Skip to main content Accessibility help
×
Home
Hostname: page-component-568f69f84b-jg9p7 Total loading time: 0.516 Render date: 2021-09-19T01:24:23.241Z Has data issue: true Feature Flags: { "shouldUseShareProductTool": true, "shouldUseHypothesis": true, "isUnsiloEnabled": true, "metricsAbstractViews": false, "figures": true, "newCiteModal": false, "newCitedByModal": true, "newEcommerce": true, "newUsageEvents": true }

Predictors of increased body weight and waist circumference for middle-aged adults

Published online by Cambridge University Press:  01 May 2013

Robert J MacInnis*
Affiliation:
Cancer Epidemiology Centre, Cancer Council Victoria, 1 Rathdowne Street, Carlton, Victoria 3053, Australia Centre for Molecular, Environmental, Genetic and Analytic Epidemiology, School of Population Health, The University of Melbourne, Carlton, Victoria, Australia
Allison M Hodge
Affiliation:
Cancer Epidemiology Centre, Cancer Council Victoria, 1 Rathdowne Street, Carlton, Victoria 3053, Australia
Helen G Dixon
Affiliation:
Centre for Behavioural Research in Cancer, Cancer Council Victoria, Carlton, Victoria, Australia
Anna Peeters
Affiliation:
Obesity & Population Health Unit, Baker IDI Heart and Diabetes Institute, Melbourne, Victoria, Australia School of Population Health and Preventive Medicine, Monash University, Melbourne, Victoria, Australia
Lucinda EA Johnson
Affiliation:
Cancer Epidemiology Centre, Cancer Council Victoria, 1 Rathdowne Street, Carlton, Victoria 3053, Australia
Dallas R English
Affiliation:
Cancer Epidemiology Centre, Cancer Council Victoria, 1 Rathdowne Street, Carlton, Victoria 3053, Australia Centre for Molecular, Environmental, Genetic and Analytic Epidemiology, School of Population Health, The University of Melbourne, Carlton, Victoria, Australia
Graham G Giles
Affiliation:
Cancer Epidemiology Centre, Cancer Council Victoria, 1 Rathdowne Street, Carlton, Victoria 3053, Australia Centre for Molecular, Environmental, Genetic and Analytic Epidemiology, School of Population Health, The University of Melbourne, Carlton, Victoria, Australia
*Corresponding
*Corresponding author: Email Robert.MacInnis@cancervic.org.au
Rights & Permissions[Opens in a new window]

Abstract

Objective

To identify predictors of increased adiposity for different measures of adiposity.

Design

Prospective cohort study, the Melbourne Collaborative Cohort Study (MCCS), with data at baseline (1990–1994) and wave 2 (2003–2007).

Setting

Participants recruited from the community.

Subjects

Australian-born participants (n 5879) aged 40 to 69 years who were not current smokers and who were free from common chronic diseases at recruitment. At baseline and at wave 2, weight and waist circumference were measured; while demographic and lifestyle variables were obtained at baseline via structured interviews.

Results

Participants who reported any recreational physical activity at baseline had lower weight and smaller waist circumference at wave 2 than those who did not, particularly for younger participants and for vigorous physical activity. Walking for leisure was not associated, and greater physical activity at work was associated, with greater adiposity measures at wave 2. A diet low in carbohydrates and fibre, but high in fat and protein, predicted greater weight and waist circumference at wave 2. Participants were less likely to have elevated weight or waist circumference at wave 2 if they consumed low to moderate amounts of alcohol.

Conclusions

Our findings indicate that promoting vigorous physical activity, encouraging a diet high in carbohydrate and fibre but low in fat and protein, and limiting alcohol intake could be promising approaches for preventing obesity in middle-aged adults. Similar interventions should successfully address the management of both weight and waist circumference, as they were predicted by similar factors.

Type
Epidemiology
Copyright
Copyright © The Authors 2013 

Understanding the individual-level predictors of weight and waist circumference gain is critical, both to identify individuals most at risk and to allow targeting of the behavioural risk factors likely to contribute the most to overweight and obesity. While consensus is forming around the role of specific dietary elements and markers of physical activity and inactivity associated with increased weight gain( Reference Mozaffarian, Hao and Rimm 1 3 ), few studies have analysed predictors of increasing waist circumference( Reference Du, van der A and Ginder 4 Reference Walls, Magliano and McNeil 7 ) and the majority of these have focused solely on dietary factors.

Waist circumference is a better indicator of excess visceral adiposity( Reference Rankinen, Kim and Perusse 8 ), which tends to be more closely associated with metabolic abnormalities and some cancers than overall adiposity as assessed by BMI( 3 ). Further, there is evidence that waist circumference is increasing at a faster rate than BMI( Reference Walls, Magliano and McNeil 7 , Reference Elobeid, Padilla and Brock 9 ), suggesting that there may be differential drivers of BMI and waist circumference. A recent analysis of cohorts in the European Prospective Investigation into Cancer and Nutrition suggested that specific dietary elements may predict changes in waist circumference independent of changes in BMI( Reference Romaguera, Angquist and Du 5 ). However, no studies have simultaneously analysed the association between a wide range of behavioural factors and changes in weight and waist circumference.

The predictors of weight gain may differ from the predictors of weight loss( Reference Fogelholm, Anderssen and Gunnarsdottir 10 ). The health benefits for overweight individuals of losing weight compared with maintaining a stable weight are still unknown( Reference Simonsen, Hundrup and Gronbaek 11 ). Exploring predictors of weight loss, in particular for elderly populations, is problematic as weight loss may be unintentional due to an undiagnosed disease( Reference Manson, Bassuk and Hu 12 ).

We evaluated potential predictors of stable and increased weight and waist circumference using a prospective cohort study in which all participants were directly measured at baseline recruitment and at a later face-to face examination (mean 11·7 years after baseline), which we refer to as wave 2. We focused on potentially modifiable predictors such as physical activity (both leisure time and work related), alcohol consumption and nutrient intake for participants who were not current smokers and were free from common chronic diseases at baseline.

Participants and methods

Participants

The Melbourne Collaborative Cohort Study (MCCS) is a prospective cohort study comprising 17 045 men and 24 469 women, aged between 27 and 75 years at baseline (99·3 % aged 40 to 69 years). Participants were recruited from the Melbourne metropolitan area between 1990 and 1994 using the electoral rolls for direct invitations, advertisements and community announcements. Southern European migrants were oversampled at the time of recruitment to help increase the variability of dietary and non-dietary risk factors. Further details of the study have been published elsewhere( Reference Giles and English 13 ). The study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving participants were approved by the Cancer Council Victoria's Human Research Ethics Committee. Written informed consent was obtained from all participants.

A randomly selected total of 11 994 participants was targeted for attendance at wave 2 between 2003 and 2007. Participants were excluded if, at baseline, they were aged less than 40 years or over 69 years (n 7), had energy intakes in the extreme centiles of the sex-specific distributions (n 244), were missing values of weight or waist circumference (n 9), had a pre-existing disease (diabetes, cancer, angina, stroke or heart attack; n 1500) or were current smokers (to avoid possible weight changes due to quitting smoking; n 1168). After these exclusions, 9066 participants were eligible.

Ascertainment of cancer status and vital status

Cancer cases were identified from notifications to the Victorian Cancer Registry of diagnoses of invasive cancer (International Classification of Diseases 9th revision rubric 140–208 or 10th revision rubric C00–C99). Vital status and place of residence were obtained from electoral rolls, electronic phone books and death records until 2010.

Direct physical measurements

All participants attended a study centre at baseline, where physical measurements and blood sampling were performed. Height, weight and waist circumference were measured according to standard procedures( Reference Lohman, Roche and Martorell 14 ). Weight and waist circumference were measured similarly at wave 2. Heart rate was measured at baseline using Dinamap automatic monitors (see MacInnis et al.( Reference MacInnis, English and Gertig 15 ) for further details).

Questionnaire data and constructed variables

Participants’ residential postcodes at baseline were used to assign them to a quintile of socio-economic status (SES) according to the Index of Relative Socio-economic Advantage and Disadvantage that was obtained from the Australian Bureau of Statistics census-based Socio-Economic Indexes for Areas( 16 ).

A structured interview schedule was used to obtain information on country of birth, smoking (‘Have you ever smoked at least seven cigarettes a week for at least a year?’ and ‘Do you now smoke at least seven cigarettes a week?’), highest level of education and self-reported history of diabetes, angina, stroke and heart attack.

Questions relating to frequency of walking (for recreation or exercise), vigorous exercise (exercise ‘making you sweat or feel out of breath, and includes such activities as swimming, tennis, netball, athletics and running’) and less-vigorous exercise (exercise ‘which did not make you sweat or feel out of breath and includes such activities as bike riding, dancing, etc.’) over the last 6 months were asked at baseline. These questions were obtained from the Risk Factor Prevalence Study conducted by the National Heart Foundation and Australian Institute of Health( 17 ). The reported frequency for each question was coded as follows: 0 (none), 1·5 (one or two times per week) or 4 (three or more times per week). Walking and less-vigorous exercise frequencies were added together along with two times the frequency of vigorous exercise to generate a physical activity score for each person. This score was then grouped based on approximate quartiles (0, >0–3, >3–4, >4). Two further questions were also asked about how much time participants spent in moderate to heavy physical exertion at work and at home.

Dietary data were collected at baseline using a self-administered 121-item FFQ specifically developed for the study( Reference Ireland, Jolley and Giles 18 ). Nutrient intakes were computed from the FFQ based on the NUTTAB95 nutrient composition data( Reference Hodge, Patterson and Brown 19 ). Intakes of protein, fat, total carbohydrates and fibre were analysed as their proportion of total energy intake, with each gram assumed to contain 17, 37, 16 and 1 kJ of total energy, respectively( Reference Holland, Welch and Unwin 20 ). Alcohol intake data were collected at baseline using beverage-specific questions of frequency and quantity.

Statistical analysis

Multivariable linear regression was used for cross-sectional analyses and to predict weight and waist circumference at wave 2. Analyses for each predictor were adjusted for possible confounders at baseline identified a priori, including baseline measures of weight (or waist circumference), sex, age, physical activity (pseudo-continuous), alcohol intake (none, 0·1–19·9, 20·0–39·9, ≥40·0 g/d), educational attainment (primary school/some high or technical school, completed high or technical school, degree/diploma), smoking status (never, past), marital status (married/de facto, single/divorced/widowed/separated), SES (pseudo-continuous, based on approximate quintiles) and total dietary energy intake (kJ/d). Additional sub-analyses were also performed, including: tests for interactions between age, sex and predictor variables; further adjustment for duration since baseline; analyses restricted to participants without disease onset before wave 2 (cancer, heart attack, stroke, angina and diabetes); and analyses excluding those who reported having hypertension at baseline.

Statistical analyses were performed using the statistical software package Stata version 12·0.

Results

After baseline, 656 of the 9066 eligible people died before wave 2 was completed and a further 1894 declined to attend for physical examination at wave 2. Twenty-one of the people who attended wave 2 were excluded as they had no weight or waist circumference measurements collected at wave 2, leaving 6495 (72 %). A further 616 participants who lost 5 kg or more in weight between baseline and wave 2 were excluded because the predictors of loss may differ from those of maintenance or gain. This left 5879 participants in the final analysis.

On average, non-participants at wave 2 (which included current smokers at baseline) were more likely to be older (56 v. 54 years old), male (44 % v. 36 %), born in Southern Europe (31 % v. 18 %), less educated (36 % v. 51 % completed high school), not married (32 % v. 27 %), heavier (males 81 v. 80 kg, females 69 v. 67 kg) and to have larger waist circumference (males 94 v. 92 cm, females 81 v. 78 cm) than participants at baseline.

At wave 2, on average 11·7 years since baseline, participants were on average 3·3 kg heavier and had 8·0 cm larger waist circumferences. Additional descriptive characteristics of the participants are shown in Table 1.

Table 1 Characteristics of participants stratified by sex: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

*Participants in the vigorous group may have also reported doing less vigorous exercise.

Table 2 shows the cross-sectional associations between both weight and waist circumference at baseline and the various exposures. Measures of weight and waist circumference were greater for men and older participants at baseline, while many of the modifiable baseline exposures had positive (protein and fat intake) or negative (exercise, walking, low alcohol, carbohydrate and fibre intakes) associations with weight and waist circumference at baseline.

Table 2 Associations of potential predictors of weight and waist circumference at baseline*: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

β, coefficient; Ref., referent category.

*Adjusted for sex, height, age, highest level of education (pseudo-continuous), physical activity score (pseudo-continuous), marital status (married/de facto v. single/divorced/widowed/separated), Index of Relative Socio-economic Disadvantage (pseudo-continuous), smoking status and dietary energy intake at baseline attendance.

†Participants in the vigorous exercise group may have also reported doing less vigorous exercise.

‡Participants in the alcohol-specific groups may have also reported drinking other types of alcoholic beverages.

Table 3 shows the associations between both weight and waist circumference at wave 2 and the various baseline exposures. Measures of weight and waist circumference were higher for women and younger participants at wave 2. Also, the longer the duration since baseline, the greater the weight and waist circumference were at wave 2. Participants born in Southern Europe had a larger waist circumference at wave 2.

Table 3 Associations of potential predictors of weight and waist circumference at wave 2: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

β, coefficient; N/A, not applicable; Ref., referent category.

*Adjusted for weight, sex, height, age, highest level of education (pseudo-continuous), physical activity score (pseudo-continuous), marital status (married/de facto v. single/divorced/widowed/separated), Index of Relative Socio-economic Disadvantage (pseudo-continuous), smoking status and dietary energy intake at baseline attendance.

†Adjusted for waist circumference, sex, height, age, highest level of education (pseudo-continuous), physical activity score (pseudo-continuous), marital status (married/de facto v. single/divorced/widowed/separated), Index of Relative Socio-economic Disadvantage (pseudo-continuous), smoking status and dietary energy intake at baseline attendance.

‡Participants in the vigorous exercise group may have also reported doing less vigorous exercise.

§Participants in the alcohol-specific groups may have also reported drinking other types of alcoholic beverages.

Participants who had completed high school or above had a lower waist circumference but not weight at wave 2 compared with those who had not. Weight and waist circumference at wave 2 were greater for people living in more disadvantaged areas. Married and de facto persons had on average lower weight at wave 2 than single, divorced or widowed people. Notably, greater waist circumferences at wave 2 were observed for divorced compared with married participants.

Positive associations were observed between waist circumference at wave 2 and baseline cardiovascular measures (systolic pressure, diastolic pressure, pulse rate) in contrast to the null associations with weight at wave 2.

Those who reported participating in recreational exercise at baseline had lower weight and smaller waist circumference at baseline and at wave 2 than those who did not. The associations strengthened with exercise intensity and frequency. The association with exercise was mainly restricted to younger participants (see Tables 4 and 5). On the other hand, moderate or heavy activity at work was positively associated with waist circumference at wave 2, even after adjustment for recreational exercise. Despite the fact that those participants who walked three times or more per week had lower baseline weight and smaller waist circumference, there was no discernible association between leisure-time walking and future weight or waist circumference.

Table 4 Selected associations of potential predictors of weight at wave 2 by age at attendance*: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

β, coefficient; Ref., referent category.

*Adjusted for weight, sex, height, age, highest level of education (pseudo-continuous), physical activity score (pseudo-continuous), marital status (married/de facto v. single/divorced/widowed/separated), Index of Relative Socio-economic Disadvantage (pseudo-continuous), smoking status and dietary energy intake at baseline attendance.

†Participants in the vigorous exercise group may have also reported doing less vigorous exercise.

Table 5 Selected associations of potential predictors of waist circumference at wave 2 by age at attendance*: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

β, coefficient; Ref., referent category.

*Adjusted for waist circumference, sex, height, age, highest level of education (pseudo-continuous), physical activity score (pseudo-continuous), marital status (married/de facto v. single/divorced/widowed/separated), Index of Relative Socio-economic Disadvantage (pseudo-continuous), smoking status and dietary energy intake at baseline attendance.

†Participants in the vigorous exercise group may have also reported doing less vigorous exercise.

There was no notable association observed with smoking status. Participants were less likely to have an elevated weight or waist circumference at wave 2 if they consumed low to moderate amounts of alcohol.

Baseline energy intake showed no association with future weight or waist circumference, but a higher proportional consumption of protein and fat at baseline was positively associated with both weight and waist circumference at wave 2, while carbohydrate and fibre consumption were negatively associated with waist circumference but not weight at wave 2. As with exercise, the associations with these nutrients were strongest for younger participants (see Tables 4 and 5).

Analyses restricted to participants without disease onset before wave 2 (cancer, heart attack, stroke, angina and diabetes) gave similar results to the main analyses (data not shown) and associations remained virtually unchanged after further adjustment for duration of follow-up time since baseline (data not shown). Analyses excluding those who reported having hypertension at baseline (1031 people) gave similar results to the main analyses (data not shown), except that systolic (β coefficient = −0·11, se = 0·05, P = 0·04) and diastolic (β coefficient = −0·16, se = 0·08, P = 0·04) blood pressure were inversely associated with weight at wave 2.

Analyses stratified by sex indicated that male former smokers were more likely to have a greater waist circumference at wave 2 (males β coefficient = 0·87, se = 0·28, P = 0·002; females β coefficient = 0·13, se = 0·29, P = 0·67; P for interaction = 0·02). Compared with abstainers, female heavy drinkers of alcohol (≥40·0 g/d) were more likely to have a smaller waist circumference at wave 2 (males β coefficient = 0·62, se = 0·51, P = 0·23; females β coefficient = −2·00, se = 0·86, P = 0·02; P for interaction = 0·003). There were no other notable differences in results between males and females (data not shown).

Discussion

We have shown that some of the factors that predict greater weight and larger waist circumference in middle-aged and older adults over a 12-year follow-up period, such as physical activity and diet, are modifiable targets for individual behaviour change. Similar interventions could successfully address the management of weight and waist circumference, as both measures are predicted by similar factors.

One strength of our study is that weight and waist circumference were directly measured on both occasions, thus minimizing potential errors that can occur with self-reported measures. Restriction of the analyses to people who were not current smokers and did not report having been diagnosed with cancer, diabetes, angina, stroke or heart disease at baseline reduced the likelihood that the results are explained by confounding due to smoking or to reverse causation due to prevalent illness( Reference Manson, Bassuk and Hu 12 ). It is possible that some participants may not have known they had diabetes, but it is unlikely that they would have changed their reported lifestyle habits at baseline because of this. On the other hand, participants who were aware they had diabetes at baseline may have altered their lifestyle in response to the diagnosis. While some studies have analysed predictors of weight change, we analysed absolute weight at wave 2 conditional on weight at baseline, as there is considerable doubt that causal knowledge may be gained by substituting a derived variable (i.e. change in weight) for a measured variable( Reference Shahar and Shahar 21 ). In addition, our approach does not restrict the relationship between values at baseline and wave 2 in the way that analysing change does.

The major limitation is the loss to follow-up of 2571 eligible people, of whom 1894 declined to attend for physical examination at wave 2. Because they differed in their baseline characteristics from those retained, selection bias is a possibility. We were unable to find any association with energy intake, which might be because FFQ are not good at measuring absolute energy. They are better at measuring composition and patterns of the diet (i.e. how much the diet is comprised of protein, fat, etc.)( Reference Willett 22 ), which might explain our findings in this regard. It should also be noted that the questions asked on physical activity pertained to frequency, but not duration. We had no information on sedentary behaviour or sleep duration, which are also associated with weight gain( Reference Mozaffarian, Hao and Rimm 1 , 3 ). We had measured behaviours at baseline, and if these changed markedly over the 12 years between waves, attenuation of associations is possible. Regression to the mean was unlikely to have had a major impact on results, since body weight and waist circumference are not highly prone to diurnal variability or reader or other measurement error( Reference Ball, Crawford and Ireland 23 ).

Like us, Ellaway and colleagues found that neighbourhood of residence was associated with waist circumference after controlling for individual measures of affluence( Reference Ellaway, Anderson and Macintyre 24 ), and living in a socio-economically deprived area is consistently associated with higher weight( 3 , Reference Giskes, van Lenthe and Avendano-Pabon 25 ). These findings suggest that targeting individual behaviour alone is unlikely to reduce the proportion of overweight individuals in the population. Changing obesity-promoting environments to increase physical activity, reduce sedentary behaviour, increase the availability of healthier food products and decrease the availability of unhealthy food products should impact on a large proportion of the population and support individual behaviour change( 26 ). While difficult and expensive, such changes might be achieved through public policy, transport infrastructure, urban design and taxation. Any public education campaigns to improve eating habits and levels of physical activity should target disadvantaged population groups.

We found that people who were married or in de facto arrangements had lower weight and waist circumference at wave 2 than people who reported being single (widowed, divorced, separated or single), suggesting the importance of intimate social relationships to weight maintenance. On the other hand, a recent publication from the Whitehall II Study showed that adverse social relationships may contribute to weight and waist circumference increases over a 12-year period( Reference Kouvonen, Stafford and De Vogli 27 ). While we did not evaluate the quality of relationships, it may be that people living with partners had generally better social relationships than those on their own, or that being single was a marker of adverse social relationships. Future studies could explore how the number and quality of social relationships of various levels of intimacy relate to weight gain.

Waist circumference (but not weight) at wave 2 was positively associated with blood pressure at baseline. Elevated blood pressure is part of the constellation of abnormalities constituting the metabolic syndrome, along with abdominal obesity, dyslipidaemia, dysglycaemia and insulin resistance( Reference Despres, Lemieux and Bergeron 28 ), and it may be that behavioural and genetic factors contributing to rising blood pressure also contribute to visceral fat accumulation. Adjusting for baseline waist circumference may not fully account for visceral adiposity as the external waist measurement also includes subcutaneous fat.

In 2007, the World Cancer Research Fund and American Institute for Cancer Research panel concluded there was convincing evidence that physical activity of all types protects against weight gain( 3 ). Conversely, a more recent systematic review by Summerbell and colleagues concluded that the evidence was weak and inconsistent( Reference Summerbell, Douthwaite and Whittaker 2 ). As there is much heterogeneity between different study designs (including age, sex and baseline weight of participants, duration of follow-up, measurement of the exposure, inclusion and exclusion of participants who smoke or have a pre-existing chronic disease), it is difficult to draw meaningful conclusions. The evidence for an association with leisure-time physical activity, however, is more convincing. Two of three intervention trials in adults (see detailed review in reference 3) showed that weight decreased with exercise aimed at weight-loss maintenance. Several epidemiological studies have also shown some benefit of leisure-time physical activity on maintenance of weight and waist circumference( Reference Mozaffarian, Hao and Rimm 1 , 3 , Reference Ekelund, Besson and Luan 29 ). Two( Reference Koh-Banerjee, Chu and Spiegelman 30 , Reference Sherwood, Jeffery and French 31 ) of the three cohort studies( Reference Koh-Banerjee, Chu and Spiegelman 30 Reference Mosca, Marshall and Grunwald 32 ) that reported results for high-intensity physical activity showed inverse associations with future weight.

Previous studies have reported mixed associations between work-related physical activity and weight change (see detailed review in reference 3). These associations are likely to differ by SES. We found a positive association between moderate to heavy activity at work and future waist circumference, but not weight, even after adjusting for area-based SES, education level and leisure-time exercise. This finding, though, could be explained in part by participants no longer performing work-related activities by wave 2 due to retirement from work.

Limited information is available for the relationship between physical activity at home and weight or waist change. One study reported no association for women( Reference Sternfeld, Wang and Quesenberry 6 ), but another showed a slightly reduced risk of becoming obese for women who spent 40 h or more per week standing or walking around at home compared with inactive women( Reference Hu, Li and Colditz 33 ). We did not observe any associations for this variable and it was likely that this would have been more difficult to recall accurately than leisure-time activity, especially vigorous activity.

Several cohort studies have reported inconsistent findings with regard to alcohol intake and weight gain (see detailed review in Summerbell et al.( Reference Summerbell, Douthwaite and Whittaker 2 )). Similarly, the evidence for an association between alcohol intake and change in waist circumference is not convincing( Reference Walls, Magliano and McNeil 7 , Reference Halkjaer, Sorensen and Tjonneland 34 , Reference Newby, Muller and Hallfrisch 35 ). Our results showed that low to moderate alcohol consumption was inversely associated with future weight and waist circumference, even after adjustment for various potential confounders.

The lack of association between total energy intake and future weight or waist circumference for our population is consistent with results from most other epidemiological studies (reviewed in Summerbell et al.( Reference Summerbell, Douthwaite and Whittaker 2 )). Obesity is determined by the imbalance of energy intake and energy expenditure( Reference Vainio and Bianchini 36 ), but it is difficult to differentiate small differences in energy intake (assuming energy expenditure is constant) using FFQ. Additionally, under-reporting food intake is more common in heavier people( Reference Seidell 37 ). The ability to predict future weight and waist circumference based on total energy intake is also limited as people may change their dietary intakes over time.

In most published studies, carbohydrate as a percentage of energy intake at baseline was not associated with change in weight or waist circumference( Reference Summerbell, Douthwaite and Whittaker 2 ), although some studies (including ours) report negative associations with weight gain( Reference Colditz, Willett and Stampfer 38 , Reference Ludwig, Pereira and Kroenke 39 ). Results from a meta-regression showed that fat intake as a percentage of total energy intake was not associated with change in weight (effect size = 0·07, 95 % CI −0·03, 0·16)( Reference Summerbell, Douthwaite and Whittaker 2 ). However, a couple of studies not included in the meta-analysis reported positive associations between fat intake and weight gain( Reference Mosca, Marshall and Grunwald 32 , Reference Coakley, Rimm and Colditz 40 ). Protein intake (typically assessed from an FFQ) was not associated with weight gain in most previous studies( Reference Summerbell, Douthwaite and Whittaker 2 ), but results were inconsistent. There is also evidence that protein is associated with satiety and could help with weight loss( Reference Paddon-Jones, Westman and Mattes 41 ). In contrast, two studies showed that protein intake was inversely associated with changes in waist circumference( Reference Ludwig, Pereira and Kroenke 39 , Reference Halkjaer, Tjonneland and Thomsen 42 ). Further research is needed to evaluate how subdivisions of energy, protein, fat and carbohydrate intakes are associated with weight gain.

Findings from the current prospective cohort study indicate that promoting vigorous exercise and encouraging a high-carbohydrate, low-fat and low-protein diet could be promising approaches for obesity prevention in middle age. These recommendations accord with current behavioural recommendations for chronic disease prevention (e.g. Dietary Guidelines for Australian Adults ( 43 ), National Physical Activity Guidelines for Adults ( 44 )). In the present study, low to moderate alcohol consumption was linked to lower weight than was abstaining from drinking. Current Australian guidelines for alcohol use recommend people limit alcohol intake if they choose to drink (average of no more than two standard drinks per day)( 45 ), because for some health outcomes, alcohol abstinence may be beneficial. Observed associations between residing in socially disadvantaged suburbs and increased weight and waist circumference highlight the importance of modifying settings to support healthy individual behaviour. Driving environmental change and supporting low-income communities to improve their levels of physical activity and healthy eating are appropriately listed as key action areas for obesity prevention in Australia( 26 ).

Acknowledgements

Sources of funding: This work was supported by infrastructure from the Cancer Council Victoria and grants from VicHealth 2001–0447 and the National Health and Medical Research Council 209057 and 251533. Conflicts of interest: The authors declare no conflict of interest. Authors' contributions: conception and design, G.G.G., D.R.E., A.M.H. and H.G.D.; analysis and interpretation of data, R.J.M., A.M.H., H.G.D., A.P., L.E.A.J., D.R.E. and G.G.G.; drafting the article, R.J.M.; revising the article critically for important intellectual content, A.M.H., H.G.D., A.P., L.E.A.J., D.R.E. and G.G.G.; final approval of the version to be published, R.J.M., A.M.H., H.G.D., A.P., L.E.A.J., D.R.E. and G.G.G. Acknowledgements: This study was made possible by the contribution of many people, including the original investigators and the diligent team who recruited the participants and who continue working on follow-up. The authors would also like to express their gratitude to the many thousands of Melbourne residents who continue to participate in the study.

References

1. Mozaffarian, D, Hao, T, Rimm, EB et al. (2011) Changes in diet and lifestyle and long-term weight gain in women and men. N Engl J Med 364, 23922404.CrossRefGoogle ScholarPubMed
2. Summerbell, CD, Douthwaite, W, Whittaker, V et al. (2009) The association between diet and physical activity and subsequent excess weight gain and obesity assessed at 5 years of age or older: a systematic review of the epidemiological evidence. Int J Obes (Lond) 33, Suppl. 3, S1S92.Google ScholarPubMed
3. World Cancer Research Fund & American Institute for Cancer Research (2007) Food, Nutrition, Physical Activity and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institute for Cancer Research.Google ScholarPubMed
4. Du, H, van der A, DL, Ginder, V et al. (2009) Dietary energy density in relation to subsequent changes of weight and waist circumference in European men and women. PLoS One 4, e5339.CrossRefGoogle ScholarPubMed
5. Romaguera, D, Angquist, L, Du, H et al. (2011) Food composition of the diet in relation to changes in waist circumference adjusted for body mass index. PLoS One 6, e23384.CrossRefGoogle ScholarPubMed
6. Sternfeld, B, Wang, H, Quesenberry, CP 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.CrossRefGoogle ScholarPubMed
7. Walls, HL, Magliano, DJ, McNeil, JJ et al. (2011) Predictors of increasing waist circumference in an Australian population. Public Health Nutr 14, 870881.CrossRefGoogle Scholar
8. Rankinen, T, Kim, SY, Perusse, L et al. (1999) The prediction of abdominal visceral fat level from body composition and anthropometry: ROC analysis. Int J Obes Relat Metab Disord 23, 801809.CrossRefGoogle ScholarPubMed
9. Elobeid, MA, Padilla, MA, Brock, DW et al. (2010) Endocrine disruptors and obesity: an examination of selected persistent organic pollutants in the NHANES 1999–2002 data. Int J Environ Res Public Health 7, 29883005.CrossRefGoogle ScholarPubMed
10. Fogelholm, M, Anderssen, S, Gunnarsdottir, I et al. (2012) Dietary macronutrients and food consumption as determinants of long-term weight change in adult populations: a systematic literature review. Food Nutr Res 56, doi:10.3402/fnr.v56i0.19103.CrossRefGoogle ScholarPubMed
11. Simonsen, MK, Hundrup, YA, Gronbaek, M et al. (2008) A prospective study of the association between weight changes and self-rated health. BMC Womens Health 8, 13.CrossRefGoogle ScholarPubMed
12. Manson, JE, Bassuk, SS, Hu, FB et al. (2007) Estimating the number of deaths due to obesity: can the divergent findings be reconciled? J Womens Health (Larchmt) 16, 168176.CrossRefGoogle ScholarPubMed
13. Giles, GG & English, DR (2002) The Melbourne Collaborative Cohort Study. IARC Sci Publ 156, 6970.Google ScholarPubMed
14. Lohman, TG, Roche, AF & Martorell, R (editors) (1988) Anthropometric Standardization Reference Manual. Champaign, IL: Human Kinetics Books.Google Scholar
15. MacInnis, RJ, English, DR, Gertig, DM et al. (2003) Body size and composition and prostate cancer risk. Cancer Epidemiol Biomarkers Prev 12, 14171421.Google ScholarPubMed
16. Australian Bureau of Statistics (1996) Index of Relative Socio-economic Advantage and Disadvantage. ABS Catalogue no. 2033.0.55.001. Canberra: Commonwealth of Australia.Google Scholar
17. National Heart Foundation of Australia & Australian Institute of Health (1991) Risk Factor Prevalence Study: Survey No. 3 1989, Cities Analysis. Canberra: National Heart Foundation of Australia and Australian Institute of Health.Google Scholar
18. Ireland, P, Jolley, D, Giles, G et al. (1994) Development of the Melbourne FFQ: a food frequency questionnaire for use in an Australian prospective study involving an ethnically diverse cohort. Asia Pac J Clin Nutr 3, 1931.Google Scholar
19. Hodge, A, Patterson, AJ, Brown, WJ et al. (2000) The Anti Cancer Council of Victoria FFQ: relative validity of nutrient intakes compared with weighed food records in young to middle-aged women in a study of iron supplementation. Aust N Z J Public Health 24, 576583.CrossRefGoogle Scholar
20. Holland, B, Welch, AA, Unwin, ID et al. (editors) (1991) McCance and Widdowson's The Composition of Foods, 5th ed. Cambridge: The Royal Society of Chemistry and Ministry of Agriculture, Fisheries and Food.Google Scholar
21. Shahar, E & Shahar, DJ (2012) Causal diagrams and change variables. J Eval Clin Pract 18, 143148.CrossRefGoogle ScholarPubMed
22. Willett, W (1998) Nutritional Epidemiology, 2nd ed. New York: Oxford University Press.CrossRefGoogle ScholarPubMed
23. Ball, K, Crawford, D, Ireland, P et al. (2003) Patterns and demographic predictors of 5-year weight change in a multi-ethnic cohort of men and women in Australia. Public Health Nutr 6, 269281.CrossRefGoogle Scholar
24. Ellaway, A, Anderson, A & Macintyre, S (1997) Does area of residence affect body size and shape? Int J Obes Relat Metab Disord 21, 304308.CrossRefGoogle ScholarPubMed
25. Giskes, K, van Lenthe, F, Avendano-Pabon, M et al. (2011) A systematic review of environmental factors and obesogenic dietary intakes among adults: are we getting closer to understanding obesogenic environments? Obes Rev 12, e95e106.CrossRefGoogle ScholarPubMed
26. National Preventative Health Taskforce (2009) Obesity in Australia: A Need for Urgent Action. Canberra: Commonwealth of Australia.Google ScholarPubMed
27. Kouvonen, A, Stafford, M, De Vogli, R et al. (2011) Negative aspects of close relationships as a predictor of increased body mass index and waist circumference: the Whitehall II study. Am J Public Health 101, 14741480.CrossRefGoogle ScholarPubMed
28. Despres, JP, Lemieux, I, Bergeron, J et al. (2008) Abdominal obesity and the metabolic syndrome: contribution to global cardiometabolic risk. Arterioscler Thromb Vasc Biol 28, 10391049.CrossRefGoogle ScholarPubMed
29. Ekelund, U, Besson, H, Luan, J et al. (2011) Physical activity and gain in abdominal adiposity and body weight: prospective cohort study in 288,498 men and women. Am J Clin Nutr 93, 826835.CrossRefGoogle ScholarPubMed
30. Koh-Banerjee, P, Chu, NF, Spiegelman, D et al. (2003) Prospective study of the association of changes in dietary intake, physical activity, alcohol consumption, and smoking with 9-y gain in waist circumference among 16 587 US men. Am J Clin Nutr 78, 719727.Google ScholarPubMed
31. Sherwood, NE, Jeffery, RW, French, SA et al. (2000) Predictors of weight gain in the Pound of Prevention study. Int J Obes Relat Metab Disord 24, 395403.CrossRefGoogle ScholarPubMed
32. Mosca, CL, Marshall, JA, Grunwald, GK et al. (2004) Insulin resistance as a modifier of the relationship between dietary fat intake and weight gain. Int J Obes Relat Metab Disord 28, 803812.CrossRefGoogle ScholarPubMed
33. Hu, FB, Li, TY, Colditz, GA et al. (2003) Television watching and other sedentary behaviors in relation to risk of obesity and type 2 diabetes mellitus in women. JAMA 289, 17851791.CrossRefGoogle ScholarPubMed
34. Halkjaer, J, Sorensen, TI, Tjonneland, A et al. (2004) Food and drinking patterns as predictors of 6-year BMI-adjusted changes in waist circumference. Br J Nutr 92, 735748.CrossRefGoogle ScholarPubMed
35. Newby, PK, Muller, D, Hallfrisch, J et al. (2003) Dietary patterns and changes in body mass index and waist circumference in adults. Am J Clin Nutr 77, 14171425.Google ScholarPubMed
36. Vainio, H & Bianchini, F (editors) (2002) IARC Handbooks of Cancer Prevention. vol. 6: Weight Control and Physical Activity. Lyon: IARC Press.Google Scholar
37. Seidell, JC (1998) Dietary fat and obesity: an epidemiologic perspective. Am J Clin Nutr 67, 3 Suppl., 546S550S.Google Scholar
38. Colditz, GA, Willett, WC, Stampfer, MJ et al. (1990) Patterns of weight change and their relation to diet in a cohort of healthy women. Am J Clin Nutr 51, 11001105.Google Scholar
39. Ludwig, DS, Pereira, MA, Kroenke, CH et al. (1999) Dietary fiber, weight gain, and cardiovascular disease risk factors in young adults. JAMA 282, 15391546.CrossRefGoogle ScholarPubMed
40. Coakley, EH, Rimm, EB, Colditz, G et al. (1998) Predictors of weight change in men: results from the Health Professionals Follow-up Study. Int J Obes Relat Metab Disord 22, 8996.CrossRefGoogle ScholarPubMed
41. Paddon-Jones, D, Westman, E, Mattes, RD et al. (2008) Protein, weight management, and satiety. Am J Clin Nutr 87, issue 5, 1558S1561S.Google ScholarPubMed
42. Halkjaer, J, Tjonneland, A, Thomsen, BL et al. (2006) Intake of macronutrients as predictors of 5-y changes in waist circumference. Am J Clin Nutr 84, 789797.Google ScholarPubMed
43. National Health and Medical Research Council (2003) Food for Health: Dietary Guidelines for Australian Adults. Canberra: Commonwealth of Australia.Google ScholarPubMed
44. Department of Health and Aged Care (1999) National Physical Activity Guidelines for Adults. Canberra: Australian Government.Google ScholarPubMed
45. National Health and Medical Research Council (2009) Australian Guidelines to Reduce Health Risks from Drinking Alcohol. Canberra: Commonwealth of Australia.Google Scholar
Figure 0

Table 1 Characteristics of participants stratified by sex: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

Figure 1

Table 2 Associations of potential predictors of weight and waist circumference at baseline*: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

Figure 2

Table 3 Associations of potential predictors of weight and waist circumference at wave 2: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

Figure 3

Table 4 Selected associations of potential predictors of weight at wave 2 by age at attendance*: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

Figure 4

Table 5 Selected associations of potential predictors of waist circumference at wave 2 by age at attendance*: adults with data at baseline (1990–1994) and wave 2 (2003–2007), Melbourne Collaborative Cohort Study

You have Access
18
Cited by

Send article to Kindle

To send this article to your Kindle, first ensure no-reply@cambridge.org is added to your Approved Personal Document E-mail List under your Personal Document Settings on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part of your Kindle email address below. Find out more about sending to your Kindle. Find out more about sending to your Kindle.

Note you can select to send to either the @free.kindle.com or @kindle.com variations. ‘@free.kindle.com’ emails are free but can only be sent to your device when it is connected to wi-fi. ‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.

Find out more about the Kindle Personal Document Service.

Predictors of increased body weight and waist circumference for middle-aged adults
Available formats
×

Send article to Dropbox

To send this article to your Dropbox account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Dropbox.

Predictors of increased body weight and waist circumference for middle-aged adults
Available formats
×

Send article to Google Drive

To send this article to your Google Drive account, please select one or more formats and confirm that you agree to abide by our usage policies. If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your <service> account. Find out more about sending content to Google Drive.

Predictors of increased body weight and waist circumference for middle-aged adults
Available formats
×
×

Reply to: Submit a response

Please enter your response.

Your details

Please enter a valid email address.

Conflicting interests

Do you have any conflicting interests? *