Molecular studies were performed on 101 cases of confined placental mosaicism (CPM) involving autosomal trisomy. The origin of the trisomic cell line was determined in 54 cases (from 51 pregnancies), 47 of which were also analyzed for the presence of uniparental disomy (UPD) in the disomic cell line. An additional 47 cases were analyzed for parental origin in the disomic cell line only. A somatic (postmeiotic) origin of the trisomy was observed in 22 cases and included the majority of cases with CPM for trisomy 2, 7, 8, 10, and 12. Most cases of CPM involving trisomy 9, 16, and 22 were determined to be meiotic. Fetal maternal UPD was found in 17 of 94 informative CPM cases, involving trisomy 2 (1 case), 7 (1 case), 16 (13 cases), and 22 (2 cases). The placental trisomy was of meiotic origin in all 17 cases associated with fetal UPD (P = .00005). A meiotic origin also correlated with the levels of trisomy in cultured chorionic villi samples (CVS) (P = .0002) and trophoblast (P = .00005). Abnormal pregnancy outcome (usually IUGR) correlated with meiotic origin (P = .0003), the presence of fetal UPD (P = 4 x 10(-7)), and the level of trisomy in trophoblast (P = 3 x 10(-7)) but not with the level of trisomy in CVS or term chorion. The good fit of somatic errors with the expected results could have been observed only if few true meiotic errors were misclassified by these methods as a somatic error. These data indicate that molecular determination of origin is a useful predictor of pregnancy outcome, whereas the level of trisomy observed in cultured CVS is not. In addition, UPD for some chromosomes may affect prenatal, but not postnatal, development, possibly indicating that imprinting effects for these chromosomes are confined to placental tissues.