Adequacy of child anthropometric indicators for measuring nutritional stress at population level: a study from Niakhar, Senegal.

OBJECTIVE To compare the responsiveness of different anthropometric indicators for measuring nutritional stress among children in developing countries. DESIGN Growth was studied within 6-month intervals in a rural Senegalese community during one dry and two rainy (hungry) seasons. Responsiveness was defined as the change divided by the standard deviation of each anthropometric indicator. Contrast was defined as the difference in responsiveness between dry and rainy seasons. SETTING The study was conducted in Niakhar, a rural area of Senegal under demographic surveillance, with contrasted food and morbidity situations between rainy and dry seasons. SUBJECTS Some 5000 children under 5 years of age were monitored at 6-month intervals in 1983–1984. The present analysis was carried out on a sub-sample of children aged 6–23 months with complete measures, totalling 2803 children-intervals. RESULTS In both univariate and multivariate analysis, mid-upper arm circumference was found to be more responsive to nutritional stress than the commonly used weight-for-height Z-score (contrast = -0.64 for mid-upper arm circumference v. -0.53 for weight-for-height Z-score). Other discriminant indicators were: muscle circumference, weight-for-height, BMI and triceps skinfold. Height, head circumference and subscapular skinfold had no discriminating power for measuring the net effect of nutritional stress during the rainy season. CONCLUSIONS The use of mid-upper arm circumference for assessing nutritional stress in community surveys should be considered and preferred to other nutritional indicators. Strict standardization procedures for measuring mid-upper arm circumference are required for optimal use.