Surfactant proteins A and B (SP-A and SP-B) were measured in human amniotic fluid by ELISA and correlated with lecithin to sphingomyelin ratio (L/S), phosphatidylglycerol (PG), and perinatal outcome. Amniotic fluid SP-A, SP-B, and L/S increased with advancing gestation. SP-A was detected at 19 wk gestation and increased dramatically in the 3rd trimester of pregnancy. SP-B was first detectable at 31 wk gestation and increased significantly to term. SP-A was a more specific predictor of non-respiratory distress syndrome (RDS) than L/S or SP-B; however, the sensitivity of SP-A in predicting RDS was less than L/S< 2.0 (26.3 versus 82.3%, respectively). In 209 pregnancies assessed within 48 h of delivery, the sensitivity of SP-B in predicting RDS (nondetectable SP-B) was comparable to the L/S, however, SP-B= 0 was frequently observed in mature infants, limiting its specificity for prediction of RDS. The greatest sensitivity and specificity were achieved with the measurement of L/S< 2.0 and negative PG, which correctly predicted 100% of the infants with RDS and 94% of those who did not develop the disorder. Measurement of SP-A or SP-B did not improve the prediction of RDS. SP-A, SP-B, and L/S were not affected by infant sex, Apgar score, rupture of membranes, size for gestational age, maternal diabetes, hypertension, or exposure to medications. SP-A, SP-B, and L/S were significantly elevated in amniotic fluid from black mothers. SP-A was significantly elevated in amniotic fluid from mothers who smoked during pregnancy. Although amniotic fluid SP-A and SP-B concentrations increased with advancing gestation, the predictive value of the L/S and PG for RDS was not enhanced by the additional measurement of SP-A and/or SP-B.