STUDY DESIGN AND PATIENT POPULATION
In cohort 1 of the P1060 trial, we enrolled HIV-infected children between 6 and 36 months of age who had had documented exposure to a single dose of nevirapine for perinatal prevention of HIV transmission. According to version 1.0 of the protocol, the child could be eligible for inclusion if either the mother or the child had received nevirapine, but the protocol was subsequently amended to require specifically that the child had received nevirapine. The protocol, including the statistical analysis plan, is available with the full text of this article at NEJM.org. The authors attest that the study was performed in accordance with the protocol and the statistical analysis plan.
To ensure that the children were infected with HIV during the time of exposure to single-dose nevirapine — and not later during breast-feeding — the inclusion criteria specified that the children had to have received the diagnosis of HIV infection or to have had an acquired immunodeficiency syndrome (AIDS)–defining event by 60 days of age or that they had to have been formula-fed exclusively from birth. In addition, children had to be eligible for treatment according to World Health Organization (WHO) criteria and to have baseline plasma HIV type 1 (HIV-1) RNA levels of more than 5000 copies per milliliter. Children who were undergoing treatment for tuberculosis were ineligible. Enrollment was stratified according to age (<12 months vs. ≥12 months), and children were randomly assigned, in a 1:1 ratio, to an NNRTI-based regimen of nevirapine, zidovudine, and lamivudine or a protease-inhibitor–based regimen of ritonavir-boosted lopinavir, zidovudine, and lamivudine. In the event of a toxic reaction or contraindication to zidovudine, stavudine was substituted.
Children were enrolled at nine sites in Africa: four in South Africa and one each in Zimbabwe, Zambia, Malawi, Uganda, and Tanzania. The study was approved by the ethics review committee at each local site, the institutional review board at each partner U.S. institution, and the Ministry of Health, where needed. Written informed consent was obtained from the children’s parents or legal guardians. Study visits were conducted 2 and 4 weeks after the initiation of treatment, every 4 weeks until week 16, at week 24, and every 12 weeks thereafter. Plasma HIV-1 RNA levels were measured at site laboratories with the use of the standard Roche Amplicor v1.5 test kit (eight sites) or the Abbott RealTime HIV-1 assay (one site), with all the laboratories participating in the Division of Acquired Immunodeficiency Syndrome (DAIDS) Quality Assurance Program. Tests for HIV resistance were performed retrospectively on plasma samples at a single laboratory with the use of the ViroSeq HIV-1 Genotyping System, versions 2.7 and 2.8 (Celera), and ViroSeq software (in which lower-end sensitivity requires that approximately 20% of the virus population possess a resistance mutation for detection).
STUDY END POINTS
The primary study objective was to compare the rates of treatment failure by 24 weeks in the two study groups. Treatment failure was defined as permanent discontinuation of the treatment regimen for any reason, including death, toxic effects, and virologic failure. Virologic failure was defined as a confirmed plasma HIV-1 RNA level of less than 1 log10 copies per milliliter below the study-entry level at 12 to 24 weeks after the initiation of treatment or a confirmed plasma HIV-1 RNA level of more than 400 copies per milliliter at 24 weeks. The need for treatment of tuberculosis during the course of the study was a reason for discontinuation of the study treatment.
Secondary end points included confirmed virologic failure or death by week 24, confirmed virologic failure (as defined above until week 24 or defined as a confirmed viral rebound of more than 4000 copies per milliliter after week 24) or death at any time during the follow-up period, and a composite of virologic failure or discontinuation of study treatment at any time during the follow-up period. Data for children who were lost to follow-up were censored at the date of the last available measurement of the HIV-1 RNA level; children whose last measurement showed virologic failure but for whom no confirmatory value was obtained were classified as having had virologic failure.
We estimated that with a sample of 288 participants, the study would have more than 90% power to detect an absolute difference of 20 percentage points in the rate of the primary end point between the two treatment groups. The P1060 trial was monitored at 6-month intervals by an independent data and safety monitoring board chartered by the National Institute of Allergy and Infectious Diseases. Stopping guidelines according to a Peto–Hay-bittle rule were proposed, with a two-sided nominal P value of less than 0.001 required before closure of enrollment in a cohort would be considered.
Reported P values have not been adjusted for interim analyses or multiple comparisons. The rate of treatment failure at week 24, with associated standard errors, was calculated from Kaplan–Meier curves for each treatment group and each age stratum separately. Rates of treatment failure and differences between study groups were calculated for the primary end point, weighted by the inverse of the variance in each age stratum. Unweighted rates (i.e., without stratification according to age) are also reported. Similar methods were used for the secondary end point of virologic failure or death at 24 weeks. We performed between-group comparisons of the time to each primary and secondary end point using Cox proportional-hazards models stratified according to age and based on the discrete version of the likelihood function, with P values calculated by means of the Wald test. Subgroup analyses were performed to assess whether the difference between treatments varied among subgroups of children; we tested for interactions of treatment with subgroup and compared treatment differences within subgroups, recognizing that the large number of statistical tests performed could result in some significant values (P<0.05) being observed by chance. Data from children who did not reach an end point were censored at the time of the last available measurement of the HIV-1 RNA level at or before the last follow-up assessment that was performed before April 20, 2009.
Safety analyses were limited to the period during which a child received a study treatment. For each type of adverse event, the highest-grade events were tabulated, and the time to a first grade 3 or higher event was compared between groups with the use of a stratified Cox proportional-hazards model.
We summarized the changes from baseline to each study visit in the percentage of CD4+ lymphocytes and in z scores for height and weight (using the Centers for Disease Control charts; www.cdc.gov/nchs/nhanes/growthcharts
) according to treatment group, using all values obtained from children who were receiving the study treatment as well as from those who had discontinued treatment (an intention-to-treat approach). A multivariate Wilcoxon-test statistic was used to test for treatment differences across the first 96 weeks of the study.