J Epidemiol Community Health. 2003 Feb; 57(2): 130-3.

Relation between body mass index and mortality in an unusually slim cohort.

Thorogood M, Appleby PN, Key TJ, Mann J.

Health Promotion Research Unit, Department of Public Health and Policy, London School of Hygiene and Tropical Medicine, UK. margaret.thorogood@lshtm.ac.uk

STUDY OBJECTIVE: To investigate the relation between body mass index and mortality in an unusually slim cohort. DESIGN: Prospective cohort study. SETTING: United Kingdom. PARTICIPANTS: About 11,000 non-meat eaters and their meat eating friends and relatives, with a median age of 33 years were recruited between 1980 and 1984. More than 20% of the cohort had a self reported BMI below 20 kg/m(2) at recruitment. There were 195,000 person years of observation after a mean of 18.0 years of follow up. MAIN RESULTS: The characteristics of participants with a BMI below 18 kg/m(2) were favourable to a lower risk of cardiovascular disease. Nevertheless, these participants had an increased all cause death rate ratio (2.07, 95% CI 1.58 to 2.70) in comparison with participants who had a BMI between 20 and 22 kg/m(2). The death rate ratio for the slimmest category was also significantly increased for circulatory diseases (including ischaemic heart disease and cerebrovascular disease), respiratory diseases, and all other causes combined excluding all malignant neoplasms. This finding was consistent across a range of subgroups. CONCLUSIONS: Lean men and women (BMI <18 kg/m(2)) experience increased all cause mortality compared with those with a BMI between 20 and 22 kg/m(2). This pattern is not seen for cancer mortality, but is found for cardiovascular and respiratory diseases. It is important that public health messages regarding healthy eating are aimed at maintaining a healthy body weight rather than just "losing weight".