Rate of Second and Third Trimester Weight Gain and Preterm Delivery Among Underweight and Normal Weight Women

Objectives Low gestational weight gain (GWG) in the second and third trimesters has been associated with increased risk of preterm delivery (PTD) among women with a body mass index (BMI) < 25 mg/m(2). However, few studies have examined whether this association differs by the assumptions made for first trimester gain or by the reason for PTD. Methods We examined singleton pregnancies during 2000-2008 among women with a BMI < 25 kg/m(2) who delivered a live-birth ≥28 weeks gestation (n = 12,526). Women received care within one integrated health care delivery system and began prenatal care ≤13 weeks. Using antenatal weights measured during clinic visits, we interpolated GWG at 13 weeks gestation then estimated rate of GWG (GWGrate) during the second and third trimesters of pregnancy. We also estimated GWGrate using the common assumption of a 2-kg gain for all women by 13 weeks. We examined the covariate-adjusted association between quartiles of GWGrate and PTD (28-36 weeks gestation) using logistic regression. We also examined associations by reason for PTD [premature rupture of membranes (PROM), spontaneous labor, or medically indicated]. Results Mean GWGrate did not differ among term and preterm pregnancies regardless of interpolated or assumed GWG at 13 weeks. However, only with GWGrate estimated from interpolated GWG at 13 weeks, we observed a U-shaped relationship where odds of PTD increased with GWGrate in the lowest (OR 1.36, 95 % CI 1.10, 1.69) or highest quartile (OR 1.49, 95 % CI 1.20, 1.85) compared to GWGrate within the second quartile. Further stratifying by reason, GWGrate in the lowest quartile was positively associated with spontaneous PTD while GWGrate in the highest quartile was positively associated with PROM and medically indicated PTD. Conclusions Accurate estimates of first trimester GWG are needed. Common assumptions applied to all pregnancies may obscure the association between GWGrate and PTD. Further research is needed to fully understand whether these associations are causal or related to common antecedents.