The relevance of nutritional indices derived from comparison with growth standards to assess the risk of dying was evaluated in a 2-year prospective study in rural Senegal. An average of 3151 children aged 6-59 months were measured twice a year and followed up during the intervening 6-month periods. Children who survived and those who died during follow-up were found to differ more by anthropometric measures directly related to absolute muscle mass (viz. weight, height or arm circumference) than by nutritional indices obtained from comparison with growth standards (weight-for-age, weight-for-height and height-for-age). The findings could not be explained by a confounding effect of age. This brings into question the current approach used to identify high-risk children.
PIP: As part of a nutritional and mortality study ongoing since 1962 in rural Senegal, the relevance of nutritional indices derived from comparison with growth curves was analyzed to assess childrens' risk of dying. During a 2-year prospective study, an average of 3151 children aged 6-59 months were measured at 6 month intervals, a total of 12,605 semesters. 301 children who died during the following 6-months had anthropometric measures taken such as height, weight, mid-upper arm circumference (MUAC), head circumference, and triceps skinfold thickness. Univariate and bivariate multiple linear regression models were constructed and compared including anthropometric indicators such as weight, weight-for- age, weight-for-height, height, height-for-age, mid-upper-arm- circumference (MUAC), MUAC-for-age and age. Measures directly related to body size, such as weight, height and MUAC were found to be more closely related to child survival than any nutritional indices derived from them. Survival is related to a measure of body size which increases with age faster than MUAC but more slowly than weight, and is not due to a confounding effect of age. Mortality peaked at 18-29 months, corresponding to weaning. For a given muscle mass, children with a larger head had a greater risk of dying. Skinfold thickness did not improve assessment, perhaps because it is difficult to measure. Stunting had an independent effect on risk which could not be explained by muscle mass. This analysis suggests that screening schemes based on comparison with growth curves or weight gain are not likely to be predictive of risk of dying; arm circumference without correction for age or height is substantially better than classical nutritional indices.