• Proportion of women in each treatment group who continue on original therapy with virologic suppression to less than 500 copies/ml at 34 weeks gestation (or the last viral load determination prior to delivery)
  
• proportion of women in each treatment group who continue on original therapy with virologic suppression to less than 500 copies/ml at 48 weeks postpartum
  

• Proportion of women in each treatment group who continue on original therapy with virologic suppression to less than 50 copies/ml at 34 weeks gestation (or the last viral load determination prior to delivery)
  
• proportion of women in each treatment group who continue on original therapy with virologic suppression to less than 50 copies/ml at 48 weeks postpartum, and to less than 500 and 50 copies/ml at 104 weeks postpartum
  
• study treatment adherence and health status by self-report, correlated with predose nelfinavir or nevirapine level at 34 weeks gestation and 8 weeks postpartum
  
• difference between postpartum and pregnancy 12-hour area under the concentration curve (AUC) for nevirapine
  
• time of trough levels in relation to the morning dose of nevirapine and nelfinavir at 34 weeks gestation and 8 weeks postpartum and correlation of trough levels with viral load
  
• incidence of HIV viral resistance by genotype among women in each treatment group at the time of virologic failure
  
• incidence of abnormal glucose tolerance, gestational diabetes, and abnormal lactate levels during pregnancy in each treatment group
  
• incidence of impaired glucose tolerance, diabetes, hyperinsulinemia, and elevated cholesterol and triglycerides at 8 weeks postpartum in each treatment group
  
• incidence of anemia, hypoglycemia, and abnormal liver function studies among infants born to women in each treatment group
  
• incidence of prematurity (less than 37 weeks), extreme prematurity (less than 32 weeks), low birth weight (less than 2.5 kg), and very low birth weight (less than 1.5 kg) among infants born to women in each treatment group
  
• perinatal HIV transmission among infants born to women in each treatment group
  

The optimal treatment strategy for women who initiate antiretroviral therapy during pregnancy is not known. Although PI-based antiretroviral regimens are prescribed with increasing frequency among pregnant women, the efficacy and safety of this approach is unknown. Pregnant women are at increased risk for glucose intolerance and insulin resistance; PIs are associated with glucose intolerance. Physiologic differences between pregnant women and nonpregnant adults may alter the pharmacokinetics of antiretroviral regimens. Fetal safety considerations and effects on perinatal HIV transmission must also be considered when selecting an antiretroviral regimen for pregnant women. This trial will compare PI-based and PI-sparing antiretroviral regimens for women initiating antiretroviral therapy in pregnancy.

Women will be stratified on the basis of viral load (50,000 or less copies/ml or greater than 50,000 copies/ml) and gestational age at entry (20 or less weeks or greater than 20 weeks) and then randomized to one of two treatment groups. Group A will receive the PI nelfinavir (NFV) with zidovudine (ZDV) and lamivudine (d4T); Group B will receive nevirapine (NVP) with ZDV and d4T. Women will have clinic visits for physical and obstetrical examinations at 2, 4, 6, and 8 weeks after entry and then every 4 weeks until delivery. After delivery, infants in both groups may receive ZDV until they are 6 weeks old. Infants are evaluated for safety and to test the infant's blood for HIV-1 at birth and at Weeks 2, 8, 16, and 24.

Women will continue on assigned antiretroviral therapy postpartum and will have 11 postpartum clinic visits over a period of 2 years. Blood samples from women will be evaluated for safety and for virologic, pharmacokinetic, and metabolic studies. The first 12 women randomized to Group A will undergo a 4-hour pharmacokinetic profile at 32 to 36 weeks gestation and at 8 weeks postpartum to determine the timing of the nelfinavir trough. The first 20 women randomized to Group B will undergo an 8-hour pharmacokinetic profile at either 16 to 24 weeks or 32 to 36 weeks gestation and then again at 8 weeks postpartum to characterize pharmacokinetics of nevirapine at steady state in pregnancy and in the postpartum period.