Article Text

Download PDFPDF

BMI, CVD and mortality
Comparison of the associations of BMI and measures of regional adiposity with coronary heart disease, diabetes and all cause mortality: a study using data from four UK cohorts
  1. A. E. Taylor1,
  2. S. Ebrahim2,
  3. Y. Ben-Shlomo1,3,
  4. R. M. Martin1,3,
  5. P. H. Whincup4,
  6. J. W. Yarnell5,
  7. S. G. Wannamethee6,
  8. D. A. Lawlor1,3
  1. 1
    Department of Social Medicine, University of Bristol, Canynge Hall, Bristol, UK
  2. 2
    Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK
  3. 3
    MRC Centre for Causal Analyses in Translational Epidemiology, Department of Social Medicine, University of Bristol, Oakfield House, Bristol, UK
  4. 4
    Division of Community Health Sciences, St. George’s, University of London, London, UK
  5. 5
    Department of Epidemiology and Public Health, Queen’s University Belfast, Belfast, UK
  6. 6
    British Regional Heart Study, UCL Department of Primary Care and Population Sciences, Royal Free and University College Medical School, London, UK

    Statistics from

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


    To compare magnitudes of association of BMI and measures of regional adiposity with coronary heart disease, diabetes and all-cause mortality.


    2 prospective cohort and 2 cross sectional studies.


    Members of 4 UK cohorts aged ⩾45 years: 3937 women from the British Women’s Heart and Health Study (BWHHS), 2367 and 1950 men from phases 1 and 3 of the Caerphilly Prospective Study (CaPS), 403 men and women from Boyd Orr, 789 men and women from Maidstone and Dewsbury (MD) study.

    Main Outcome Measures

    Incident all cause mortality, coronary heart disease (CHD) and diabetes (in CaPS and BWHHS), cross sectional measures of arterial plaques (in Boyd Orr and MD), blood lipids, blood pressure (in all 4 cohorts) and fasting glucose and insulin (in BWHHS, phase 1 CaPS and Boyd Orr).


    In BWHHS women, all measures of adiposity were strongly positively associated with incident diabetes, the strongest relationships seen with waist circumference (WC) (HR 2.35, 95% CI 2.03 to 2.73) and waist: height ratio (WHtR) (HR 2.29, 95% CI 1.98 to 2.66). There was statistical evidence that both of these were more strongly associated with diabetes than BMI (HR 1.80 (95% CI 1.59 to 2.04) (p for heterogeneity both <0.02). In phase 3 CaPS men, there was no strong evidence for differences in the strengths of association with incident diabetes between BMI and WC, waist: hip ratio (WHR) and WHtR (p all >0.49). Pooling estimates from BWHHS and CaPS revealed no evidence for differences in strengths of associations between BMI, WC, WHR or WHtR and CHD. All cause mortality was associated with WC, WHR and WHtR (HRs: 1.07 to 1.11) but not with BMI (HR 0.98) in pooled estimates (BWHHS, CaPS). Pooled odds ratios (Boyd Orr, MD) revealed no strong evidence for differences between anthropometric measures in their associations with arterial plaques. There was little evidence of differences between associations of BMI and other measures of adiposity with blood lipids, blood pressure, glucose or insulin in any of the cohorts.


    Measurements of central adiposity (WC, WHtR) were more strongly associated with diabetes than BMI in women but not men. There was no strong evidence for measurements of central adiposity being more strongly associated with CHD and stronger associations with all cause mortality were small. Further studies are required to determine whether measuring central adiposity in clinical practice would improve prediction of diabetes risk.