Geriatric Original Article

International Journal of Obesity (2012) 36, 1180–1186; doi:10.1038/ijo.2012.99; published online 26 June 2012

The relationship between body mass index prior to old age and disability in old age

K Backholer1,2,5, K Pasupathi1,2,5, E Wong1,2, A Hodge3, C Stevenson4 and A Peeters1,2

  1. 1Baker IDI Heart and Diabetes Institute, Melbourne, Victoria, Australia
  2. 2Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, Victoria, Australia
  3. 3Cancer Epidemiology Centre, Cancer Council Victoria, Melbourne, Victoria, Australia
  4. 4School of Health and Social Development, Deakin University, Melbourne, Victoria, Australia

Correspondence: Dr K Backholer, Baker IDI Heart and Diabetes Institute, 99 Commercial Road, Melbourne, Victoria 3004, Australia. E-mail: Kathryn.backholer@bakeridi.edu.au

5These authors contributed equally.

Received 14 February 2012; Revised 4 May 2012; Accepted 5 May 2012
Advance online publication 26 June 2012

Top

Abstract

Objectives:

 

To analyse the relationship between body mass index (BMI) in middle-age and disability status in old-age using data from the Melbourne Collaborative Cohort Study (MCCS).

Methods:

 

A total of 41514 participants enroled in the MCCS between 1990–1994. Height and weight were measured at baseline and disability, defined as limitations to self-care activities of daily living (ADLs) and self-care plus mobility activities, was identified at follow-up (2003–2007). In all, 6300 participants were <65 years at baseline, greater than or equal to70 years at follow-up and not missing BMI at baseline or ADLs at follow-up. The association between BMI in six categories (BMI 18.5–22.5; 22.5–25; 25–27.5; 27.5–30; 30–35; 35+) and disability status was analysed using logistic regression. Models were stratified by sex, and sequentially adjusted for age, education, country of birth, then smoking, alcohol, fruit and vegetable intake, and physical activity.

Results:

 

Adjusted odds ratios for composite self-care ADL and mobility limitations compared with BMI 18.5–22.5kgm−2 were 1.73 (95%CI 1.14–2.64) for BMI 30–35kgm−2 and 3.46 (1.78–6.73) for BMI 35+kgm−2 in males. In females, adjusted odds ratios were 1.29 (1.00–1.68) for BMI 22.5–25kgm−2, 1.74 (1.35–2.24) for BMI 25–27.5kgm−2, 2.58 (1.98–3.36) for BMI 27.5–30kgm−2, 2.74 (2.10–3.58) for BMI 30–35kgm−2 and 4.21 (3.12–5.88) for BMI 35+kgm−2.

Conclusion:

 

A graded relationship was observed between BMI and disability in males and females, across the continuum of BMI. These results highlight the importance of a healthy body weight at middle age in order to reduce the risk of disability in old age.

Keywords:

activities of daily living; overweight; disability; middle age

Top

Introduction

With mortality decreasing and life expectancy increasing, the challenge for public health has shifted from prolonging life to prolonging healthy life. Disability, an umbrella term that covers impairment, activity limitations and participation restrictions, is a commonly used marker of overall health status.1 Disability can be measured in a number of ways. One of the more common measures, recommended by the World Health Organisation,2 is through an index of activities of daily living (ADL), attempting to recognise the differences between normal functional ability and the current demands imposed by the environment. Obesity, through its association with a number of health risks, including diabetes, cardiovascular disease, cancer and osteoarthritis, has been shown to lead to an increased risk of disability in the elderly, while the relationship for overweight is not clear.3, 4, 5

Most studies to date measure weight status in the elderly, and use body mass index (BMI) as their marker of weight status. However, it has been shown that the accuracy of BMI as a measure of body fat decreases with age due to a natural increase in the ratio of fat to lean body mass that occurs in the elderly.6 Consequently, elderly people with adiposity levels associated with obesity are more likely to be classified as overweight when compared with a younger person with the same adiposity levels. Furthermore, the association between BMI and mortality has been shown to attenuate as age increases.7, 8 It is therefore possible that BMI measured at mid-life is a better predictor of ill-health in the elderly than BMI measured later in life. Further support for this notion was recently demonstrated in a review of the relationship between bodyweight and dementia. This study concluded that while there was a relationship between mid-life BMI and later life dementia, there was no relationship with late-life BMI.9

To assess the potential risk of future disability associated with excess weight, it is essential that this be studied in cohorts large enough to permit analysis of the relationship along the continuum of BMI and with sufficient follow-up time to distinguish between mid-life risk and later life disability. Furthermore, such cohorts need comprehensive measures of disability and to have evaluated possible confounders and effect modifiers. Although several studies have analysed the relationship between mid-life BMI and disability,10, 11, 12, 13, 14 data of this kind are lacking. With the prevalence of overweight and obesity increasing across all age groups, it is important to know to what extent excess weight prior to old age is associated with disability in old age.

In this study, we analyse the relationship between six BMI categories in mid-life and disability in old age. We use the 14-year follow-up of a large Australian cohort study, the Melbourne Collaborative Cohort Study (MCCS). We analyse the relationship between BMI measured at baseline (1990–1994) in participants aged less than 65 years and disability measured at follow-up (2003–2007) in those aged 70 years or over.

Top

Materials and methods

Data source

The MCCS is a prospective cohort study with subjects recruited from metropolitan Melbourne through the electoral roll, advertisements and announcements in community papers, ethnic radio and churches. The study deliberately oversampled migrants from Southern European countries to extend the range of lifestyle and genetic factors. A total of 41514 individuals were recruited between 1990 and 1994. During a face-to-face interview, a questionnaire was administered with a range of demographic, lifestyle and medical history questions, and trained professionals took physical measurements including height (cm) and weight (kg). At 4 years post-baseline, participants were followed up with a postal questionnaire similar to the baseline questionnaire (response rate 87%). Between 2003 and 2007, a further face-to-face interview was conducted, similar to baseline and with some additional questions including those pertaining to disability (response rate among those eligible for the current study was 57%). Further details of the study have been described elsewhere.15

For the current study, we included participants with complete measures of BMI and potential confounders at baseline (n=41443), who were less than age 65 at baseline (n=34335), and alive and aged 70 years or over at the 2003–2007 follow-up (n=7668). We also excluded the 39 participants with BMI<18.5 as this category was too small to analyse. Of the remaining participants, 1365 were missing ADL information at the 2003–2007 follow-up and were excluded (final n=6264, 2518 males and 3745 females).

Measure of adiposity

At baseline, trained interviewers measured height (with a stadiometer, cm) and weight (with a digital scale, kg). BMI was calculated as weight in kilograms divided by height in metres squared. BMI was categorised into six categories based on World Health Organisation recommendations:16 BMI 18.5–<22.5kgm−2 (lower normal weight), 22.5–<25kgm−2 (upper normal weight), 25–<27.5kgm−2 (lower overweight), 27.5–<30kgm−2 (upper overweight), 30–<35kgm−2 (lower obese) and 35+kgm−2 (upper obese).

Measures of disability

A number of questions about disability were asked at the 2003–2007 follow-up. For this analysis, we included data on: bathing, dressing, eating, getting out of a chair or bed, going to or using the toilet at home and walking about 200–300m (round the block). Subjects were asked if their health limited each of these activities and response options were ‘none’, ‘some’, ‘a lot’ and ‘cannot do’. When assessing each activity independently, disability was defined as a response other than none. We analysed two composite measures of disability, defined as disability in any of the six activities (five ADLs and walking) and any disability in the five basic self-care ADLs only.

Other covariates

Baseline covariates collected through the questionnaire included education level, country of birth, fruit and vegetable intake, alcohol intake, physical activity, smoking status and presence of chronic disease. Education was ascertained from the question ‘What is the highest level of education you completed?’, and was dichotomised into those who had completed high school and above and those who had not finished high school. Country of birth was categorised as Southern European (Italian/Greek/Maltese) versus not (Australian, New Zealand, British, Dutch, Irish, Scottish and Welsh). Fruit and vegetable intake was collected using a self-administered food frequency questionnaire specifically developed for use in the MCCS,17 and was dichotomised into those meeting and not meeting national guidelines, where meeting guidelines was determined as consuming more than two and five servings per day of fruit and vegetables, respectively.18 Alcohol intake was ascertained by asking participants their usual quantity and frequency per day of alcohol consumption for the current decade, and analysed according to meeting current guidelines of no more than two standard drinks per day (up to 20g per day).19 Physical activity status was derived from three questions asking participants how many times a week they walked for recreation or exercise and did vigorous or non-vigorous activities in their leisure time in the 6 months before the questionnaire. The physical activity score was then derived by combining the three activity types and used as a continuous variable.20 For smoking status, participants were asked if they were never, past or current smokers, and we analysed our data using the same categories. Self-reported past medical history was obtained for cardiovascular disease (including hypertension, angina, acute myocardial infarction and strokes), diabetes, arthritis, asthma and cancer.

At the 4-year post-baseline, postal questionnaire participants were asked ‘In the past 12 months, to what extent did health problems limit your everyday physical activity?’, with response options of ‘not at all’, ‘a little bit’, ‘moderately’, ‘quite a bit’ and ‘extremely’. The response to this question was used in the sensitivity analysis discussed below. In this question, physical activity was subjectively defined by each participant and was not related to the questions used to derive physical activity status.

Analyses

Descriptive statistics were used to compare baseline characteristics between responders and non-responders, and across BMI categories.

Logistic regression analysis was used to analyse the relationship between BMI category at baseline and disability at follow-up. All analyses were stratified by sex. The variable used for BMI encompassed the six BMI categories defined above, and the BMI category 18.5–<22.5kgm−2 was used as the reference category. Three models were used: the first adjusting for demographic variables including age, education status and country of birth; the second additionally adjusting for behavioural variables including smoking status, physical activity, alcohol intake, and fruit and vegetable intake; and the third additionally adjusting for chronic diseases including the prevalence of angina, asthma, diabetes, arthritis, cancer, heart attack, stroke and hypertension at baseline.

Analyses were also performed for subgroups categorised according to smoking status (never smokers), country of birth (Southern European-born participants) and chronic disease status (those free of chronic diseases at baseline that directly limit ADL’s, angina, arthritis, heart attack and stroke). In further sensitivity analyses, to test the effect of possible reverse causation, we tested the effect of repeating the analysis in a population whose physical activities were not at all limited by health problems at the 4-year postal follow-up (n=997 males and 1767 females).

All analyses were performed using Stata 11.1 (Stata Corp. LP., College Station, TX, USA) and significance was considered to be P<0.05.

Ethics

The MCCS study protocol was approved by the Cancer Council Victoria’s Human Research Ethics Committee and subjects gave written consent to participate. The current study was approved by the MCCS data committee and the Monash University standing committee on ethics in research involving humans (CF11/2075 - 2011001130).

Top

Results

Those who returned to the follow-up exam and completed the disability measures within the questionnaire were more likely to be higher educated, to meet alcohol guidelines and to have prevalent cancer, and less likely to be Southern European-born, be physically inactive, current smokers and have prevalent diabetes or arthritis (Table 1). No differences in baseline age, sex or BMI were observed.


As BMI increased, the prevalence of high education and current smoking status significantly decreased. The prevalence of Southern European-born, physical inactivity and prevalent angina, diabetes, arthritis and hypertension significantly increased as BMI increased. Significant trends in sex and alcohol were less clear. No significant differences in age and prevalent heart attack and stroke were observed across BMI categories (Table 2).


The prevalence of each of the individual activity limitations measured at follow-up, with the exception of eating, generally increased in both males and females as baseline BMI increased (Figure 1). Composite measures for disability, regardless of whether mobility was included or not, also increased as BMI increased (Figure 1). Among males, the prevalence of individual activity limitations and composite disability measures were generally lowest in the BMI category 22.5–25.0kgm−2, with prevalence increasing from either category 25.0–27.5 or 27.5–30kgm−2. Among women, a continuous increase in prevalence of individual and composite measures of disability was seen from BMI category 18.5–22.5kgm−2 (with the exception of eating). The highest prevalence of disability was observed with the composite measure of self-care ADL plus mobility limitations in those with BMI 35+kgm−2, with a prevalence of 62% in males and 74% in females.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Prevalence of disability across BMI categories, (a) males and (b) females.

Full figure and legend (104K)

After adjustment for the demographic variables strong, graded relationships were observed between BMI category at baseline and disability status at follow-up, independent of the measure of disability status used, and in both males and females (Table 3). A significant dose response relationship was observed for all analyses of the relationship between BMI category and disability (Table 3). Among males, an increased risk of self-care ADL limitations was observed from BMI category 30–35kgm−2, and from BMI category 27.5–30kgm−2 when combining self-care ADL and mobility limitations. Among females, a relationship was observed from BMI category 22.5–25kgm−2 for both composite measures of disability. For males, the BMI category 35+kgm−2 was associated with around a 3.5-fold increase in the odds of both composite measures of disability. For females, the BMI category 35+kgm−2 was associated with a 4.4-fold increase in the odds of self-care ADL limitations and a 6.6-fold increase in the odds of composite self-care ADL and mobility limitations.


After additional adjustment for the behavioural variables, the strength of the relationships decreased slightly and the significance of the relationship for some of the lowest BMI categories became marginal (Table 3). After additional adjustment for the presence of chronic disease at baseline, the strength of the relationships between BMI category and both measures of disability decreased further, and the relationship for some of the lower BMI categories lost significance. However, the dose response relationships remained strong and significant for all outcomes (Table 3).

We examined potential effect modification by smoking and country of birth by repeating the logistic regression analyses in never smokers (n=1065 for males, 2756 for females) and in those not born in Southern Europe (n=1736 for males, 2863 for females; Figure 2 for the composite outcome measure of self-care ADL and mobility limitations). When never smokers were compared to the total population, a stronger relationship between BMI and both composite measures of disability were observed, but the same basic patterns were still evident. Risk of composite self-care ADL and mobility limitations increased from BMI 27.5 to 30kgm−2 in males and from BMI 22.5 to 25kgm−2 in females (Figure 2). In non-Southern European-born participants, the magnitude of the odds ratios was very similar to those in the total population, although significance was lost for some owing to the smaller sample size.

Figure 2.
Figure 2 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Logistic regression analysis (Odds ratios from a model adjusting for age, education status, country of birth, smoking status, physical activity, alcohol intake, and fruit and vegetable intake) of BMI category at baseline on composite self-care and mobility disability at follow-up in: the total population; never smokers; and those free of chronic disease (chronic disease includes angina, diabetes, arthritis, cancer, stroke and hypertension at baseline. Comparisons are shown for males and females. Note: y axis is logarithmic.

Full figure and legend (70K)

We also repeated the logistic regression analyses in those without disability-associated chronic disease at baseline (n=1493 for males, 1837 for females; Figure 2 for the composite outcome measure of self-care ADL and mobility limitations). In this group, the magnitude of the relationship was generally slightly reduced for males and wide confidence intervals were observed for each BMI category. There was no appreciable change in the odds of disability for females. A dose response relationship was significant for both males and females.

Sensitivity analyses conducted in a population who were not limited by health problems at the 4-year postal follow-up did not appreciably alter any results.

Top

Discussion

In this analysis of the relationship between BMI in middle age and disability in old age, we demonstrate a strong dose response relationship for both males and females. This relationship is evident whether disability is defined as any limitation to the five basic self-care ADLs or when also including mobility-related disability. For males, an increased risk of disability was observed from the upper overweight or lower obese category, whereas for females it was observed from the upper normal weight or lower overweight category. For males, the upper overweight category was associated with an approximately 50% increase in the odds of disability, and the upper obese category with around a threefold increase. For females, the upper overweight category was associated with an approximately 2.5-fold increase in the odds of disability, and the upper obese category with around a 4- to 6-fold increase.

Our finding that obesity confers an increased risk of disability is consistent with three recent reviews on this topic.3, 4, 5 However, in contrast to these reviews, that demonstrate a modest or no increased risk of disability associated with overweight, we also observe an increased risk of disability associated with overweight females, and to a lesser extent, overweight males. Moreover, the magnitude of association between obesity and disability appears to be considerably greater in our study compared to the point estimates included in these reviews. Several methodological differences may explain these discrepancies including the diverse populations and definitions of disability utilised. However, it is also possible that heterogeneity between studies arises from the age at which BMI was measured. We analysed BMI measured in middle age to avoid the possible confounding of illness-associated weight loss in old age. Conversely, the majority of studies included in the three reviews measure BMI in older aged participants. As we hypothesised, it may be that BMI measured at middle age is a better marker of ill health in later life than is BMI measured in older age. Support for this hypothesis is evident from a study by Bussetto et al.21 who examined the association of both recalled BMI at age 50 and contemporaneous BMI with disability for an elderly Italian population. This study demonstrated a 1.20-fold (95% CI 1.11–1.30) and 2.19-fold (95% CI 2.00–2.40) increased risk of ADL disability associated with overweight and obesity, respectively, when weight was recalled from age 50, which was consistently higher than the risks associated with contemporary BMI. Further support for this hypothesis was highlighted in a recent review on the risk of dementia associated with overweight and obesity, which illustrated a graded positive risk relationship across all BMI categories when BMI was measured in mid-life, whereas no association was derived from a measure of BMI in late-life.9

The risk relationship between increased BMI and disability appears to be changing over time. Analysis based on US data collected at two time points (1988–1994 and 1999–2000) demonstrated a stronger relationship between overweight, obesity and ADL limitations in more recent periods.22 Although our study uses baseline data collected ~20 years ago, this lag time is required to allow middle-aged participants to reach old-age and develop disability. Nevertheless, it is essential that analyses such as those conducted herein are repeated as more contemporary data arise.

Our results are also consistent with those of Sun et al.23 who observed a linearly reduced odds of healthy survival after age 70 associated with increasing mid-life BMI among females. Like ours, this study also used finer categorisation of BMI, and observed an increased risk from the upper normal weight range. Although not directly comparable to our study due to the different outcome measures (healthy survival in the study by Sun et al. was defined as having no history of major chronic disease and no substantial cognitive, physical or mental limitations), these studies highlight that the longer-term risks associated with excess weight may occur below the current thresholds for overweight and obesity.

We additionally observe a consistently stronger relationship between BMI and disability among females compared with males. These sex differences have been demonstrated in other studies4, 24, 25 and may reflect known sex differences in fat distribution,26 recovery from disability27 and/or sex-specific causal pathways between obesity and disability.28 Furthermore, as women are more likely to have conditions that disable (for example, musculoskeletal disease) and men are more likely to have conditions that result in mortality (for example, cardiovascular disease), it is thus possible that the males that are still alive at follow-up represent a healthier group than that of females at follow-up.

Smoking has been shown to be a confounder of the association between BMI and disability.29 However, we demonstrate that the results presented herein are not excessively influenced by smoking as we observe similar relationships in smoking-adjusted analyses. Our results do, however, suggest a role of smoking in modifying the relationship between BMI and disability. We observe a stronger relationship between BMI and disability among male never smokers when compared with all males combined. This is likely due to the higher proportion of smokers in the normal weight group who are at a greater risk of disability, and highlights the need to stratify by smoking status when assessing BMI as a risk factor for ill health.7, 14 The moderating effect of smoking was not evident in females, likely due to the lower prevalence of smoking among females.

The key strength of our study was the long follow-up period and the large sample size that allowed evaluation of dose response risk across small BMI categories. Other strengths included objective measurement of height and weight and the rich data source that allowed for extensive adjustments in multiple regression models.

The major limitation of this study was the lack of disability data at baseline, and thus our inability to exclude the possibility of reverse causation—that BMI may be a consequence rather than a cause of disability. To address this limitation, we conducted a secondary analysis examining the relationship between BMI and disability in a population free of chronic disease at baseline. Although results from this analysis were attenuated and lost significance in males, an overall dose response relationship was nonetheless evident for both sexes. Such a sensitivity analysis likely removes some of the true effect of increased bodyweight on future disability (where increased body weight has already resulted in onset of chronic disease), suggesting that the causal effect is likely to lie between the two sets of estimates. This causal pathway of action is supported with our chronic disease adjusted models whereby the magnitude of association between excess body weight and disability was reduced. We additionally examined the limitation of reverse causation by assessing the association between BMI and disability in a population where health problems did not limit everyday physical activity at 4 years post-baseline. This analysis did not appreciably change any of the results.

A further limitation to this analysis is the loss to follow-up in the MCCS cohort, with only 57% of eligible participants providing ADL information. It is possible that the reasons for not attending follow-up included limitations in mobility or ADLs, and indeed non-responders were found to be less healthy at baseline. Such an effect is likely to be associated with an underestimation of the relationship. However, we observed no difference in mean baseline BMI between responders and non-responders and thus have no a priori reason to expect differences in follow up disability.

It is important to recognise that the disability measures used here are only one series of constructs for measuring disability. Disability can be defined in a number of ways from limitations in mobility, physical function and instrumental ADLs (which assess the ability of an individual to live within a community) and basic self-care ADLs, all of which represent varying severity and positions along the spectrum of the disablement process. Our results show that BMI was associated individually with each ADL except eating, and also with our two composite measures of disability, suggesting that excess body weight is a likely risk factor for a wide range of disability measures.

In conclusion, this study highlights the importance of maintaining a healthy body weight in middle age to reduce the risk of disability in later life. The associations observed are greater in magnitude, and significant increases in the risk of disability start from a lower BMI, than previously reported. This discrepancy may be due to the age at which BMI was measured, and whether measuring BMI at mid-life is a better predictor of disability later in life warrants further investigation. If such risks are borne out, the combination of the increasing prevalence of overweight and obesity with a rapidly aging population will lead to large increases in the prevalence of disability in coming decades. These findings highlight the need to better treat and manage obesity related comorbidities to prevent the progression to disability, and more importantly to prevent weight gain in midlife.

Top

Conflict of interest

The authors declare that they have no conflict of interest.

Top

References

  1. World Health OrganisationHealth topics: Disabilities, 2010.
  2. de Bruin A, Picavet HS, Nosikov A. Health Interview Surveys: towards international harmonization of methods and instruments. WHO Reg Publ Eur Ser 1996; 58: i–161. | PubMed |
  3. Vincent HK, Vincent KR, Lamb KM. Obesity and mobility disability in the older adult. Obes Rev 2010; 11: 568–579. | Article | PubMed |
  4. Backholer K, Wong E, Freak-Poli R, Walls HL, Peeters A. Increasing body weight and risk of limitations in activities of daily living: a systematic review and meta-analysis. Obes Rev 2012; 13: 456–468. | Article | PubMed |
  5. Rejeski WJ, Marsh AP, Chmelo E, Rejeski JJ. Obesity, intentional weight loss and physical disability in older adults. Obes Rev 2010; 11: 671–685. | Article | PubMed | ISI |
  6. Prentice AM, Jebb SA. Beyond body mass index. Obes Rev. 2001; 2: 141–147. | Article | PubMed | CAS |
  7. Stevens J. Impact of age on associations between weight and mortality. Nutr Rev 2000; 58: 129–137. | Article | PubMed | CAS |
  8. Vasconcelos Fde A, Cordeiro BA, Rech CR, Petroski EL. Sensitivity and specificity of the body mass index for the diagnosis of overweight/obesity in elderly. Cad Saude Publica 2010; 26: 1519–1527. | Article | PubMed |
  9. Anstey KJ, Cherbuin N, Budge M, Young J. Body mass index in midlife and late-life as a risk factor for dementia: a meta-analysis of prospective studies. Obes Rev 2011; 12: e426–e437. | Article | PubMed |
  10. Ebrahim S, Wannamethee SG, Whincup P, Walker M, Shaper AG. Locomotor disability in a cohort of British men: the impact of lifestyle and disease. Int J Epidemiol 2000; 29: 478–486. | Article | PubMed |
  11. Stenholm S, Sainio P, Rantanen T, Koskinen S, Jula A, Heliovaara M et al. High body mass index and physical impairments as predictors of walking limitation 22 years later in adult Finns. J Gerontol A Biol Sci Med Sci 2007; 62: 859–865. | Article | PubMed |
  12. Launer LJ, Harris T, Rumpel C, Madans J. Body mass index, weight change, and risk of mobility disability in middle-aged and older women. The epidemiologic follow-up study of NHANES I. JAMA 1994; 271: 1093–1098. | Article | PubMed | ISI | CAS |
  13. Houston DK, Stevens J, Cai J. Abdominal fat distribution and functional limitations and disability in a biracial cohort: the Atherosclerosis Risk in Communities Study. Int J Obes (Lond) 2005; 29: 1457–1463. | Article | PubMed | CAS |
  14. Peeters A, Bonneux L, Nusselder WJ, De Laet C, Barendregt JJ. Adult obesity and the burden of disability throughout life. Obes Res 2004; 12: 1145–1151. | Article | PubMed | ISI |
  15. Giles GG, English DR. The Melbourne Collaborative Cohort Study. IARC Sci Publ: Melbourne, 2002; 156: 69–70.
  16. World Health OrganisationObesity. Preventing and managing the global epidemic. Geneva: World Health Organisation, 1998, 3–5 June 1997.
  17. Hodge A, Patterson AJ, Brown WJ, Ireland P, Giles G. 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 NZ J Public Health 2000; 24: 576–583. | Article | CAS |
  18. National Health and Medical Research Council. Food for HealthDietary Guidelines for All Australians. National Health and Medical Research Council: Canberra, Australia, 2003.
  19. National Health and Medical Research CouncilAustralian Alcohol Guidelines: Health Risks and Benefits. National Health and Medical Research Council: Canberra, Australia, 2011.
  20. MacInnis RJ, English DR, Hopper JL, Haydon AM, Gertig DM, Giles GG. Body size and composition and colon cancer risk in men. Cancer Epidemiol Biomarkers Prev 2004; 13: 553–559. | PubMed | ISI |
  21. Busetto L, Romanato G, Zambon S, Calo E, Zanoni S, Corti MC et al. The effects of weight changes after middle age on the rate of disability in an elderly population sample. J Am Geriatr Soc 2009; 57: 1015–1021. | Article | PubMed |
  22. Alley DE, Chang VW. The changing relationship of obesity and disability, 1988-2004. JAMA 2007; 298: 2020–2027. | Article | PubMed | ISI | CAS |
  23. Sun Q, Townsend MK, Okereke OI, Franco OH, Hu FB, Grodstein F. Adiposity and weight change in mid-life in relation to healthy survival after age 70 in women: prospective cohort study. BMJ 2009; 339: b3796. | Article | PubMed |
  24. Chen H, Bermudez OI, Tucker KL. Waist circumference and weight change are associated with disability among elderly Hispanics. J Gerontol A Biol Sci Med Sci 2002; 57: M19–M25. | Article | PubMed |
  25. Larrieu S, Peres K, Letenneur L, Berr C, Dartigues JF, Ritchie K et al. Relationship between body mass index and different domains of disability in older persons: the 3C study. Int J Obes Relat Metab Disord 2004; 28: 1555–1560. | Article | PubMed | CAS |
  26. Wu CH, Yao WJ, Lu FH, Yang YC, Wu JS, Chang CJ. Sex differences of body fat distribution and cardiovascular dysmetabolic factors in old age. Age Ageing 2001; 30: 331–336. | Article | PubMed | ISI | CAS |
  27. Leveille SG, Penninx BW, Melzer D, Izmirlian G, Guralnik JM. Sex differences in the prevalence of mobility disability in old age: the dynamics of incidence, recovery, and mortality. J Gerontol B Psychol Sci Soc Sci 2000; 55: S41–S50. | Article | PubMed |
  28. Wray LA, Blaum CS. Explaining the role of sex on disability: a population-based study. Gerontologist 2001; 41: 499–510. | Article | PubMed | CAS |
  29. Majer IM, Nusselder WJ, Mackenbach JP, Kunst AE. Life expectancy and life expectancy with disability of normal weight, overweight, and obese smokers and nonsmokers in Europe. Obesity (Silver Spring) 2011; 19: 1451–1459. | Article | PubMed |
Top

Acknowledgements

AP was supported by the VicHealth fellowship, EW was supported by the Monash University Australian Postgraduate Award and CS was supported by the National Health and Medical Research (Grant No. 465130). AH supported by NHMRC No. 520316. We thank the Melbourne Collaborative Cohort Study investigators and participants, and Damien Jolley for assistance with graph presentation.

Extra navigation

.

naturejobs

natureevents

ADVERTISEMENT