Fatness and muscularity as risk indicators of child mortality in rural Congo.

Abstract

To examine the relationship of anthropometrical indicators of fatness and muscularity with mortality in children in a rural African community. A prospective cohort study was carried out in the rural health zone of Bwamanda, Northern Congo using a random cluster sample of 5167 children, aged 0-5 years. Short- and long-term mortality rates, being deaths within 3 months and deaths in 3-month periods observed 3-30 months after enrolment. Rates of all cause mortality and of mortality from kwashiorkor or marasmus, by level of baseline fatness and muscularity. Indicators of fatness and muscularity were obtained by correcting anthropometric arm fat and arm muscle areas for age, sex, weight and height. The relationship of both the fatness and muscularity scores with short-term mortality was marked by a clear threshold (-0.5 SDS) below which there was a significant rise in mortality from all causes as well as from kwashiorkor and marasmus. These excess mortalities were also found in normal weight children. Fatness and muscularity scores remained significant determining factors of short-term mortality in a multiple logistic regression analysis with sex, age, season and weight-for-age. A ROC curve analysis showed that fat and muscularity scores had better predictive abilities than weight-for-age. Low fat status had a bad prognosis on the long-term in underweight children. Measures of current nutritional status should not be based on weight indices alone. Objective and/or clinical evaluation of fat and muscle status (also in normal weight children) should be added in order to detect a higher proportion of malnourished children and to more accurately evaluate mortality risk.