Twin Birth Weight Chart Comparisons in Korea

Birth weight deviations from gestational age-specific standards affect neonatal outcomes. While singleton references are established, their applicability to twins is debated. With the increasing incidence of twin and preterm births, twin-specific growth standards remain limited in Korea. This study systematically analyzed twin growth trajectories, considering sex, mode of conception, and chorionicity, to establish accurate references.

This study analyzed data from the Korean Neonatal Network (2013-2020; ≤ 28 weeks, twins: 2,789; singletons: 6,171) and Samsung Medical Center (1995-2023; 22-37 weeks; twins: 4,201; singletons: 10,883). Birth weight distributions were compared between twins and singletons, and among subgroups defined by sex, mode of conception, and chorionicity. Smoothed percentile curves were constructed using a generalized additive model for location, scale, and shape, and validated by worm plots and bootstrap analysis.

At 32 weeks of gestation, the median birth weight of twins was 1,750 g (interquartile range, 1,600-1,950), and that of singletons was 2,050 g (1,850-2,250) (P < 0.001). Male twins had a higher median birth weight than females (1,800 g [1,650-1,980] vs. 1,700 g [1,550-1,880]). Twins conceived by in vitro fertilization had a lower median birth weight than those conceived spontaneously (2,173 g [1,940-2,390] vs. 2,340 g [2,100-2,560]; P < 0.001) in the unadjusted analysis; however, this difference diminished and was not statistically significant in subgroup analyses by gestational age, sex, or chorionicity. The growth model demonstrated good fit to the data across gestational ages.

Subgroup analyses by chorionicity, sex, and mode of conception did not reveal statistically significant differences. The twin-specific birth weight charts demonstrated a good model fit and internal consistency within the study population. These findings support their potential utility as reference standards for the clinical evaluation of twin growth in Korea. Further validation in larger and more diverse cohorts is warranted to confirm their generalizability.