Evaluation of anthropometric measurements at birth in predicting birthweight less than 2000 g in African and Asian newborns: a meta-analysis.

Background: Particularly in developing countries, lower birthweight may be associated with higher neonatal mortality, and deliveries frequently take place at home where scales are not always available. Therefore, surrogate measurements for birthweight are necessary as a primary screening measure. The aim of this study was to determine whether newborn chest and arm circumferences can predict birthweight less than 2000 g.
Methods: The selection criteria were studies published in English that could provide all the true- and false-positive and true- and false-negative results with regard to the prediction of birthweight less than 2000 g by other anthropometric measurements among apparently healthy neonates. Ten bibliographic databases (e.g., PubMed) were searched and a bivariate meta-analysis was conducted with hierarchical summary receiver operating characteristic (ROC) curves. A total of 36,987 participants in 24 studies for chest circumference and 16,164 participants in 15 studies for arm circumference were included. The study regions were limited to Africa and Asia.
Results: For chest and arm circumferences (24 and 15 studies, respectively), pooled sensitivity (0.94 and 0.89, respectively) and specificity (0.94 and 0.96, respectively), and diagnostic odds ratios (263 and 174, respectively) were sufficiently high to allow good predictions. The diagnostic odds ratio for chest circumference was significantly higher than for arm circumference (P<0.001). The generalizability of the findings is to some extent guaranteed.
Conclusion: Newborn chest and arm circumferences may be useful predictors of birthweight less than 2000 g, with chest circumference possibly better.
(Copyright © 2014 Elsevier Masson SAS. All rights reserved.)