Is cardiothoracic ratio in healthy middle aged men an independent predictor of coronary heart disease mortality? Whitehall study 25 year follow up

BMJ 1998; 316 doi: (Published 02 May 1998) Cite this as: BMJ 1998;316:1353
  1. Harry Hemingway, senior lecturer in epidemiology (h.hemingway{at},
  2. Martin Shipley, senior lecturer in medical statisticsb,
  3. David Christie, professor of environmental and occupational healthc,
  4. Michael Marmot, professor of epidemiology and public healthb
  1. a Department of Research and Development, Kensington and Chelsea and Westminster Health Authority, London W2 6LX
  2. b International Centre for Health and Society, Department of Epidemiology and Public Health, University College London Medical School, London WC1E 6BT
  3. c Department of Environmental and Occupational Health, Box 13, Hunter Building, University of Newcastle, Newcastle, New South Wales 2308, Australia
  1. Correspondence to: Harry Hemingway, International Centre for Health and Society, Department of Epidemiology and Public Health, University College London Medical School, London WC1E 6BT
  • Accepted 24 June 1997

Aetiological studies of myocardial ischaemia have tended to concentrate on factors which influence atherothrombotic processes in the coronary arteries rather than myocardial pathophysiology.1 The commonest clinical measure of heart size—cardiothoracic ratio—was included in the original Whitehall study of healthy middle aged civil servants. Cardiothoracic ratio is associated with left ventricular mass2 and left ventricular systolic function; since left ventricular mass determined by echocardiography has been shown to predict coronary heart disease in elderly people,1 we hypothesised that increased cardiothoracic ratio would independently predict mortality from coronary heart disease. Unlike previous studies3 we did not include mortality from stroke since it may be related to heart size through different pathophysiological mechanisms.

Adjusted hazard ratios (95% confidence intervals) for the effect of cardiothoracic ratio on all cause and coronary heart disease mortality

View this table:

Subjects, methods, and results

We studied the 1203 male British civil servants aged 40-69 years who participated in the original Whitehall study and were randomly selected (by random number tables) for measurement of cardiothoracic ratio from 100 mm chest radiographs. The rate ratio for all cause mortality among those in the random sample compared with those not in the sample was 1.01 (95% confidence interval 0.93 to 1.11) making a serious selection bias unlikely. Details of the standardised methods of risk factor, electrocardiographic and radiographic measurements and their quality control have been reported. 4 5 Cardiothoracic ratio was calculated as the ratio of the maximal transverse diameter of the cardiac silhouette to the distance between the internal margins of the ribs at the level of the right hemidiaphragm. In all, 1191 (99%) of the subjects were flagged at the NHS Central Registry and there were 534 deaths over 25 years, 196 of which were due to coronary heart disease (ICD-8 codes 410-414). Adjusted mortality hazard ratios and their confidence intervals were estimated by Cox's proportional hazards regression models.

The table shows the extent to which cardiothoracic ratio affects the risk of death from all and coronary causes independently of potential confounders. After age, blood pressure, heart rate, total cholesterol concentration, smoking, prevalent symptoms of coronary heart disease, and electrocardiographic evidence of ischaemia were adjusted for, men with a cardiothoracic ratio in the highest fifth of the distribution had a hazard ratio of 1.65 (95% confidence interval 1.01 to 2.70) for coronary heart disease mortality compared with men with a cardiothoracic ratio in the lowest fifth. When men with a ratio ≥0.5 were excluded, the top fifth (≥0.47 and <0.5) was associated with an increased risk of coronary death of 1.67 (0.99 to 2.82) after age and blood pressure were adjusted for.


The cardiothoracic ratio in a healthy middle aged population predicted coronary mortality over 25 years independent of blood pressure and other risk factors. A ratio of ≥0.5 has by convention been defined as a threshold of pathological enlargement. In our healthy population of civil servants a ratio of 0.47 to <0.5 was associated with increased risk of death from coronary heart disease, questioning this convention. The results of recently established population based echocardiographic studies are therefore awaited to establish the relative contribution of left ventricular mass and left ventricular systolic dysfunction in predicting coronary heart disease among healthy middle aged subjects. Until then the Whitehall study offers the advantage of a prolonged follow up.

Does lowering cardiothoracic ratio reduce the risk of coronary heart disease? Among hypertensive patients, drug treatment and exercise may reduce cardiothoracic ratio. However, further studies are required to investigate whether such effects lead to a reduction in subsequent coronary heart disease events and therefore constitute a worthwhile therapeutic goal. In the meantime the prognostic information provided by the cardiothoracic ratio should be considered in risk stratification of healthy middle aged men.


Contributors: HH posed the research question and wrot the first draft of the paper. All the authors commented on drafts of the paper which were incorporated by HH. MS carried out all the statistical analyses. DC collected the radiographic data and his doctoral thesis formed the basis of this paper. MM is codirector of the original Whitehall study. HH is the guarantor.

Funding: MM is supported by an Medical Research Council research professorship. MS is supported by the British Heart Foundation.

Conflict of interest: None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract