The objective of this secondary analysis was to identify maternal characteristics that modified the effect of maternal supplements on newborn size. Participants included 1465 maternal–newborn dyads in Guatemala, India, and Pakistan. Supplementation commenced before conception (Arm 1) or late 1st trimester (Arm 2); Arm 3 received usual care. Characteristics included body mass index (BMI), stature, anemia, age, education, socio-economic status (SES), parity, and newborn sex. Newborn outcomes were z-scores for length (LAZ), weight (WAZ), and weight to length ratio-for-age (WLRAZ). Mixed-effect regression models included treatment arm, effect modifier, and arm * effect modifier interaction as predictors, controlling for site, characteristics, and sex. Parity (para-0 vs. para ≥1), anemia (anemia/no anemia), and sex were significant effect modifiers. Effect size (95% CI) for Arm 1 vs. 3 was larger for para-0 vs. ≥1 for all outcomes (LAZ 0.56 (0.28, 0.84, p < 0.001); WAZ 0.45 (0.20, 0.07, p < 0.001); WLRAZ 0.52 (0.17, 0.88, p < 0.01) but only length for Arm 2 vs. 3. Corresponding effects for para ≥1 were >0.02. Arm 3 z-scores were all very low for para-0, but not para ≥1. Para-0 and anemia effect sizes for Arm 1 were > Arm 2 for WAZ and WLRAZ, but not LAZ. Arm 1 and 2 had higher WAZ for newborn boys vs. girls. Maternal nulliparity and anemia were associated with impaired fetal growth that was substantially improved by nutrition intervention, especially when commenced prior to conception.