The association between birthweight, sociodemographic variables and maternal anthropometry in an urban sample from Dhaka, Bangladesh.

The relationship between birthweight, sociodemographic variables and maternal anthropometry was examined in a sample from an inner urban area of Dhaka, Bangladesh. About 21% of babies were of low birthweight (LBW) using the World Health Organization cut-off of < 2500 g. LBW was more common in younger (< 20 years) and older (> 30 years) mothers, the low-income group and those with little or no education. The mean birthweights of the higher-educated, higher-income group and male children were on average 290, 260 and 120 g, respectively, higher than uneducated, lower-income groups and female children. The best cut-offs for detecting LBW and normal-weight infants was maternal weight of 50 kg (odds ratio = 4.6), maternal arm circumference of 23 cm (odds ratio = 5.0) and body mass index of 20.5 (odds ratio = 6.5). The sensitivity and specificity were best for maternal weight (69% and 68%, respectively). Logistic regression analyses show that mothers' weight at term was the best single predictor of LBW (31%), while maternal weight along with age, educational level and income group correctly predicted just over 35% of LBW. Regression analyses also confirmed that mothers' weight at term was the best predictor of birthweight, with a correlation coefficient of 0.49.