Concordance among measures of pregnancy outcome based on fetal size and duration of gestation.

Epidemiologic and clinical studies of pregnancy outcome often consider a variety of related, overlapping outcome measures. The overlap among these measures was analyzed using data from the Mount Sinai Hospital Perinatal Data Base, New York City, New York. A total of 52,621 births from 1986 through 1996 were included, with information on gender, ethnicity, birth weight, and gestational age assigned based on last menstrual period or early ultrasound. The authors considered very low birth weight (VLBW) (<1,500 g), low birth weight (LBW) (<2,500 g), degrees of preterm delivery (less than 32, 34, and 37 weeks' gestation), and small for gestational age (less than the 10th percentile of weight for gestational age) births. Infants at the extremes of gestational age (<32 or 34 weeks' gestation) were almost always LBW (97.6 and 91.7%, respectively), and those who were VLBW were almost always preterm (99.2%). However, only 69.2% of LBW infants were preterm, and 50.2% of preterm infants were LBW (kappa = 0.54). Only for VLBW and less than 32 weeks' gestation were both measures of overlap at least 70% (kappa = 0.98). The lack of concordance among measures suggests that multiple outcome measures be considered and that results from analyses using disparate measures not be compared directly.