• Augustine Okwudili Maduekwe National Postgrduate Medical College of Nigeria (NPMCN)


In developing countries including Nigeria, about 60-80% of deliveries take place in rural communities, either at home
or in primary health care centres where there may be no facilities to weigh the babies at birth. About 40- 50% of
neonatal mortality is attributed to low birth weight in most of these countries. Early identification and classification of
babies as low birth weight is thus a crucial step in recognition of the at risk newborn infants for intervention and
referral to specialised centres, which is vital in preventing neonatal morbidity and mortality.
The objectives were to determine the validity of anthropometric parameters such as chest circumference,
occipitofrontal circumference, mid arm circumference and foot length as proxies to identify low birth weight babies.
A cross sectional study was carried out involving 762 neonates delivered at National Hospital, Maitama and Asokoro
District Hospitals between May and December 2009. Birth weight, chest circumference, occipitofrontal circumference,
mid arm circumference and foot length were measured within 24 hours of delivery using standard techniques. Each
parameter was correlated with birth weight. A regression model was fitted for birth weight and each anthropometric
parameter. The optimum cut-off point for the anthropometric parameters was based on the closest predicted value for
the anthropometric parameter on linear regression. Non parametric Receiver Operating Characteristics curve analysis
with 95% confidence interval (95% CI) of area under the curve was used for validation, sensitivity and specificity. The
parameter with the least misclassification and best combination of sensitivity and specificity was chosen to identify
low birth weight babies.
Out of 762 newborn babies studied, 393 (51.57%) were males and 369 (48.43 %) were females. Ninety nine (99)
(13%) newborns were low birth weight. The mean birth weight was 3141± 603 grams. Using ROC- AUC analysis,
chest circumference with the best correlation coefficient r =0.825 (AUC = 0.98, 95 CI 0.97 – 0.99) was identified as
the best possible proxy to identify low birth weight babies at ChC ≤ 28.5 cm for birth weight of < 2500 grams and
ChC ≤ 26.0 cm for birth weight of ≤ 2000 grams. Occipitofrontal circumference with the next best correlation
coefficient r of 0.771(AUC 0.95, 95% CI 0.90- 0.99) had cut-off points of OFC ≤ 31.5cm for < 2500 grams and OFC
≤ 29.0 for ≤ 2000 grams. MAC had r of 0.687 and FL had r of 0.661 and consequently had lower AUC values, and
poor combination of sensitivity and specificity.
It is therefore recommended that chest circumference measurement be used to identify this group of at risk babies, in
rural areas where weighing scales may not be available.