Maternal anthropometry to predict small for gestational age: a meta-analysis

Simple, rapid, inexpensive, and reliable diagnostic tools for predicting small for gestational age newborns would be beneficial to reduce the rates of neonatal mortality and morbidity. This study was performed to evaluate the diagnostic performance of maternal anthropometric measurements as predictors of small for gestational age newborns.

Bivariate diagnostic meta-analysis was performed searching ten databases, e.g., PubMed (MEDLINE) (August 2014), including English language studies providing all numbers of true positive, false positive, true negative, and false negative results for prediction of small for gestational age newborns. This meta-analysis constructed hierarchical summary receiver operating characteristic curves. Study quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies.

Sufficient numbers of studies to evaluate maternal height, weight, body mass index, and weight gain during pregnancy (=39, 39, 61, and 99, respectively) were included. The quality of studies was relatively well controlled. None of the maternal parameters showed sufficiently high sensitivity (=0.37-0.73), specificity (=0.35-0.72), or diagnostic odds ratios (=1-2). The informational value and diagnostic accuracy were categorized as small (and rarely important) based on the positive likelihood ratio (=1-2) and negative likelihood ratio (=0.5-1) and low based on the area under the curve (≤0.7), respectively. No studies evaluating other anthropometric measurements were subjected to meta-analysis.

This meta-analysis did not provide evidence for the usefulness of maternal anthropometric measurements in predicting small for gestational age newborns.