Gestational weight gain and birth outcomes in U.S. and Tanzanian populations

Numerous studies have linked inadequate gestational weight gain to preterm birth and small-for-gestational age births (SGA) and excessive gestational weight gain to large-for-gestational age births, but gaps remain in our understanding of these relationships. Study 1 evaluated predictors and outcomes associated with area under the gestational weight gain curve. Compared to the middle quintile, the highest quintile of accumulated pound-days was associated with a decreased hazard of spontaneous preterm birth [Hazard Ratio (HR): 0.54, 95% Confidence Interval (CI) 0.30, 0.96)] and small-for-gestational age (SGA) births (HR 0.52, 95% CI 0.36, 0.76) overall and an increased hazard of LGA births among normal and underweight women (HR 3.78, 95% CI 1.76, 8.13). Study 2 evaluated potential predictors of high and low gestational weight gain using the INTERGROWTH 21st standards among pregnant women in Dar es Salaam, Tanzania and examined the relationship between gestational weight gain and preterm birth, perinatal death, small-for-gestational age births, and large-for-gestational age births. Weight gain below the 3rd percentile was associated with increased hazards of perinatal death [Hazard Ratio (HR): 1.94, 95% Confidence Interval (CI) 1.01, 3.73)] and small-for-gestational age (SGA) births (HR 1.51, 95% CI 1.02, 2.24), and a decreased hazard of preterm birth (HR 0.65, 95% CI 0.48, 0.88). Weight gain above the 97th percentile showed a trend toward an increased hazard of large-for-gestational age births (HR 1.77, 95% CI 0.96, 3.25). Study 3 assessed the accuracy of interpolated weight values obtained from regression models when compared to measured weight values at 28 (n=764) and 40 (n=358) weeks of gestation using data from the LIFECODES cohort at Brigham and Women’s Hospital. Mean differences in pounds between observed and estimated weights derived from the 28-week linear model (0.18 (SD 6.92), Bland Altman limits of agreement = -13.66–14.01) and 40-week linear models (-0.40 (5.43), limits of agreement -11.26–10.23), while weight values derived from the 40-week spline model underestimated observed weights (mean difference -2.50 (SD 6.91), limits of agreement 16.31–11.30 and -3.38 (SD 5.44), limits of agreement -14.26–7.49, respectively). Estimated weights derived from all models showed near perfect correlation with observed values (r=0.97–0.99)