The Surveillance Monitoring for ART Toxicities (SMARTT) cohort of the Pediatric HIV/AIDS Cohort Study includes over 3,500 HIV-exposed but uninfected infants and children at 22 sites in the US, including Puerto Rico. The goal of the study is to determine the safety of in utero exposure to antiretrovirals (ARVs) and to estimate the incidence of adverse events. Domains being assessed include metabolic, growth and development, cardiac, neurological, neurodevelopmental (ND), behavior, language, and hearing. SMARTT employs an innovative trigger-based design as an efficient means to identify and evaluate adverse events. Participants who met a predefined clinical or laboratory threshold (trigger) undergo additional evaluations to define their case status. After adjusting for birth cohort and other factors, there was no significant increase in the likelihood of meeting overall case status (case in any domain) with exposure to combination ARVs (cARVs), any ARV class, or any specific ARV. However, several individual ARVs were significantly associated with case status in individual domains, including zidovudine for a metabolic case, first trimester stavudine for a language case, and didanosine plus stavudine for a ND case. We found an increased rate of preterm birth with first trimester exposure to protease inhibitor-based cARV. Although there was no overall increase in congenital anomalies with first trimester cARV, a significant increase was seen with exposure to atazanavir, ritonavir, and didanosine plus stavudine. Tenofovir exposure was associated with significantly lower mean whole-body bone mineral content in the newborn period and a lower length and head circumference at 1 year of age. With ND testing at 1 year of age, specific ARVs (atazanavir, ritonavir-boosted lopinavir, nelfinavir, and tenofovir) were associated with lower performance, although all groups were within the normal range. No ARVs or classes were associated with lower performance between 5 and 13 years of age. Atazanavir and saquinavir exposure were associated with late language emergence at 1 year, but not at 2 years of age. The results of the SMARTT study are generally reassuring, with little evidence for serious adverse events resulting from in utero ARV exposure. However, several findings of concern warrant further evaluation, and new ARVs used in pregnancy need to be evaluated.