Associations of morbidity in the underweight

There is little research on the demographic characteristics and morbidity of people categorized as 'underweight' from their body mass index (BMI) although they have often been shown to have greater mortality. This uncertainty makes it difficult to determine whether to include or exclude these individuals when estimating the health and mortality impacts of BMI. This project compares the demographic characteristics and morbidity patterns of the underweight with those of acceptable weight and the overweight.

Data on 10 243 community-living residents from the Health Survey for England (2003) were used. Logistic regression models were constructed to compare demographic, biochemical and anthropometric factors in the underweight (BMI<18.5) with those classified as acceptable weight (BMI 18.5-24.9) or overweight (BMI 25.0-29.9).

Univariate analyses found, when compared with other BMI categories, underweight individuals were significantly younger, more likely to smoke, alcohol abstainers, inactive, poorer and were less likely to be ethnically white (all P<0.001). U-shaped relationships between BMI and activities of daily living, respiratory disease, physical activity and mental health variables were seen. In multivariate analysis, the fewest number of significant differences in demographic and morbidity factors were between the underweight and those of acceptable weight.

We recognize that these are cross-sectional data and exclude individuals in institutional settings, but these findings are important. Overall, we could not conclude that the underweight were less healthy than individuals in the other BMI categories. We cannot therefore recommend that the underweight should be excluded from analyses that examine the effects of obesity on mortality.