Relation Between Blood Glucose and Coronary Mortality Over 33 Years in the Whitehall Study

  1. Michael G. Marmot, FRCP
  1. From the Department of Epidemiology and Public Health, University College London, London, U.K
  1. Address correspondence and reprint requests to Eric J. Brunner, PhD, 1-19 Torrington Pl., London WC1E 6BT, England. E-mail: e.brunner{at}


OBJECTIVE—Glucose intolerance is a risk factor for coronary disease, but there is uncertainty about the shape of the dose-response relationship between glucose level and risk of coronary mortality. We examined the prospective relation of 2-h postload blood glucose (2hBG) with coronary and other major causes of mortality over 33 years.

RESEARCH DESIGN AND METHODS—A 50-g oral glucose tolerance test (OGTT) was performed at baseline (1967–1969) in 17,869 male civil servants aged 40–64 years.

RESULTS—There were 3,561 coronary deaths during 451,787 person-years of observation. All-cause, cardiovascular, and respiratory mortality were elevated among participants with glucose intolerance. The hazard of coronary mortality rose from 2hBG = 4.6 mmol/l (83 mg/dl [95% CI 4.2–5.3]). The dose-response relation was best fitted by a single slope above this level, with no evidence of nonlinearity, compared with Cox models using other threshold levels, and those containing log 2hBG terms. There was no evidence for a dose-response relationship below 2hBG = 4.6 mmol/l. Between this level and 11.1 mmol/l (200 mg/dl), the age-adjusted hazard ratio was 3.62 (95% CI 2.3–5.6). The graded relationship was attenuated by 45% after adjustment for baseline coronary heart disease (CHD), BMI, systolic blood pressure, blood cholesterol, smoking, physical activity, lung function, and employment grade.

CONCLUSIONS—A threshold model with linear slope best described the dose-response relationship between postload blood glucose and CHD mortality risk.

  • Received July 27, 2005.
  • Accepted October 3, 2005.
| Table of Contents