While Apgar scores provide a valid prediction of mortality in term infants (primarily reflecting asphyxia), intervening variables in premature newborns complicate interpretation. Physiologic states normal to preterm infants (such as, decreased muscle tone) can depress scores but may not influence survival significantly. Therefore the relationship between Apgar scores and survival in term and preterm infants differs. Because of the paucity of studies on preterm infants, we tested Apgar scores, as well as birthweight and gestational age, as outcome predictors in 748 low-birthweight infants (500-1800 gm). Our purpose was to assess the relationship between 1- and 5-minute Apgar scores and survival, and to evaluate all combinations of the four variables as outcome predictors. Univariate analysis showed a significant relationship between each of the four variables and survival; however, no single variable accounted for more than 32% of the variance in outcome, thus no single factor could be invoked as the major determinant of survival. Logistic regression analyses demonstrated the interrelationships of the four variables to survival. While both Apgar scores were related to survival, independent of the effects of birthweight and gestational age, they were slightly less predictive than either of these variables alone. However, when 1- and 5-minute Apgar scores were combined with gestational age, the predictive value was slightly better than any of the four variables alone or in other possible combinations.