Intended for healthcare professionals


Alcohol consumption and mortality from all causes, coronary heart disease, and stroke: results from a prospective cohort study of Scottish men with 21 years of follow up

BMJ 1999; 318 doi: (Published 26 June 1999) Cite this as: BMJ 1999;318:1725
  1. Carole L Hart, research fellowa,
  2. George Davey Smith, professor of clinical epidemiology (zetkin{at},
  3. David J Hole, principal epidemiologistc,
  4. Victor M Hawthorne, professor of epidemiologyd
  1. aDepartment of Public Health, University of Glasgow, Glasgow G12 8RZ
  2. bDepartment of Social Medicine, University of Bristol, Bristol BS8 2PR
  3. cWest of Scotland Cancer Surveillance Unit, University of Glasgow, Glasgow G12 8RZ
  4. dUniversity of Michigan, School of Public Health, Department of Epidemiology, Ann Arbor, MI 48109, USA
  1. Correspondence to: Professor Davey Smith
  • Accepted 18 February 1999


Objectives: To relate alcohol consumption to mortality.

Design: Prospective cohort study.

Setting: 27 workplaces in the west of Scotland.

Participants: 5766 men aged 35-64 when screened in 1970-3 who answered questions on their usual weekly alcohol consumption.

Main outcome measures: Mortality from all causes, coronary heart disease, stroke, and alcohol related causes over 21 years of follow up related to units of alcohol consumed per week.

Results: Risk for all cause mortality was similar for non-drinkers and men drinking up to 14 units a week. Mortality risk then showed a graded association with alcohol consumption (relative rate compared with non-drinkers 1.34 (95% confidence interval 1.14 to 1.58) for 15-21 units a week, 1.49 (1.27 to 1.75) for 22-34 units, 1.74 (1.47 to 2.06) for 35 or more units). Adjustment for risk factors attenuated the increased relative risks, but they remained significantly above 1 for men drinking 22 or more units a week. There was no strong relation between alcohol consumption and mortality from coronary heart disease after adjustment. A strong positive relation was seen between alcohol consumption and risk of mortality from stroke, with men drinking 35 or more units having double the risk of non-drinkers, even after adjustment.

Conclusions: The overall association between alcohol consumption and mortality is unfavourable for men drinking over 22 units a week, and there is no clear evidence of any protective effect for men drinking less than this.


  • Funding NHS Management Executive (Cardiovascular Disease and Stroke Research and Development Initiative), Chest Heart and Stroke Scotland, and the Stroke Association.

  • Competing interests None declared.

View Full Text