• Policy & prevention
  • Advocacy
  • Projects
    • The DAPHNE Project
    • The Spotlight Project
    • INFORMAS
    • Marketing to children
    • Conflicts of Interest
    • Past EU projects
      • The HOPE Project
      • The ENERGY Project
      • The DYNAMO-HIA Project
        • Criteria for estimating relative risks
        • Estimates of relative risk by BMI Category
        • Estimates of relative risk of disease per unit of BMI above 22kg/m2
        • Data Sources for estimates of relative risk
      • ToyBox-study
      • Benefits of modest weight loss
  • World Obesity Publications
  • Advisory Network (STAN)
• Clinical care
• Action Initiative
• Journals
• Events
• World Obesity Day 2016
• The Lancet Commission on Obesity
• About Obesity
• External resources

Data sources

1. Criteria for selecting sources of RRs

There are many ways of assessing adiposity, e.g. waist circumference, waist-to-hip ratio, skin-fold thickness and Body Mass Index (using height and weight). In the present report, the search was restricted to studies and meta-analyses that used BMI as a way of categorising adiposity, and which reported the risk in comparison with a ‘normal’ BMI of between 18 kg/m2 and 24.9 kg/m2.

2. Search strategy

Computerised databases, library and internet searches

The PubMed database (http://www.ncbi.nlm.nih.gov/sites/entrez) was searched using the free search terms ‘Obesity’, ‘Overweight’, ‘BMI’, ‘adiposity’, ‘bodyweight’, ‘lifestyle’, in conjunction with the disease of interest, in order to identify relevant studies and researchers who could be contacted to obtain data or further information about the studies described.
Examples of meta-analyses included in the present review are:
• Bogers et al. Association of overweight with increased risk of CHD partly independent of blood pressure and cholesterol levels. Arch Intern Med, 2007; 167:1720-1728.
• McGee et al. Body Mass Index and mortality: a meta-analysis based on person level data from twenty-six observational studies. Ann Epidemiol, 2005; 15:87-97.
• McTigue et al. Obesity in older adults: a systematic review of the evidence for diagnosis and treatment. Obesity 2006; 14:1485-1497.
• Prospective Studies Collaboration. BMI and cause-specific mortality in 900,000 adults: collaborative analyses of 57 prospective studies. Lancet, 2009; 373:1083-1096.
• Renehan et al. BMI and incidence of cancer: a systematic review and meta-analysis of prospective observational studies. The Lancet 2008; 371:569-578.
• Vazquez G et al. Comparison of BMI, WC, WHR in predicting incident diabetes: a meta-analysis. Epd Revs 2007; doi:10.1093/epirev/mxm008.

Examples of large surveys also referred to in the present review are:
• Adams et al. Overweight, obesity and mortality in a large prospective cohort of persons 50-71 years old. NEJM 2006; 355:763-778.
• Baker et al. Childhood BMI and the risk of CHD in adulthood. NEJM 2007; 357:2329-2337.
• Banegas et al Mortality attributable to obesity in Europe, EJCN, 2003; 57:201-208.
• Calle et al BMI and mortality in a prospective cohort of US adults. NEJM 1999; 341:1097-1105.
• Carey VJ et al. Body fat distribution and risk of NIDDM in women. Am J Epid. 1997;145:614-619.
• Chen et al. BMI and mortality from IHD in a lean population: 10 year prospective study of 220,000 adult men. Int J Epidemiol, 2006; 35:141-150.
• Engeland A et al. Height and body mass index in relation to colorectal and gallbladder cancer in two million Norwegian men and women. Cancer Causes Control. 2005 Oct;16(8):987-96.
• Engeland A, Tretli S, Bjørge T. Height and body mass index in relation to esophageal cancer; 23-year follow-up of two million Norwegian men and women. Cancer Causes Control. 2004;15:837-43.
• Freedman DM et al. The mortality risk of smoking and obesity combined. Am J Prev Med, 2006; 31: 355-362.
• Hippisley-Cox et al. Predicting risk of type 2 diabetes in England and Wales: prospective derivation and validation of QDScore. BMJ 2009;338:b880. doi:10.1136/bmj.b880.
• Pischon T et al General and abdominal obesity and risk of death in Europe. NEJM 2008;359:2105-2120.
• Reeves et al. Cancer incidence and mortality in relation to BMI in the Million Women Study. BMJ 2007; 335:1134-1144.
• Schienkiewitz A et al. BMI history and risk of type 2 diabetes: results from the EPIC-Potdsdam Study. Am J Clin Nutr 2006; 84:427-433.
• Stevens et al, The effect of age on the association between BMI and mortality. NEJM 1998; 338:1-7..
• Wang Y et al. Comparison of abdominal adiposity and overall obesity in predicting risk of type 2 diabetes among men. Am J Clin Nutr 2005;81:555-563.

Additional studies were found using known sources of expertise, such as the UK Foresight study Tackling Obesities: Future Choices published by the UK government in 2007 (http://www.foresight.gov.uk/OurWork/ActiveProjects/Obesity/Obesity.asp), and the World Cancer Research Fund series of background meta-analyses conducted for the report Food, Nutrition, Physical Activity, and the Prevention of Cancer: a Global Perspective. Second Expert Report (2007) (http://www.dietandcancerreport.org/).

Contacts with experts
Contacts were made with experts in the field for references to published or unpublished data sources or for the identification of appropriate contact persons. Experts were defined as contact authors for large studies that examined the association between BMI and the selected outcomes, or authors of meta-analyses in the same field of research.

3. Characteristics of excluded studies of relative risks
Studies or meta-analyses were normally excluded if any one of the following criteria was satisfied:
• The measurement of exposure differed from that used for this project;
• The outcome measure was prognosis, pre-cancerous lesions or pre-disease markers rather than incident cases or mortality;
• The statistical analyses of the study did not examine any major confounding factors such as age, sex or smoking.

4. Characteristics of included studies of relative risks
Relative risk estimates for all-cause mortality, IHD, stroke, diabetes mellitus, lung cancer, colorectal cancer, mouth/oropharynx cancer, oesophagal cancer, breast cancer, endometrial cancer, kidney cancer and gallbladder cancer were obtained from literature searches and additional material described above.
No convincing data were found for a relationship between BMI and COPD and so it was assumed for the present analysis that there is no association (RR=1 for both genders and all age groups). Subsequent evidence may lead to a revision of this assumption.

5. Approach for selecting age and sex specific relative risks
When reporting the relative risk estimates for this specific workpackage, the following assumptions were made:

a) For each outcome, the same relative risk estimates were applied to all countries, assuming no interaction between an individual’s BMI and that individual’s country of residence on the associations. However, it is not possible to verify whether this assumption is true, as the study populations covered by the literature reviews were from limited numbers of countries, which did not allow comparisons between countries, although some evidence for differential relative risks may be found for some diseases when comparing Asian and Far Eastern population groups with European population groups;

b) Due to the limited evidence available for children, it was decided to apply a relative risk of one (1.0) for individuals under the age of 20 years; and

c) For the other age groups, we assumed that the relative risk estimates are the same for all age groups, except where we were able to obtain data by age group. In these cases the relative risks were adjusted to conform with the known effects of age on the association between BMI and disease outcome. The adjustments are shown in the tables in the next section.

6. Potential sources of uncertainty related to the choice of data sources used

A potential source of uncertainty in the data used to derive the estimates of relative risk shown above was the inconsistency in reporting of how BMI was obtained. In most individual studies the source of BMI was usually stated, and allowed the reader to assess the findings in terms of whether the BMI was based on measured or self-reported heights and weights. In the reviews and meta-analyses there was some inconsistency in reporting, and some mixing of data from surveys using different approaches to obtaining BMI.

A second potential uncertainty was the underlying population ethnicity. There is evidence that some ethnic groups – such as populations predominantly in Asia – are more sensitive to the health effects of adiposity than others (1,2),  possibly because they tend to show higher abdominal fat deposits for a given BMI than non-Asian populations (3).  In addition, differences in health status associated with BMI may also be found among immigrants to Europe from non-European regions, compared with indigenous Europeans (4).

Further sources of uncertainty should be mentioned. In the present report, the same relative risk relationships have been assumed for all EU member state population groups. The paucity of data on levels of obesity and the relationships between obesity and health in the various member states makes it impossible to assess whether this assumption is justified. When further evidence becomes available it can be used to refine the present model and improve the accuracy of the relative risk estimates.

In the present report a relative risk of 1.0 has been given to all people under the age of 20 years. There is a severe lack of information on the relationship between obesity and health outcomes for younger people, not least because most of the diseases of interest take several years to develop. There are many papers that indicate that children even under the age of 10 years have evidence of early signs of disease associated with increased BMI (5), but there is no clear evidence of relative risk of actual disease for these younger age groups. In the last decade a few children have been diagnosed with Type II diabetes, a disease previously so rare in this age groups that it was referred to as ‘adult onset diabetes’. The new cases of Type II diabetes in children are closely associated with excess body weight, but there are inadequate data to establish a relative risk estimate for these age groups. It is to be hoped that better evidence will be accumulated that can allow more accurate relative risks estimates to be applied to people under age 20 years.

References:
1) Lee ZSK et al. Obesity and cardiovascular risk factors in Hong Kong Chinese. Obes Reviews 2002, 3:173-182.
 
2) Li G, et al. Obesity, coronary heart disease risk factors and diabetes in Chinese: an approach to the criteria of obesity in the Chinese population. Obes Reviews 2002, 3:167-172

3) Deurenberg P et al. Asians are different from Caucasians and from each other in their body mass index/body fat per cent relationship. Obes Reviews 2002, 3:141-146.

4) Ujcic-Voortman JK, et al. Diabetes prevalence and risk factors among ethnic minorities. Eur J Public Health. 2009;19:511-5.

5) Lobstein T, Jackson-Leach R. Estimated burden of paediatric obesity and co-morbidities in Europe. Part 2. Numbers of children with indicators of obesity-related disease. Int J Pediatr Obes. 2006;1:33-41.