Our objective was to test if protease inhibitors (PIs) increase the incidence of fetal growth restriction (FGR). Human immunodeficiency (HIV)-seropositive women were studied. At birth the neonatal weight percentile was assigned by predicted growth potential (GP), accounting for race, parity, weight, height, gestational age, birthweight, and gender (Gardosi, 1992). FGR was defined as GP < 10% percentile. Maternal age, CD4 count, viral load, weight gain, prenatal care, tobacco, alcohol, substance abuse, and PI use were related to FGR using chi-square and multiple regression analysis. Ninety-three of 191 women received PI. In these, FGR occurred in 27 (29%) compared with 15 (15.3%) in the non-PI group ( P = 0.02). Maternal CD4 count ( P < 0.0001) was the primary determinant, and smoking ( P = 0.037) was an independent cofactor for FGR (Nagelkerke r2 = 0.24). Twenty-six of 82 (31.7%) smokers had FGR, versus 16 of 109 (14.7%) of nonsmokers (odds ratio, 2.69; 95% confidence interval, 1.33 to 5.46; P = 0.005). After exclusion of the CD4 count, PI became a cofactor for FGR ( P = 0.021 and Nagelkerke r2 = 0.104). We concluded that maternal HIV status and smoking determine the risk for FGR. Although PIs increase the risk for FGR, this effect appears to depend on maternal disease severity.