Table 4.

Three measures of the additive interactions between prepregnancy overweight/obesity and passive smoking for macrosomia and LGA

	RERI (95% CI)	AP (95% CI)	S (95% CI)
Macrosomia as the outcome
Model 1	0.54 (0.11–0.97)	0.21 (0.05–0.36)	1.50 (1.03–2.19)
Model 2	0.52 (0.05–0.98)	0.20 (0.03–0.35)	1.44 (1.00–2.06)
Model 3	0.44 (0.00–0.89)	0.17 (0.01–0.34)	1.40 (0.95–2.07)
Model 4	0.44 (0.00–0.88)	0.18 (0.01–0.34)	1.41 (0.96–2.08)
LGA as the outcome
Model 1	0.49 (0.05–0.93)	0.19 (0.03–0.35)	1.44 (0.99–2.01)
Model 2	0.42 (-0.02 to 0.85)	0.16 (-0.01 to 0.33)	1.39 (0.95–2.03)
Model 3	0.35 (-0.07 to 0.76)	0.15 (-0.03 to 0.33)	1.36 (0.90–2.06)
Model 4	0.35 (-0.07 to 0.76)	0.15 (-0.03 to 0.33)	1.36 (0.90–2.06)

Model 1: univariable analysis. Model 2: multivariable analysis, adjusted for age, height, parity, education >12 years, gestational age at delivery (only for macrosomia), gender (only for macrosomia), weight gain from prepregnancy to GCT, insulin treatment, alcohol drinker during pregnancy, multiple pregnancies, and systolic blood pressure at first antenatal care visit. Model 3: further adjusted for GDM, in addition to the variables listed in model 2. Model 4: further adjusted for GCT value, in addition to the variables listed in model 2. GDM, gestational diabetes mellitus; CI, confidence interval; LGA, large for gestational age; RERI, relative excess risk due to interaction; AP, attributable proportion due to interaction; S, synergy index. RERI > 0, AP > 0, or S > 1 suggest significant additive interaction.