Use of a simple anthropometric measure to predict birth weight

Diamond, I., McDonald, J. and Guidotti, R. (1993) Use of a simple anthropometric measure to predict birth weight. Bulletin of the World Health Organization, 71 (2), 157-163.

Abstract

The study was undertaken in 22 centers throughout the world to collect data on a consecutive sample of 400 births. 3 measurements were made for each baby: birth weight, mid-arm circumference, and chest circumference. In addition, the baby's sex and gestational age at birth were recorded. The main problem with data quality was a tendency for weights to be recorded in round hundred grams and circumferences in whole centimeters. The primary objective was to identify cut-off points below which a baby is diagnosed to be at risk for conditions associated with low birth weight. Centers in South Asia, such as Delhi and Chandigarh, had the lowest average birth weight and anthropometric measures, whereas those in Europe, such as St. Petersburg (Russia), Szeged (Hungary), and Yerevan (Armenia) has some of the highest. In 18 of the 22 centers, the correlations between birth weight and chest circumference were greater than those for arm circumference. Regression analyses demonstrated that the best model in each center was birth weight predicted by chest circumference. However, a different regression equation had to be estimated for each center. The estimated regression coefficients varied between the extremes of Islamabad and Chandigarh, where an increase of 1 cm in chest circumference predicted birth weight increases of 260 and 156 gm, respectively. For practical use in developing countries, cut-off points for chest circumference and end-points for birth weight need to be defined for the prediction of low birth weight. Therefore, the standard WHO end-point of 2500 gm was adopted, and babies below this were defined as having low birth weight. Cut-off points of 29 and 30 cm are proposed. Babies with a chest circumference <29 cm would be diagnosed as highly at risk, and they should be referred to a health center immediately. Those with a chest circumference of 29-30 cm would be diagnosed as at risk, and their progress should be monitored carefully.

This record has no associated files available for download.

Contributors

Download statistics