A Phase IIB randomized, controlled trial to evaluate the safety, tolerance, and HIV vertical transmission rates associated with short course nevirapine (NVP) vs. short course zidovudine (ZDV) in HIV infected pregnant women and their infants in Uganda

Executive Summary

This is the preliminary report, based on data from HIVNET 012. The database cutoff date for this report was June 30, 1999.

ABSTRACT

Background

HIVNET 012 was originally designed as a 1500 person randomized double blind Phase III placebo-controlled trial to determine the safety and efficacy of oral zidovudine (ZDV) and the efficacy of oral nevirapine (NVP) for the prevention of vertical transmission of HIV-1 infection in pregnant women and their neonates in Uganda. However, after the announcement of the results of the CDC sponsored Thai ZDV trial in February 1998, the placebo arm was dropped after 49 enrolled mothers had given birth. After Institutional Review Board approval in Uganda and the United States, randomized enrollment continued into the two active arms in order to provide screening information to select the more promising of the two active arms for inclusion in a revised protocol having an appropriate control arm.

Methods

Eligible patients were HIV-infected pregnant women in Uganda (after 36 weeks gestation) who had no antiretroviral treatment during the current pregnancy. The target sample size was a total of 556 women (500 fully evaluable mother-infant pairs) in the two active arms; this was chosen so as to provide at least 80% probability of correctly choosing NVP as the more promising active arm for inclusion in the revised protocol if NVP in truth was at least equivalent to ZDV in terms of the HIV transmission rate. Women were randomized to receive either NVP (200 mg dose orally in labor and 2 mg/kg to her infant within 72 hr after birth) or ZDV (600 mg orally plus 300 mg every 3 hr thereafter during labor and 4 mg/kg orally twice a day to her infant for 7 days after birth). The primary endpoints for this analysis were rate of HIV infection and HIV free survival at 6-8 weeks and 14-16 weeks of age. Infants were defined as HIV infected based on a positive qualitative HIV RNA assay (Roche AMPLICOR assay) which was confirmed by either a quantitative HIV RNA assay (Roche AMPLICOR MONITOR assay) or HIV culture on a second blood specimen. In the case of an infant death where there was only one positive RNA assay on the specimen preceding death, the infant was considered to be HIV-1 infected.

Results

Six hundred forty-five women were enrolled as of the April 30, 1999 enrollment cutoff date. Of these 645 mothers, 19 were randomized to placebo, 313 to ZDV and 313 to NVP. In turn, 5 mothers randomized to ZDV and 3 to NVP were lost to follow-up before delivery. With eleven sets of twins and one set of triplets, 312 infants were born to the 308 ZDV mothers, while 319 infants were born to the 310 NVP mothers. Unless indicated otherwise, this Executive Summary will focus on demographics and efficacy results in these 308 ZDV and 310 NVP mothers and in their first born infants. (At the time of this report, toxicity data were available from the first 556 mother/infant pairs who were randomized to ZDV and NVP.)

The mothers' median age was 24 years (1st and 3rd quartiles 21, 27), the median CD4 cell count at approximately 36 weeks gestation was 448 cells/uL (1st and 3rd quartiles 275, 643) and the median plasma HIV-1 RNA level was 26943 copies/mL (1st and 3rd quartiles 7872, 81550).

The frequency of cesarean sections was slightly higher on ZDV than on NVP (i.e. 13.3% vs. 11.0%). Membranes ruptured more that 4 hours before delivery slightly less frequently in the ZDV mothers than in the NVP mothers (i.e., 12.5% vs. 16.2%). Other baseline characteristics of the women, including maternal CD4 cell count, plasma HIV-1 RNA level, and duration of labor were reasonably well balanced between the two treatment arms.

There were 3 stillbirths, (2 in the ZDV group and 1 in NVP). The median one minute Apgar was 10 (quartiles 9, 10), the 5 minute Apgar was 10 (quartiles 10, 10), and the median birth weight was 3100 grams (quartiles 2800, 3400). While no infants weighed less than 1500 grams at birth, 8 (1.3%) weighed between 1500-2000 grams, and 33 (5.3%) weighed between 2000 and 2500 grams.

Infant gender was well balanced, with 51.3% of ZDV infants and 49.4% of NVP infants being female. The NVP infants tended to have somewhat lower birthweight, with 68 (23.0%) of ZDV infants having birthweights below 2800 grams, compared to 91 (30.0%) of NVP infants. At 14 weeks of age 95.2% of infants were still breastfeeding (95.4% in the ZDV arm and 95.0% in the NVP arm).

Two infants were lost to follow-up before dosing and did not contribute any information to the analysis of HIV-free survival. These two plus seven additional infants who died before HIV testing did not contribute information to the analysis of time to HIV-infection. Thus, the analysis of HIV-free survival includes data on 307 ZDV infants and 309 NVP infants, while the time to HIV infection analysis includes data on 302 ZDV and 307 NVP infants.

At the time of analysis, at 6-8 weeks, 59 infants in the ZDV group and 35 infants in the NVP group were HIV infected. At 14-16 weeks, 65 infants in the ZDV group and 37 infants in the NVP group were HIV infected. The estimated percent infected based on a Kaplan-Meier analysis for the ZDV and NVP groups respectively were 10.4% and 8.2% by day 3, 21.3% and 11.9% by 6-8 weeks, and 25.1% and 13.1% by 14-16 weeks. The two-sided p-values for these differences at 3 days, 6-8 weeks and 14-16 weeks were 0.354, 0.0027 and 0.0006 respectively.

With regard to HIV-free survival, at 6-8 weeks 66 infants in the ZDV group and 38 infants in the NVP group were HIV infected or had died. At 14-16 weeks, 74 infants in the ZDV group and 41 in the NVP group were HIV infected or had died. The Kaplan-Meier (KM) estimated percent of infants that were HIV infected or had died, for the ZDV and NVP groups, were 12.2% and 8.8% at 3 days, 23.1% and 12.8% at 6-8 weeks, and 27.6% and 14.4% at 14-16 weeks. The two-sided p-values for these differences at 3 days, 6-8 weeks and 14-16 weeks were 0.175, 0.0012 and 0.0002 respectively.

Reported maternal and infant deaths and side effects (including anemia and rash) were balanced between the two randomized groups. Infection was the largest cause of maternal and infant serious adverse events including deaths in both groups. Of the 59 serious adverse events reported among infants within 56 days of birth, there were 4 (1.4%) in the ZDV arm and 2 (0.7%) in the NVP arm which were judged to all be possibly, but unlikely related to study drug. Of the 25 serious adverse events reported among mothers within 6 weeks of birth, there was 1 event (ZDV arm) judged to all be possibly, but unlikely related to study drug.

Conclusion
A treatment regimen consisting of a single oral 200 mg dose of NVP given to the mother at onset of labor and a single 2mg/kg dose given to the infant within 72 after birth significantly reduced the risk of perinatal transmission in a breast feeding population during the first 14 weeks of life compared with a treatment regimen consisting of oral ZDV given in labor and to the infant during the first week of life. The ZDV and the NVP regimens appeared to have similar low rates of serious adverse events and non-serious adverse events in both mothers and infants. It is unknown whether there are any long-term toxicities for the ZDV or NVP treated group. The relative impact of these regimens on long-term survival and on later transmission associated with breastfeeding is also unknown. Therefore, long-term follow-up of all HIVNET 012 infants remains a high priority.

Press releases, fact sheets and other NIAID-related materials are available on the NIAID Web site at http://www.niaid.nih.gov.