Context: Because of methodological limitations, the results of the few prospective studies assessing long-term cognitive effects of prenatal cocaine exposure are inconsistent. Objective: To assess effects of prenatal cocaine exposure and quality of caregiving environment on 4-year cognitive outcomes. Design: Longitudinal, prospective, masked comparison cohort study from birth (September 1994-June 1996) to 4 years. Setting: Research laboratory of a US urban county teaching hospital. Participants: A total of 415 consecutively enrolled infants identified from a high-risk population screened for drug use through clinical interview, urine, and meconium screens. Ninety-three percent retention for surviving participants at 4 years of age resulted in 376 children (190 cocaine-exposed and 186 nonexposed). Main Outcome Measure: The Wechsler Preschool and Primary Scales of Intelligence-Revised. Results: After control for covariates, prenatal cocaine exposure was not related to lower full-scale IQ (cocaine exposed [80.7] vs nonexposed [82.9]; P=.09) scores or summary verbal (cocaine exposed [79.9] vs nonexposed [81.9]; P=.11) or performance (cocaine exposed [85.5] vs nonexposed [87.5]; P=.18) IQ scores at age 4 years. However, prenatal cocaine exposure was related to small but significant deficits on several subscales (mean [SE]): visual-spatial skills (cocaine exposed [7.3 (0.22)] vs non-exposed [8.2 (0.22)]; P=.01), general knowledge (cocaine exposed [6.1 (0.18)] vs non-exposed [6.7 (0.17)]; P=.04), and arithmetic skills (cocaine exposed [6.2 (0.20)] vs nonexposed [6.8 (0.20)]; P=.05). Prenatal cocaine exposure was also associated with a lower likelihood of achievement of IQ above normative means (odds ratio, 0.26 [95 confidence interval, 0.10-0.65] P=.004). The quality of the caregiving environment was the strongest independent predictor of outcomes. Cocaine-exposed children placed in nonrelative foster or adoptive care lived in homes with more stimulating environments and had caregivers with better vocabulary scores, and they attained full-scale and performance IQ scores (83 and 87, respectively) similar to nonexposed children in biological maternal or relative care (full-scale IQ, 82; performance IQ, 88) and higher than cocaine-exposed children in biological maternal or relative care (full-scale IQ, 79; performance IQ, 84). Conclusions: Prenatal cocaine exposure was not associated with lower full-scale, verbal, or performance IQ scores but was associated with an increased risk for specific cognitive impairments and lower likelihood of IQ above the normative mean at 4 years. A better home environment was associated with IQ scores for cocaine-exposed children that are similar to scores in nonexposed children. (Author' s abstract)