The ninety-fifth percentile for growth discordance predicts complications of twin pregnancy

tive: Our purpose was to define twin growth discordance on the basis of perinatal outcome. Study Design: Twins delivered at >23 weeksʹ gestation between 1995 and 2000 were identified by use of our computerized perinatal database. Birth weight (BW), chorionicity, and the following outcomes were recorded: cesarean delivery for nonreassuring fetal status (NRFS), umbilical artery (UA) pH, 5-minute Apgar score, neonatal intensive care unit (NICU) admission, and stillbirth. Discordance was calculated as 100 ×(Larger BW − Smaller BW)/Larger BW. The distribution of discordance values for the study population was determined. Outcomes were compared for twins with and without discordance at thresholds defined by population percentile. Results: The mean discordance for 346 twins was 11.4% ± 10.5%. The 75th percentile was 16% discordance, the 90th was 23%, and the 95th was 31%. Discordance was independently associated with outcomes. The 95th percentile was the threshold most predictive of cesarean section NRFS, UA pH <7.1, 5-minute Apgar score <7, and NICU admission. Conclusion: Defining pathologic discordance as the 95th percentile, or >30% difference, enhances its predictive value. (Am J Obstet Gynecol 2002;187:667-71.)