The univariate and multivariate logistic regression analyses for
outcome in relation to exposure to ETS are presented in Table
3. The crude odds ratios were significant for pre-term birth
and small-for-gestation babies. The multiple logistic regression
analyses, including all the cofactors and confounders that
were significant in the antenatal period, showed that exposure
to ETS during pregnancy was significantly associated with a
higher risk of small-for-gestation babies (OR 2.10; 95% CI: 1.27-
3.48). The linear regression modelling for birth weight and
APGAR at one minute did not show any significant association
with exposure to ETS (adjusted r2 = 0.055 and 0.011 respectively
indicating a poor fit; variables entered in the model
were age, education, occupation, birth order, number of live
issues and anaemia).