Identifying the Energy Gap: Magnitude and Determinants of 5-Year Weight Gain in Midage Women

Authors


  • 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.

School of Human Movement Studies, University of Queensland, St. Lucia, Queensland 4072, Australia. E-mail: wbrown@hms.uq.edu.au

Abstract

Objective: The aims of this study were to estimate average yearly weight gain in midage women and to identify the determinants of weight gain and gaining weight at double the average rate.

Research Methods and Procedures: The study sample comprised 8071 participants (45 to 55 years old) in the Australian Longitudinal Study on Women's Health who completed mailed surveys in 1996, 1998, and 2001.

Results: On average, the women gained almost 0.5 kg per year [average 2.42 kg (95% confidence interval, 2.29 to 2.54) over 5 years]. In multivariate analyses, variables associated with energy balance (physical activity, sitting time, and energy intake), as well as quitting smoking, menopause/hysterectomy, and baseline BMI category were significantly associated with weight gain, but other behavioral and demographic characteristics were not. After adjustment for all of the other biological and behavioral variables, the odds of gaining weight at about twice the average rate (>5 kg over 5 years) were highest for women who quit smoking (odds ratio = 2.94; 95% confidence interval, 2.17, 3.96). There were also independent relationships between the odds of gaining >5 kg and lower levels of habitual physical activity, more time spent sitting, energy intake (but only in women with BMI > 25 at baseline), menopause transition, and hysterectomy.

Discussion: The average weight gain equates with an energy imbalance of only about 10 kcal or 40 kJ per day, which suggests that small sustained changes in the modifiable behavioral variables could prevent further weight gain.

Introduction

Overweight and obesity increase the risk of many leading causes of morbidity and mortality, including coronary heart disease, stroke, type 2 diabetes, and some forms of cancer (1), as well as symptoms such as back ache and surgical procedures such as cholecystectomy (2, 3, 4, 5). The increasing prevalence of overweight and obesity, therefore, poses a major threat to the health of populations throughout the world. A recent World Health Organization report estimated that more than one billion adults worldwide are overweight and at least 300 million are clinically obese (1). In developed countries including the U.S., United Kingdom, and Australia, the prevalence of overweight and obesity has escalated in the last 2 decades; currently, between 50% and 65% of adults in these countries have a BMI above the healthy range (6, 7, 8).

The current obesity epidemic reflects a long-term upward shift in the distribution of BMI across whole populations (1, 6, 9), which is thought to be attributable to energy imbalance resulting from the adoption of more sedentary lifestyles, without equivalent reduction in energy intake (EI)1 (10, 11). The magnitude of this imbalance is thought to be relatively small, with energy accumulation in the order of 15 kcal/d accounting for weight gain in U.S. population studies (12). Although the determinants of this energy imbalance clearly relate to EI and expenditure, the relative contributions of the many potential demographic, biological, and behavioral determinants of weight gain are poorly understood.

Over the past 20 years or so, studies in several developed countries have suggested that women are at greater risk than men of experiencing major weight gain (13, 14, 15, 16). Women's heightened risk of weight gain is particularly pronounced at middle age (17), but the reasons for this remain poorly understood. Prospective cohort studies in North America, Europe, and Australia have suggested that weight gain at this life stage is due to aging (18, 19), lifestyle changes such as diet, physical activity (PA), or smoking (14, 20), or menopause (21).

The analyses presented in this paper aim to identify the major determinants of weight gain in midage women during a 5-year period at a life stage that has been identified as being a high-risk time for weight gain. The first objective was to quantify the amount of weight gained by midage women and to estimate the associated daily energy imbalance and the demographic and behavioral factors associated with it. The second objective was to identify factors associated with gaining weight at twice the average rate.

Research Methods and Procedures

The Australian Longitudinal Study on Women's Health (ALSWH)

The ALSWH is a prospective study of factors affecting the health and well-being of three cohorts of women, 18 to 23, 45 to 50, and 70 to 75 years old at the time of the baseline surveys in 1996. Women were selected randomly from the national Medicare health insurance database (which includes all permanent residents of Australia regardless of age, including immigrants and refugees) with intentional overrepresentation of women living in rural and remote areas. Further details of the recruitment methods and response rates have been described elsewhere (22). The study collects self-reported data using mailed surveys at 2- to 3-year intervals from ∼40, 000 women living in all states and territories of Australia. The surveys include questions about: health conditions, symptoms and diagnoses; use of health services; health-related quality of life, including measures of physical and mental health; social circumstances, including work and time use; demographic factors; and health behaviors. Informed consent was obtained from all participants in 1996, with ethical clearance by the University of Newcastle. This paper includes data from the midage cohort only.

Participants

Midage women who responded to the first (S1, 1996), second (S2, 1998), and third (S3, 2001) surveys were considered for inclusion in these analyses (n = 11, 196). Women who reported weight change greater than ±32 kg (n = 26), being pregnant at S3 (n = 69), being unable to walk 100 m (at S1, S2, or S3, n = 411), or having cancer (S3, n = 780) were excluded, leaving 9910 women eligible for this study. Of these, 1839 had missing data for one or more of the following variables: weight at S1 or S3, BMI at S1, education, alcohol consumption, hours worked, marital status, cigarette smoking, EI, area of residence, menopause transition, or hysterectomy. These women were also excluded, leaving data from 8071 women for inclusion. This analysis sample was similar to the entire cohort for all of the variables considered in this paper (data not shown).

Measures

Complete details of each survey are on the study website (www.newcastle.edu.aucentrewha). Based on our previous work with this cohort (4) and a review of the literature on determinants of weight gain, the following variables were selected for inclusion in this study.

Weight and BMI

BMI was calculated as reported weight (kilograms) divided by the square of reported height (meters squared) and categorized as: underweight (BMI < 20 kg/m2), healthy weight (BMI 20 to 25 kg/m2), overweight (BMI > 25 to 30 kg/m2), or obese (BMI > 30 kg/m2) according to the Australian National Health and Medical Research Council classification system (17). [The WHO categorization, which uses BMI < 18.5 as underweight, was not used because only 139 (1.7% of the analysis sample) had BMI < 18.5 at S1]. Self-reported weights at S1 and S3 were used to calculate 5-year weight change in kilograms (with negative values indicating weight loss and positive values weight gain). Weight change was then categorized into five groups: maintainer (−2.25 to +2.25 kg), weight loser (weight loss of >2.25 kg), low gainer (>2.25 to 5 kg), moderate gainer (>5 to 10 kg), or high gainer (>10 kg) (23).

PA

At S1, two questions about PA were used. They asked how many times in a normal week women engaged in vigorous exercise (e.g., aerobics, jogging) or less vigorous exercise (e.g., walking, swimming) lasting for 20 minutes or more (24). Responses were used to derive a PA score based on frequency of participation in vigorous [7.5 metabolic equivalent tasks (METs)] and less vigorous (4 METs) PA lasting at least 20 minutes. [PA score = ∑[frequency × 20 × 4 (less vigorous) + frequency × 20 × 7.5 (vigorous)]]. The items are known to have acceptable test-retest reliability (25).

At S2 and S3, PA was assessed using questions based on those developed for the evaluation of the national Active Australia campaign in 1997 and for national monitoring of PA in Australia (26). The questions ask about the frequency and total duration of walking (for recreation or transport) and of vigorous (e.g., aerobics, jogging) and moderate-intensity (e.g., swimming, golf) activity in the last week. These items have been shown to have acceptable reliability and validity for population measurement of PA (27, 28). A PA score was derived from reported duration of time spent in each form of PA during the last week [∑[(walking minutes × 3.5) + (moderate minutes × 4.0) + (vigorous minutes × 7.5)] MET/min] (29).

A composite score to represent habitual PA was derived from the PA scores for each survey (sum of S1, S2, and S3 PA scores) and categorized as: none (<180), very low (180 to <680), low (680 to <1600), moderate (1600 to <2960), and high (≥2960). The cut-off points were based on current Australian and U.S. PA guidelines such that a score of 1600 (for example, 400 at S1 and 600 at S2 and S3) approximately indicates compliance with current national guidelines of 150 minutes of moderate intensity PA per week (30), and a score of 680 approximates to an hour a week of moderate intensity PA.

EI

Information about daily EI was collected only at S3, using the Cancer Council of Victoria food frequency questionnaire (FFQ) (31). This FFQ is a validated instrument that assesses usual consumption of 74 food and 6 alcoholic beverage items over the previous 12 months (32). Respondents recorded their usual frequency of consumption of each item on a 10-point scale, ranging from never to three or more times per day. Nutrient intakes, including total EI, were computed using software developed by the Cancer Council of Victoria, based on the NUTTAB95 nutrient composition data (33). In light of likely underreporting of EI (32), approximate quintiles of EI (kilojoules per day) were created, with the following cut-off points: very low (≤4800), low (>4800 to 5800), moderate (>5800 to 6800), high (>6800 to 8300), and very high (>8300).

Sitting Time

An estimate of average daily sitting time was calculated from two questions about sitting during weekdays and weekend days at S3. These questions, which asked about time spent sitting while doing things like visiting friends, driving, reading, watching television, or working at a desk or computer, were not included in earlier surveys. The questions have been shown to have acceptable reliability (34). Average sitting time per day was categorized into approximate quintiles: very low (<3 hours), low (3 to <4.5 hours), moderate (4.5 to <6 hours), high (6 to <8 hours), or very high (≥8 hours).

Smoking and Alcohol

Data on tobacco smoking at S1 and S3 were used to categorize smoking transition status as: never smoked, ex-smoker (quit before S1), quitter (smoker at S1 and not at S3), or smoker (at S3, including new adopters and restarters since S1). Only S1 alcohol consumption data were used because the frequency and quantity of consumption reported at S1 and S2 were very similar, and comparable questions were not asked at S3. The categories used were: no risk (never or rarely drink alcohol), low risk (up to 14 drinks per week), or intermediate/high risk (>14 drinks per week) (35).

Menopause Transition and Hysterectomy

At both S1 and S3, women were categorized as having had a hysterectomy and/or oophorectomy or as being pre-, peri-, or postmenopausal (36). These categories were used to create six menopause transition categories: prior hysterectomy (before S1), recent hysterectomy (between S1 and S3), premenopausal (at both S1 and S3), postmenopausal (at both S1 and S3), pre-/peri- to postmenopausal (pre- or perimenopausal at S1 and postmenopausal at S3), or perimenopausal (pre- or perimenopausal at S1, perimenopausal at S3, or any other combination not already defined).

Sociodemographic Variables

Education level, based on highest qualification achieved, was assessed at S1 and categorized as: low (school certificate or less), intermediate (higher school certificate), technical (having a trade certificate or diploma), or university (completed a university degree). Australian Standard Coding of Occupations (37) was used to define women's occupation at S3, with the categories: no paid job, blue collar (e.g., in production, transport, cleaning, etc.), skilled (e.g., in a trade or advanced clerical work), or professional (e.g., manager, teacher, etc.). Hours worked per week at S3 were categorized as: none (not in paid work), 1 to 34 hours, or 35 hours or more. Marital status at S3 was categorized as: sole (i.e., separated, divorced, widowed, never married) or partnered (i.e., married or living with a partner). Area of residence at S3 was classified as urban, large rural center, other rural area, or remote, based on an index of distance to the nearest urban center (38).

Statistical Analyses

All statistical analyses were performed using SAS version 8 (39). To maximize the number of women whose data could be included in these analyses, missing categories were created for the PA, sitting time, and occupation variables.

Exploratory analyses of variables associated with categories of weight change were conducted using χ2 tests. Because there was a strong interaction between BMI at S1 and EI at S3, a combined variable, BMI × EI, was created and used in subsequent analyses in place of the original variables. Step-wise multiple regression was used to estimate average weight change associated with each variable, after adjustment for all other variables. The observed margins least squares means option in the SAS generalized linear model procedure was used to estimate means and 95% confidence intervals (CIs) for weight change between S1 and S3, for the entire sample and for each category of the independent variables, after adjustment for all other variables in the model. Similarly, step-wise logistic regression was used to identify factors that increased the odds for weight gain among the 2046 women who gained >5 kg, compared with the 3077 women who maintained their weight within ±2.25 kg.

Results

Mean (SD) weight and BMI at S3 for the 8071 women included in these analyses were 70.9 (14.45) kg and 26.6 (5.20) kg/m2 and for the entire cohort of 11, 196 were 71.2 (14.84) kg and 26.7 (5.33) kg/m2, respectively. Sociodemographic, biological, and behavioral characteristics of women who maintained their weight (within ±2.25 kg) and women who lost or gained >2.25 kg are shown in Table 1. Overall, about one in seven women reported weight loss of >2.25 kg, and almost one-half reported weight gain of >2.25 kg. Almost 40% were defined as weight maintainers; of these, 25% (N = 772) reported no weight change during the 5-year period. As expected, weight gain was associated with lower levels of habitual PA, more time spent sitting, and higher BMI × EI. Greater weight gain was also associated with quitting smoking, higher alcohol intake, and having had a hysterectomy. In contrast, with the exception of marital status, the demographic variables were not strongly statistically significantly associated with weight change. In general, weight losers and high gainers had similar characteristics—women in both groups having had high BMI at S1.

Table 1. . Percentage distributions of sociodemographic, biological, and behavioral factors across weight change categories
 pN = 8071Maintainer (±2.25 kg) N = 3077 (38.1%)Weight loser (−>2.25 kg) N = 1204 (14.9%)Low gainer (+2.25 to <5 kg) N = 1744 (21.6%)Mod gainer (+5 to <10 kg) N = 1451 (18.0%)High gainer (+>10 kg) N = 595 (7.4%)Overall distribution (%)
Habitual physical activity (S1, S2, and S3)<0.0001       
 High 197027.423.123.321.422.224.4
 Moderate 191724.824.223.223.719.023.8
 Low 189821.722.825.324.825.923.5
 Very low 98311.111.712.614.012.912.2
 None 3733.76.04.35.06.64.6
 Missing 93011.212.211.211.113.511.5
Average sitting time each day (S3)<0.0001       
 Very low 139618.819.416.614.813.517.3
 Low 160220.918.720.619.615.519.9
 Moderate 186522.924.022.824.021.223.1
 High 134616.414.418.015.920.516.7
 Very high 116113.413.214.215.919.014.4
 Missing 7017.710.37.99.810.48.7
Smoking transition (S1→S3)<0.0001       
 Never smoked 441756.549.757.655.146.654.7
 Ex-smoker 230528.829.027.827.232.128.6
 Quitter 2862.52.52.95.29.13.5
 Smoker 106312.318.911.712.412.313.2
Alcohol risk (S1)<0.0001       
 No risk 347040.147.641.645.147.443.0
 Low risk 418355.346.752.649.747.251.8
 Intermediate/high risk 4184.65.75.85.25.45.2
Menopause transition (S1→S3) and hysterectomy<0.0001       
 Premenopausal (no change) 95213.311.912.69.18.111.8
 Perimenopausal 230429.425.829.129.526.228.6
 Postmenopausal (no change) 4505.76.85.34.55.75.6
 Pre-/peri- to postmenopausal 214426.425.127.327.426.426.6
 Recent hysterectomy (S1 to S3) 5156.26.86.26.46.76.4
 Prior hysterectomy (before S1) 170619.023.719.523.226.921.1
BMI (S1) × energy intake (S3 FFQ)<0.0001       
 Healthy weight (<25 kg/m2)        
 Very low 93713.18.812.910.39.211.6
 Low 88012.67.811.310.57.910.9
 Moderate 84611.95.512.211.07.110.5
 High 93614.26.313.010.67.411.6
 Very high 77010.95.910.79.07.99.5
Overweight (>25 to 30 kg/m2)        
 Very low 4725.57.74.95.47.95.9
 Low 4424.97.44.85.46.65.5
 Moderate 4424.76.65.65.76.45.5
 High 4835.15.76.08.06.46.0
 Very high 4714.56.16.36.88.75.8
Obese (>30 kg/m2)        
 Very low 2191.96.61.72.42.92.7
 Low 2522.36.51.63.34.73.1
 Moderate 2672.66.02.82.84.23.3
 High 3332.77.43.84.64.74.1
 Very high 3213.15.82.74.28.14.0
Education (S1)0.1673       
 Low 379745.747.647.749.146.147.0
 Intermediate 137316.817.416.717.018.217.0
 Trade/certificate/diploma 162020.718.819.020.122.520.1
 University degree 128116.816.216.613.913.315.9
Occupation (S3)0.0636       
 No paid job 182021.821.822.523.026.922.6
 Blue collar 96811.314.012.612.68.212.0
 Skilled 195824.224.124.724.523.524.3
 Professional 275335.832.633.532.733.634.1
 Missing 5727.07.56.77.27.77.1
Hours worked (S3)0.0219       
 None 170319.821.421.321.425.921.1
 1 to 34 h/wk 289937.434.237.034.931.135.9
 35+ h/wk 346942.844.441.743.643.043.0
Marital status (S3)0.0008       
 Sole 186521.626.721.723.626.623.1
 Partnered 620678.473.378.376.473.576.9
Area of residence (S3)0.7726       
 Urban 308638.436.639.537.638.238.2
 Large rural centre 110613.613.513.015.113.613.7
 Other rural area 346643.144.042.941.942.542.9
 Remote 4134.95.84.65.45.75.1

Average weight gain for all women in the analysis sample was 2.42 kg (95% CI, 2.29 to 2.54), or just under 0.5 kg per year, over the 5 years between S1 and S3. For the weight gain analysis, stepwise selection of variables significantly associated with weight change in Table 1 eliminated all of the sociodemographic factors. Adjusted mean weight changes relating to the remaining behavioral and biological variables are shown in Table 2. Women who gave up smoking gained 5.06 kg (95% CI, 4.39 to 5.74), whereas women who smoked gained only 1.48 kg, on average (95% CI, 1.13 to 1.83). Even after adjustment for the other variables, average weight gain was significantly lower than the sample average among women who remained premenopausal, among those who reported sitting <3 h/d, and among all women with BMI > 30 except those who reported very high EI. In contrast, weight gain was significantly higher than average among women who reported sitting time of >8 h/d and among women with BMI of 25 to 30 who reported very high EI (see Table 2).

Table 2. . Mean 5-year weight change (kilograms)* for women in each behavioral, biological, and sociodemographic category (N = 8071), adjusted for all other variables in the table
 NMean (95% CI)*p
  • *

    Observed margins least square means, adjusted for all other variables shown in the table.

  • p value for differences between categories.

Overall weight change80712.42 (2.29, 2.54) 
Habitual physical activity (S1, S2, and S3)  0.0022
 High19702.21 (1.95, 2.47) 
 Moderate19172.08 (1.82, 2.34) 
 Low18982.72 (2.46, 2.99) 
 Very low9832.74 (2.37, 3.10) 
 None3732.81 (2.22, 3.41) 
 Missing9302.42 (2.04, 2.79) 
Average sitting time each day (S3)  <0.0001
 Very low13961.80 (1.50, 2.11) 
 Low16022.25 (1.97, 2.54) 
 Moderate18652.27 (2.01, 2.54) 
 High13462.78 (2.47, 3.09) 
 Very high11613.04 (2.71, 3.38) 
 Missing7012.66 (2.23, 3.09) 
Smoking transition (S1→S3)  <0.0001
 Never smoked44172.43 (2.26, 2.60) 
 Ex-smoker23052.49 (2.25, 2.73) 
 Quitter2865.06 (4.39, 5.74) 
 Smoker10631.48 (1.13, 1.83) 
Menopause transition (S1→S3)  <0.0001
 Premenopausal (no change)9521.71 (1.34, 2.08) 
 Perimenopausal23042.57 (2.33, 2.81) 
 Postmenopausal (no change)4501.77 (1.23, 2.31) 
 Pre-/peri- to postmenopausal21442.42 (2.18, 2.67) 
 Recent hysterectomy (S1 to S3)5152.54 (2.03, 3.04) 
 Prior hysterectomy (before S3)17062.73 (2.45, 3.00) 
BMI group (S1) × energy intake (S3 FFQ)  <0.0001
Healthy weight (<25 kg/m2)   
 Very low9372.55 (2.18, 2.93) 
 Low EI8802.57 (2.19, 2.96) 
 Moderate EI8462.84 (2.45, 3.24) 
 High EI9362.72 (2.34, 3.09) 
 Very high EI7702.84 (2.43, 3.26) 
Overweight (>25 to 30 kg/m2)   
 Very low4722.28 (1.76, 2.81) 
 Low4422.08 (1.53, 2.62) 
 Moderate4422.54 (2.00, 3.08) 
 High4832.91 (2.39, 3.43) 
 Very high4713.24 (2.72, 3.77) 
Obese (>30 kg/m2)   
 Very low219−0.29 (−1.06, 0.49) 
 Low2520.94 (0.22, 1.66) 
 Moderate2670.89 (0.19, 1.59) 
 High3331.56 (0.93, 2.19) 
 Very high3212.26 (1.62, 2.89) 

When women who gained >5 kg (N = 2046) were compared with women who maintained their weight within 2.25 kg (N = 3007), significant relationships were found between weight gain > 5 kg and the behavioral and menopause variables but not with the remaining sociodemographic variables (see Figure 1). Quitting smoking was associated with an almost 3-fold increase in odds of gaining >5 kg, compared with women who had never smoked. For PA, women who reported low, very low, or no PA (i.e., <150 minutes per week) were significantly more likely to gain >5 kg than women in the high PA category (equivalent to >300 minutes of PA a week). Similarly, women who reported >4.5 hours sitting a day were more likely to gain >5 kg than those reporting <3 hours sitting time. Women who were perimenopausal and those who became postmenopausal during the study period, as well as those who had had a recent or prior hysterectomy, were also more likely to gain >5 kg than women who remained premenopausal. EI was not associated with weight gain > 5 kg in women who were in the healthy BMI category at S1. However, the odds of gaining >5 kg were higher for almost all of the BMI × EI categories for women who were overweight or obese at S1.

Figure 1.

Odds ratios and 95% confidence intervals (95% CI) for gaining more than 5 kg (N = 2046), compared with maintaining weight (±2.25 kg, N = 3077). Each variable is adjusted for all the other variables in the Figure. Asterisk indicates reference category.

Discussion

The findings confirm those of previous studies that have shown that weight gain is largely a reflection of energy imbalance. However, in addition to the usual culprits of PA and EI, we found that, even after adjustment for the other energy balance variables, sitting time, quitting smoking, menopause, and having a hysterectomy were also independently associated with weight gain over a 5-year period in this cohort of midage Australian women.

Assuming that each kilogram of weight gained represents 7700 kcal, the amount of weight gained over a 5-year period (∼0.5 kg per year, on average) reflects a cumulative positive energy accumulation of only ∼10 kcal (40 kJ) a day. This estimate does not account for inefficiencies in storage of excess energy, increases in energy expenditure caused by weight gain, or the fact that the measures used in this study are relatively imprecise, compared with direct measures of EI and expenditure (40). Although direct measures of energy cannot be used in large population studies, the estimate of ∼10 kcal/d is consistent with that suggested by Hill and colleagues (10, 11) These authors proposed that energy imbalance of this magnitude is responsible for the current obesity epidemic (10, 11). By similar calculations, the women in this study who gained 5 to 10 kg would have had an imbalance of 20 to 40 kcal/d over this 5-year period. Allowing for the potential inefficiency in storage of excess energy, changes in EI and expenditure amounting to 50 to 100 kcal a day would not be unrealistic targets for the prevention of further weight gain in this group (11).

The amount of weight gained by women who quit smoking during the 5-year period of this study is disturbing because this weight gain is seen by women as a barrier to quitting smoking (41). A small U.S. study has found that smoking cessation for just 4 weeks is associated with changes in adipose cell metabolism that contribute to weight gain (42). However, other studies have shown that weight gain associated with quitting is transient (43, 44), and this is borne out by the data for the ex-smokers in this study, who did not gain weight at a higher than average rate.

The independent effect of hysterectomy on weight gain is also interesting because women who had a hysterectomy during the follow-up period, as well as those who had had this surgery before the baseline survey, were more likely to gain weight at the higher rate (>5 kg in 5 years). Hysterectomy has been shown in animal studies to induce significant decreases in energy requirements so that weight gain can be avoided only by substantial changes in energy balance (45).

The data presented here suggest that EI was predictive of weight gain primarily in those women who were overweight or obese at S1. The results, therefore, highlight the need to address the modifiable energy balance variables (EI, PA, and sitting time), particularly in women who are already overweight or obese, in an attempt to prevent further weight gain. Moreover, public health and clinical advice for women who are going through menopause, facing hysterectomy, or quitting smoking should routinely include the fact that these events are associated with increased risk of weight gain and that it may be necessary to preemptively avoid this by making simultaneous small reductions in EI and small increases in PA, and by decreasing sitting time. Women whose occupations require them to sit for long hours (like the authors of this paper) might also benefit from more focused strategies to prevent weight gain, such as promotion of more active transport to and from work and mandatory activity breaks while at work.

The main strength of this study is that it includes a large population-based sample of women from all walks of Australian life. The major limitation is that all of the data are self-reported, and it is well known that women tend to underestimate their weight and overestimate their height (46). However, if the women underestimated their weight consistently, weight change data might be reasonably accurate, although if awareness of increasing weight led them to underreport at S3, then the average weight change calculated in this paper may be an underestimate. Another limitation is that the measures of both PA and EI used here do not provide absolute estimates of total daily activity or of total EI. The PA measure does not capture incidental or unstructured PA (28), which can be substantial in some occupations in which women are on their feet all day (e.g., teaching, nursing), and it is well known that FFQs underestimate EI (47).

Nonetheless, the data on relative weight gain across different PA and EI categories provide useful insight into the potential reduction in weight gain that might be expected if there were small population shifts in the modifiable factors (PA, sitting time, and EI) that underlie weight gain at this life stage. For example, walking for 30 minutes a day would use 120 MET/min or 140 kcal/d for the average (71 kg) woman in this study. Even smaller increases of 15 minutes a day, or ∼1500 steps at a brisk pace, would, therefore, make a significant contribution to reducing further weight gain and the serious and costly chronic health problems that are associated with it (48).

Acknowledgement

The ALSWH, which was conceived and developed by groups of interdisciplinary researchers at the Universities of Newcastle and Queensland, is funded by the Australian Department of Health and Ageing. We would like to thank Anne Young for her advice on statistical issues, and all of the participants for their valuable contribution to this project. K.B. is supported by a National Health and Medical Research Council/National Heart Foundation of Australia Career Development Award.

Footnotes

  • 1

    Nonstandard abbreviations: EI, energy intake; PA, physical activity; ALSWH, Australian Longitudinal Study on Women's Health; MET, metabolic equivalent task; S1, first survey; S2, second survey; S3, third survey; FFQ, food frequency questionnaire; CI, confidence interval.

Ancillary