Customized growth charts for twin gestations to optimize identification of small-for-gestational age fetuses at risk of intrauterine fetal death

Objective Customized growth charts developed for singleton pregnancies have been shown to be more effective than population-based ones at identifying small-for-gestational age (SGA) fetuses at risk for intrauterine fetal death (IUFD). We sought to compare the association between SGA and IUFD in twins using customized growth charts designed for twin gestations compared to those designed for singletons. Methods This was a retrospective cohort study using a database including singleton and twin pregnancies undergoing ultrasound examination between 16 and 20 weeks' gestation. After excluding preterm births < 34 weeks, congenital anomalies and stillbirths, we identified 51 150 singleton births. Coefficients for significant physiological and pathological variables affecting birth weight for singletons were derived using backward stepwise multiple regression. The same process was repeated for twin births (1608 pairs), also adjusting for chorionicity. Customized growth charts for each pregnancy were derived using these two regression models for optimal birth weight at term and a proportionality equation. The association between SGA < 10^th percentile, defined using the twin and singleton-customized charts, and IUFD were compared. Statistical analysis, including calculation of adjusted odds ratios (OR) for IUFD and screening accuracy using each chart, was performed. Results The derived coefficients for optimal birth weight for twins were different from those for singletons, with lower constants and root mean square error (3422 and 288.9, respectively, in twins vs 3543 and 416 in singletons). Among 3786 twin infants, IUFD was seen in 123 (3.2%). The numbers of fetuses identified as SGA were 575 (15.2%) and 504 (13.3%) by the singleton and twin charts, respectively. Fetuses classified as SGA by the twin-specific customized charts were at a significantly increased risk for IUFD (adjusted OR, 2.3 (95%CI, 1.4-3.5)), whereas those classified as SGA by the singleton-customized charts were not (adjusted OR, 1.2 (95% CI, 0.7-2.0)). Conclusion Customized charts designed specifically for twins are more effective at identifying twin pregnancies at risk for IUFD than are those derived using singleton birth data.