Anthropometric measurements as surrogate for identifying low birth weight: comparison with gold standard of measuring weight among neonates at a tertiary hospital in Tanzania

Reducing neonatal mortality is crucial for achieving Sustainable Development Goal 3.2. In Tanzania and other low- and middle-income countries, many babies are born at home or in primary health care settings without access to weighing scales. Other anthropometric measurements apart from weight may serve as effective surrogates for identifying low birth weight (LBW) infants and prioritizing them for care. This study aimed to assess the sensitivity, specificity, and accuracy of four anthropometric measurements chest circumference (CHC), occipitofrontal circumference (OFC), foot length (FL), and mid-upper arm circumference (MUAC) and determine the optimal cut-off values for detecting LBW infants. A hospital-based cross-sectional study was conducted in the neonatal unit at Muhimbili National Hospital (MNH). Newborns admitted within 24 h of birth were enrolled. A total of 471 newborns had CHC, OFC, FL, MUAC, and birth weight (BW) measured. Pearson’s correlation coefficient assessed the relationship between these measurements and birth weight. Receiver Operating Characteristics (ROC) analysis was used to determine optimal cut-off points. Of 471 newborns, 52.2% were female. The mean birth weight was 2639 g (SD = 851), with 43.1% (203) classified as LBW. A strong positive correlation of 0.90 was found between CHC and BW (p < 0.01), followed by OFC (0.82, p < 0.01). ROC analysis showed CHC (AUC = 0.97, 95% CI 0.97–0.98) and OFC (AUC = 0.94, 95% CI 0.92–0.96) as the most reliable indicators for detecting LBW. The optimal cut-off values were ≤ 29.45 cm for CHC and ≤ 32.9 cm for OFC. Our study demonstrated that CHC with a cutoff ≤29.45 cm and OFC with a cutoff ≤ 32.9 cm has higher sensitivity and specificity for predicting low birth weight among neonates compared to FL or MUAC, indicating greater accuracy in detecting LBW in Tanzania. Our findings apply specifically to LBW newborns; we did not classify IUGR or SGA separately, and further studies are required to validate these cut-offs in those populations.