Data from 2543 HIV-infected women were analyzed to correlate antiretroviral therapy (ART) used during pregnancy with maternal and pregnancy outcomes. ART was analyzed according to class of agents used and according to monotherapy versus combination ART containing neither protease inhibitors (PIs) nor nonnucleoside reverse transcriptase inhibitors versus highly active ART. Timing of ART was classified according to early (recorded at or before 25-week gestation study visit) and late (recorded at 32-week gestation or delivery visit) use. Maternal outcomes assessed included hematologic, gastrointestinal, neurologic, renal, and dermatologic complications; gestational diabetes; lactic acidosis; and death. Adverse pregnancy outcomes assessed included hypertensive complications; pre-term labor or rupture of membranes; preterm delivery (PTD); low birth weight; and stillbirth. Logistic regression analyses controlling for multiple covariates revealed ART to be independently associated with few maternal complications: ART use was associated with anemia (odds ratio [OR] = 1.6, 95% confidence interval [CI]: 1.1-2.4), and late use of ART was associated with gestational diabetes (OR = 3.5, 95% CI: 1.2-10.1). Logistic regression analyses revealed an increase in PTD at <37 weeks for 10 women with late use of ART not containing zidovudine (ZDV; OR = 7.9, 95% CI: 1.4-44.6) and a decrease in adverse pregnancy outcomes as follows: late use of ART containing ZDV was associated with decreased risk for stillbirth and PTD at <37 weeks (OR = 0.06, 95% CI: 0.02-0.18; OR = 0.5, 95% CI: 0.3-0.8, respectively), and ART containing nucleoside reverse transcriptase inhibitors but not ZDV during early and late pregnancy was associated with decreased risk for PTD at <32 weeks (OR = 0.3, 95% CI: 0.2-0.7). Benefits of ART continue to outweigh observed risks.