal women support the idea that participating in physical activities and sedentary behavior represent separate aspects of lifestyle. Therefore, to prevent weight gain among normal-weight postmenopausal women, public health strategies may need to focus on both reducing the time women spend engaging in certain sedentary behaviors, such as TV viewing, and increasing the time they spend engaging in recreational PA.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the agencies with which they are affiliated. We thank Kimberly Walker-Thurmond for help in data management and analytic sample creation and Janet Fulton and Harold W. Kohl III for input on data analysis. There was no funding/outside support for this study.

Footnotes
  • 1

    Nonstandard abbreviations: PA, physical activity; TV, television; ACS, American Cancer Society; CPS, Cancer Prevention Study; HRT, hormone replacement therapy; MET, metabolic equivalent; OR, odds ratio; CI, confidence interval.

  • The costs of publication of this article were defrayed, in part, by the payment of page charges. This article must, therefore, be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Ogden, C. L., Carroll, M. D., Curtin, L. R., McDowell, M. A., Tabak, C. J., Flegal, KM (2006) Prevalence of overweight and obesity in the United States, 1999–2004. JAMA 295: 15491555.
  • 2
    Ogden, C. L., Fryar, C. D., Carroll, M. D., Flegal, KM (2004) Mean body weight, height, and body mass index, United States 1960–2002. Adv Data 347: 117.
  • 3
    Kumanyika, S. (1987) Obesity in black women. Epidemiol Rev. 9: 3150.
  • 4
    Colditz, G. A., Willett, W. C., Stampfer, M. J., London, S. J., Segal, M. R., Speizer, FE (1990) Patterns of weight change and their relation to diet in a cohort of healthy women. Am J Clin Nutr. 51: 11001105.
  • 5
    Hu, F. B., Li, T. Y., Colditz, G. A., Willett, W. C., Manson, JE (2003) Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women. JAMA 289: 17851791.
  • 6
    Willett, W. C., Manson, J. E., Stampfer, M. J., et al (1995) Weight, weight change, and coronary heart disease in women: risk within the “normal” weight range. JAMA 273: 461465.
  • 7
    Resnick, H. E., Valsania, P., Halter, J. B., Lin, X. (2000) Relation of weight gain and weight loss on subsequent diabetes risk in overweight adults. J Epidemiol Community Health 54: 596602.
  • 8
    Colditz, G. A., Willett, W. C., Rotnitzky, A., Manson, JE (1995) Weight gain as a risk factor for clinical diabetes mellitus in women. Ann Intern Med. 122: 481486.
  • 9
    Rexrode, K. M., Hennekens, C. H., Willett, W. C., et al (1997) A prospective study of body mass index, weight change, and risk of stroke in women. JAMA 277: 15391545.
  • 10
    Huang, Z., Hankinson, S. E., Colditz, G. A., et al (1997) Dual effects of weight and weight gain on breast cancer risk. JAMA 278: 14071411.
  • 11
    Lahmann, P. H., Lissner, L., Gullberg, B., Olsson, H., Berglund, G. (2003) A prospective study of adiposity and postmenopausal breast cancer risk: the Malmo Diet and Cancer Study. Int J Cancer 103: 246252.
  • 12
    Feigelson, H. S., Jonas, C. R., Teras, L. R., Thun, M. J., Calle, EE (2004) Weight gain, body mass index, hormone replacement therapy, and postmenopausal breast cancer in a large prospective study. Cancer Epidemiol Biomarkers Prev. 13: 220224.
  • 13
    U.S. Department of Health & Human Services (Accessed April 30, 2007). Reports of the Surgeon General, U.S. Public Health Service. Report on Physical Activity and Health, 1996. http:www.surgeongeneral.govlibraryreports.htm
  • 14
    Fogelholm, M., Kukkonen-Harjula, K. (2000) Does physical activity prevent weight gain: a systematic review. Obes Rev. 1: 95111.
  • 15
    Saris, W. H. M., Blair, S. N., van Baak, M. A., et al (2003) How much physical activity is enough to prevent unhealthy weight gain? Outcome of the IASO 1st Stock Conference and consensus statement. Obes Rev. 4: 101114.
  • 16
    Livingstone, M. B., Robson, P. J., Wallace, J. M., McKinley, MC (2003) How active are we? Levels of routine physical activity in children and adults. Proc Nutr Soc. 62: 681701.
  • 17
    Sternfeld, B., Wang, H., Quesenberry, C. P., Jr, et al (2004) Physical activity and changes in weight and waist circumference in midlife women: findings from the Study of Women's Health Across the Nation. Am J Epidemiol. 160: 912922.
  • 18
    Calle, E. E., Rodriguez, C., Jacobs, E. J., et al (2002) The American Cancer Society Cancer Prevention Study II Nutrition Cohort: rationale, study design, and baseline characteristics. Cancer 94: 24902501.
  • 19
    Thun, M. J., Calle, E. E., Rodriguez, C., Wingo, PA (2000) Epidemiological research at the American Cancer Society. Cancer Epidemiol Biomarkers Prev. 9: 861868.
  • 20
    Dziura, J., de Leon, Mendes C., Kasl, S., DiPietro, L. (2004) Can physical activity attenuate aging-related weight loss in older people? The Yale Health and Aging Study, 1982–1994. Am J Epidemiol. 159: 759767.
  • 21
    Williamson, D. F., Kahn, H. S., Remington, P. L., Anda, RF (1990) The 10-year incidence of overweight and major weight gain in U.S. adults. Arch Intern Med. 150: 665672.
  • 22
    Ainsworth, B. E., Haskell, W. L., Leon, A. S., et al (1993) Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc. 25: 7180.
  • 23
    Patel, A. V., Callel, E. E., Bernstein, L., Wu, A. H., Thun, MJ (2003) Recreational physical activity and risk of postmenopausal breast cancer in a large cohort of US women. Cancer Causes Control 14: 519529.
  • 24
    National Heart, Lung, and Blood Institute, National Institute of Diabetes and Digestive and Kidney Diseases (1998) Obesity education initiative. Clinical Guidelines on the Identification, Education, and Treatment of Overweight and Obesity in Adults: The Evidence Report: NIH Publication No. 98-4083. Bethesda, MD: National Heart, Lung, and Blood Institute, National Institute of Diabetes and Digestive and Kidney Diseases.
  • 25
    Block, G., Hartman, A. M., Dresser, C. M., Carroll, M. D., Gannon, J., Gardner, L. (1986) A data-based approach to diet questionnaire design and testing. Am J Epidemiol. 124: 453469.
  • 26
    Flagg, E., Coates, R., Calle, E., Potischman, N., Thun, M. (2000) Validation of the American Cancer Society Cancer Prevention Study II Nutrition Survey Cohort food frequency questionnaire. Epidemiology 11: 462468.
  • 27
    Coakley, E. H., Rimm, E. B., Colditz, G., Kawachi, I., Willett, W. (1998) Predictors of weight change in men: results from the Health Professionals Follow-up Study. Int J Obes Relat Metab Disord. 22: 8996.
  • 28
    Simkin-Silverman, L. R., Wing, R. R., Boraz, M. A., Kuller, LH (2003) Lifestyle intervention can prevent weight gain during menopause: results from a 5-year randomized clinical trial. Ann Behav Med. 26: 212220.
  • 29
    Kahn, H. S., Tatham, L. M., Rodriguez, C., Calle, E. E., Thun, M. J., Heath, C. W., Jr (1997) Stable behaviors associated with adults’ 10-year change in body mass index and likelihood of gain at the waist. Am J Public Health 87: 747754.
  • 30
    Brown, W. J., Williams, L., Ford, J. H., Ball, K., Dobson, AJ (2005) Identifying the energy gap: magnitude and determinants of 5-year weight gain in midage women. Obes Res. 13: 14311441.
  • 31
    Lahmann, P. H., Lissner, L., Gullberg, B., Berglund, G. (2000) Sociodemographic factors associated with long-term weight gain, current body fatness and central adiposity in Swedish women. Int J Obes Relat Metab Disord. 24: 685694.
  • 32
    Haapanen, N., Miilunpalo, S., Pasanen, M., Oja, P., Vuori, I. (1997) Association between leisure time physical activity and 10-year body mass change among working-aged men and women. Int J Obes Relat Metab Disord. 21: 288296.
  • 33
    Guthrie, J. R., Dennerstein, L., Dudley, EC (1999) Weight gain and the menopause: a 5-year prospective study. Climacteric 2: 205211.
  • 34
    U.S. Department of Agriculture/Department of Health and Human Services (Accessed February 2005). Dietary Guidelines for Americans.http:www.health.govdietaryguidelinesdga2005document.
  • 35
    Teixeira, P. J., Going, S. B., Houtkooper, L. B., et al (2002) Weight loss readiness in middle-aged women: psychosocial predictors of success for behavioral weight reduction. J Behav Med. 25: 499523.
  • 36
    Centers for Disease Control and Prevention/Department of Health and Human Services (Accessed March 2006). Physical Activity for Everyone: Getting Started http:www.cdc.govnccdphpdnpaphysicalstartingindex.htm
  • 37
    Wier, L. T., Ayers, G. W., Jackson, A. S., Rossum, A. C., Poston, W. S., Foreyt, JP (2001) Determining the amount of physical activity needed for long-term weight control. Int J Obes Relat Metab Disord. 25: 613621.
  • 38
    Kahn, E. B., Ramsey, L. T., Brownson, R. C., et al (2002) The effectiveness of interventions to increase physical activity. A systematic review. Am J Prev Med. 22 (4 suppl): 73107.
  • 39
    Evenson, K. R., Wilcox, S., Pettinger, M., et al (2002) Vigorous leisure activity through women's adult life: the Women's Health Initiative Observational Cohort Study. Am J Epidemiol. 15: 945953.
  • 40
    Williams, P. T., Wood, PD (2006) The effects of changing exercise levels on weight and age-related weight gain. Int J Obes (London) 30: 543551.
  • 41
    Tudor-Locke, C. E., Myers, AM (2001) Challenges and opportunities for measuring physical activity in sedentary adults. Sports Med. 31: 91100.
  • 42
    Weller, I., Carey, P. (1998) The impact of excluding non-leisure energy expenditure on the relation between physical activity and mortality in women. Epidemiology 9: 632635.
  • 43
    Prentice, A. M., Goldberg, G. R., Murgatroyd, P. R., Cole, TJ (1996) Physical activity and obesity: problems in correcting expenditure for body size. Int J Obes Relat Metab Disord. 20: 688691.
  • 44
    Wolf, A. M., Hunter, D. J., Colditz, G. A., et al (1994) Reproducibility and validity of a self-administered physical activity questionnaire. Int J Epidemiol. 23: 991999.