| |
A Randomized, Double-Blind, Placebo-Controlled Trial of Combined Nevirapine and AZT Compared with Nevirapine Alone in the Prevention of Perinatal Transmission of HIV in Zimbabwe
|
| |
| |
Clinical Infectious Diseases Jan 1, 2007;44:111-119
Paul Thistle,1,4 Rachel F. Spitzer,4 Richard H. Glazier,5,6,9,10 Richard Pilon,12 Gordon Arbess,5,11 Andrew Simor,7 Eleanor Boyle,11 Inam Chitsike,2 Tsungai Chipato,3 Maureen Gottesman,5,11 and Michael Silverman8
1The Salvation Army Howard Hospital, Glendale, and Departments of 2Paediatrics and 3Obstetrics and Gynaecology, University of Zimbabwe, Harare, Zimbabwe; and Departments of 4Obstetrics and Gynaecology, 5Family and Community Medicine, 6Public Health Sciences, and 7Laboratory Medicine and Pathobiology and 8Lakeridge Heath Center, Division of Infectious Diseases, University of Toronto, 9Institute for Clinical Evaluative Sciences, 10Centre for Research on Inner City Health, and 11Department of Family and Community Medicine, St. Michael's Hospital, Toronto, and 12National HIV and Retrovirology Laboratories, Public Health Agency of Canada, Ottawa, Canada
ABSTRACT
Background. A single dose of nevirapine (sdNVP) administered to both mother and infant can decrease mother-to-child transmission of human immunodeficiency virus (HIV) by 47%, compared with ultra-short course zidovudine therapy (usZDV). There is limited data about the benefit of usZDV added to sdNVP to prevent mother-to-child transmission.
Methods. We performed a double-blind, randomized, placebo-controlled trial to determine whether usZDV combined with sdNVP improved neonatal outcome, compared with sdNVP alone. Mothers were randomized to 1 of 2 treatment groups. Mothers in the usZDV/sdNVP group received a loading dose of zidovudine (600 mg administered orally) and continued to receive 300-mg doses of zidovudine orally every 3 h while in labor, and their infants received zidovudine at a dosage of 2 mg per kg of body weight 4 times per day orally for 72 h. Mothers and infants in the sdNVP group received zidovudine placebo dosed in the same manner. All mothers also received nevirapine at a dosage of 200 mg orally while in labor, and all infants received nevirapine 2 mg per kg of body weight orally within 72 h of delivery.
Results. The study was stopped on the basis of futility, because interim data showed that, at present trends, superiority would not be demonstrated. Results at 6 weeks of age were available for 609 infants. The primary end point of HIV RNA positivity or death occurred in 21.8% of infants in the usZDV/sdNVP arm and 23.6% of the infants in the sdNVP arm.
Conclusion. usZDV, when added to a standard 2-dose regimen of sdNVP, did not demonstrate a clinically important decrease in the combined end point of mother-to-child transmission or infant death. High rates of adverse maternal and infant outcome in both study arms suggest that improved approaches are necessary.
Introduction
Various short and ultra-short course therapies to decrease mother-to-child transmission (MTCT) of HIV in developing countries have been studied [1-7]. Specifically, single-dose nevirapine (sdNVP) administered to both the mother and the neonate and an ultra-short regimen of zidovudine (usZDV) have been compared with one another in short-course treatment, with the sdNVP group showing a 47% decrease in transmission [1]. These studies have served as the basis for the World Health Organization (WHO) recommendation for antenatal and intrapartum antiretroviral therapy for HIV-seropositive women [8,9]. Despite such research, very few eligible women in sub-Saharan Africa currently receive any regimen for the prevention of MTCT of HIV [10]. Recent WHO recommendations include sdNVP monotherapy when there is only the capacity to deliver a minimal number of antiretroviral drugs for the prevention of MTCT. If there is no capacity to deliver maternal HAART, but there is capacity to deliver other antiretrovirals, then the addition of ZDV with or without additional lamivudine to the mother is recommended to prevent the emergence of NVP-resistance mutations and to possibly further reduce transmission [8].
To date, there are limited data on combination short-course therapy that includes sdNVP for the prevention of MTCT of HIV. A study from Thailand has shown a benefit to adding sdNVP (compared with placebo) to a ZDV regimen commenced at 28 weeks of gestation [11]. Furthermore, in the absence of antepartum or intrapartum therapy, adding ZDV (compared with placebo) to neonatal-only NVP therapy in the first 7 days of life has a protective efficacy of 36% [12]. Combination ZDV and lamivudine given antenatally, intrapartum, and postpartum to the infant is considered to be more efficacious than NVP monotherapy [13]. Short courses of ZDV or a combination of ZDV and lamivudine did not differ from NVP monotherapy in efficacy [13].
We sought to evaluate whether an inexpensive regimen of sdNVP and usZDV, given both intrapartum to the mother and postpartum for 3 days to the neonate, is superior in the prevention of MTCT of HIV, compared with a regimen of sdNVP alone, in a breast-feeding population in rural Zimbabwe. Both of these regimens are within the financial means of most Zimbabweans ($13 and $7 US for the combination regimen and the sdNVP-only regimen, respectively). We previously demonstrated that, in Zimbabwe, usZDV therapy was not inferior to a course of 4 weeks of antenatal ZDV therapy [2]. In our setting, many women do not present for antenatal care, and therefore, this combination has the advantage of being practical and feasible. Combination therapy was studied for its potential ability to reduce MTCT further than sdNVP monotherapy. Our hypothesis was that combination therapy would be more effective than sdNVP monotherapy.
DISCUSSION
Although combination regimens of ZDV and sdNVP have shown benefit over therapy with ZDV alone, it is important to clarify whether the addition of usZDV produces a further benefit in reduction of transmission, compared with the common approach of therapy with sdNVP alone. Unfortunately, in this study, usZDV/sdNVP short-course therapy for the prevention of MTCT of HIV did not show significant benefit over treatment with sdNVP alone.
A recent meta-analysis suggested that there was no benefit from short-course ZDV therapy in the prevention of MTCT [13]. However, this analysis had to compare outcomes from studies using short-course ZDV with outcomes from the placebo arm in other studies (i.e., no direct comparisons were available). Furthermore, this analysis did not include the 2 randomized studies that did find benefit when directly assessing ultra-short regimens in countries in which clade C viruses were endemic [2, 12].
In a recent study, adding postpartum ZDV therapy to antepartum and postpartum NVP therapy (i.e., sdNVP therapy without any antepartum ZDV treatment) was not of any benefit [18]. In this latter study, however, the ZDV therapy was not started until a median of 9.3 h postpartum [18]. Our study differs significantly from this previous work, in that we added ZDV to the regimen both during labor and postpartum. Thus, at the time of birth, ZDV loading had already occurred. This may explain the slight (but statistically nonsignificant) trend towards benefit seen in our study, compared with the slight trend towards inferiority seen in the study by Taha et al. [18]. Neither approach demonstrated statistically significant benefit when administered in conjunction with maternal and infant NVP therapy (i.e., sdNVP). However, when no maternal NVP therapy was given and only infant NVP was administered postpartum, usZDV therapy did demonstrate benefit [12]. Therefore, usZDV helps to prevent MTCT if the infant does not receive NVP until after birth; however, if the mother receives NVP prepartum, ZDV therapy does not further reduce MTCT, regardless of whether the infant is treated only postpartum or is also treated intrapartum (via the mother). This can be explained by the ability of usZDV regimens to show an incremental benefit only when the transmission rates are higher, such as when maternal NVP therapy is absent (table 6).
Several studies suggest significant expression of NVP-resistance mutations in individuals exposed to short-course therapy, as well as poorer viral suppression with the use of subsequent nonnucleoside reverse-transcriptase inhibitor-based antiretroviral therapy, if such therapy is started soon after delivery [19-22]. Recently, it has been shown that the addition of ZDV and lamivudine therapy, administered to the mother and the child for 4 or 7 days, can reduce the emergence of NVP resistance in both the mothers and infants [23]. Short-course ZDV may inhibit the development of NVP resistance in the neonate, as suggested in another study [24], although our data were inconclusive because of problems with sample storage. In this study, we do not have any data on maternal NVP resistance.
Study limitations. This study was performed at a single health care center. However, this hospital is relatively better funded (receiving funding both from the Zimbabwe Ministry of Health and from the Salvation Army) and, therefore, has more staff than most rural Zimbabwean health care centers. This hospital likely reflects the best possible care in this region.
During the course of this study, Zimbabwe experienced severe socioeconomic difficulties. In particular, a gasoline shortage led to severe disruption of public and private transportation and, thus, to difficulties in patient follow-up and outreach visits.
Study strengths. This study had very liberal inclusion criteria and very narrow exclusion criteria. This led to enrollment of 70.8% of all screened HIV-positive mothers. Only 6 (19%) of 32 Adult AIDS Clinical Trials Group or Terry Beirn Community Programs for Clinical Research on AIDS trials funded by the National Institute of Allergy and Infectious Disease enrolled as high a percentage of screened patients [25]. Therefore, although a gasoline shortage led to a high dropout rate, 37.8% of all HIV-positive screened mothers were enrolled and had their infants followed-up to 6 weeks postpartum. Although some studies did not record the number of HIV-positive women who were screened [11, 26], this study had a higher total percentage of patients with follow-up data than did many landmark developing-world MTCT studies [1, 3, 4, 7, 12, 14, 18, 27]; therefore, our study likely reflects real-world effectiveness in a rural African population. The high rates of the combined end point of HIV transmission or neonatal death (as well as the high rate of maternal death) seen in both arms of this study (table 5 and figure 1) are, therefore, especially concerning and necessitate development of improved interventions. Previous studies by our group and others, conducted in areas in which HIV infection is endemic and which are without access to antiretroviral agents, have found similar maternal and neonatal mortality rates [28-30]. It is felt that these high mortality rates are related to the direct effects of HIV infection, as well as to opportunistic infections resulting from immunosuppression [28-30]. Therefore, decreasing this mortality rate and the HIV-transmission rate will require intensified strategies. This may involve maternal CD4+ cell testing, antepartum initiation of HAART, and administration of more-prolonged antiretroviral therapy to the neonate.
RESULTS
Study recruitment was conducted between December 2002 and August 2004. An interim analysis was conducted in July 2004, after the publication of a study by Taha et al. [18] that examined a different regimen that combined sdNVP and ZDV and that showed no benefit in ZDV therapy to the neonate alone. As a result of this interim analysis, study recruitment was terminated in August 2004 on the basis of futility, because the analysis indicated that the likelihood of detecting a significant difference between the 2 arms of the study at present enrollment rates was extremely low.
There were 7467 pregnant women who consented to HIV testing after pretest counseling. Of these, 1610 had test results that indicated that they were HIV seropositive, indicating a prevalence of HIV infection of 21.6% at our study site. A total of 1140 study participants were randomized to receive either sdNVP and usZDV or sdNVP and placebo (figure 1). No mothers had a history of receiving previous antiretroviral therapy or clinical evidence of hepatic disease.
The characteristics of the women in each arm are shown in table 1. The characteristics of the infants in each group are presented in table 2. Infants in the usZDV/sdNVP arm were born at an earlier gestational age than were infants in the sdNVP arm. However, there was no difference between groups in the number of premature infants (i.e., infants born before 37 weeks of gestation) (P = .39). The difference between the 2 arms with respect to the mean values for duration of gestation among premature infants was small (3.5 days), and both mean values were >37 weeks. This difference may, however, have led to the statistically significant, increased incidence of admission to the at-risk nursery for the infants in the usZDV/sdNVP arm. A physician blinded to the study allocation determined the admissions to the at-risk nursery. Because of difficulties with patient follow-up, we also compared characteristics of the mothers (table 3) and infants (table 4) who had 6-week follow-up data with characteristics of those who did not.
Results are available for a total of 609 neonates at 6 weeks, 312 in the usZDV/sdNVP group and 297 in the sdNVP group (table 5). At 6 weeks, 45 of the 312 infants in the usZDV/sdNVP group had test results that were positive for HIV RNA (a prevalence of 14.4%), and 49 of the infants in the sdNVP group had test results that were positive for HIV RNA (a prevalence of 16.5%). The primary outcome measure (HIV RNA-positive status or death at 6 weeks) occurred in 21.8% of infants treated with usZDV/sdNVP, compared with 23.6% of infants treated with sdNVP (P = .06; percentage difference, 1.8%; 95% CI, -4.9% to 8.4%). There was no significant difference between the 2 treatment arms before and after adjusting for prognostic covariates. The unadjusted and adjusted ORs (adjusted for age, gestational age, marital status, premature rupture of membranes, mode of delivery, maternal opportunistic infection, and sexually transmitted infection) were 1.259 (95% CI, 0.743-2.132) and 1.282 (95% CI, 0.751-2.188) respectively. At present trends, 7500 patients would have to be enrolled in each group to have a power of 0.8 to detect this difference at P < .05.
There were 19 sets of multiple births, including 18 sets of twins and 1 set of triplets. Only data for the first-born infant was included in the analysis. No cases of severe drug toxicity were observed. Because of difficulties in drug-resistance genotyping caused by the long-term storage of dried blood spot samples at ambient temperature and humidity, drug-resistance genotypes were obtained for only 31 of the HIV-infected infants. The NVP-resistance mutation Y181C (which, in most cases, was present as a wild-type/mutant mixture) was detected in 1 (10%) of 10 infants in the usZDV/sdNVP group and 7 (21%) of 21 infants in the sdNVP group (relative risk, 0.30; 95% CI, 0.04-2.12; P = .2216).
METHODS
This was a randomized, double-blinded, placebo-controlled trial that compared 2 short-course regimens of antiretroviral prophylaxis. After voluntary counseling before and after the HIV serological testing (which is offered to all antenatal mothers at the study site), pregnant, HIV-seropositive women were approached for consent to participate in the trial. All counseling was conducted and consent was obtained in the local Shona language. Those patients who consented were randomized to 1 of 2 interventions.
In the usZDV/sdNVP group, women received a loading dose of 600 mg of ZDV orally and then received 300 mg of ZDV orally every 3 h while in labor. Infants received ZDV at a dosage of 2 mg per kg of body weight orally 4 times per day for 72 h after delivery. In group B (the sdNVP group), a placebo was substituted for ZDV and was dosed in a manner identical to that of the dosing in the first group, to ensure blinding. In both groups, women also received a single 200-mg dose of NVP administered orally while in labor, and their infants received NVP at a dosage of 2 mg per kg of body weight once within 72 h of delivery (this was given before discharge from the hospital).
Study site. The study was conducted at the Salvation Army Howard Hospital, a 140-bed hospital in the rural Chiweshe communal land, 80 km north of Harare. This is the referral center for the Mazowe District of Mashonaland Central Province and has a catchment area of >250,000 people. There is a family and child health clinic on site and a satellite network of rural health clinics in villages as far as 100 km away. The on-site clinical and research team consists of a board-certified obstetrician, 3 trained counsellors, and state-registered midwives.
At the first antenatal visit to the family and child health clinic, demographic, obstetric, and general medical history data were gathered, and the findings of a physical examination were recorded on case report forms. Routine antepartum and postpartum care was provided to all patients, including screening and treatment for syphilis and anemia, as per Zimbabwean Ministry of Health and Child Welfare national guidelines. Similarly, sexually transmitted infection treatment was provided as necessary using a syndromic approach. Antenatal maternal vitamin supplements were also offered free of charge, including iron and folate.
Trained study midwives supervised administration of medication to the mother during labor and delivery and, subsequently, to her infant; thus, there was good adherence to the assigned regimen. Cesarean deliveries were performed for obstetrical indications at the discretion of the attending physician, who was blinded to patient allocation. Elective cesarean deliveries for HIV positivity alone were not performed.
Inclusion criteria included being an HIV-positive pregnant woman, having the ability and willingness to give informed consent for study participation, and the ability and willingness to have infants involved in follow-up. Exclusion criteria included the inability to give or the refusal of informed consent, clinical evidence of significant hepatic disease (e.g., jaundice), and receipt of previous antiretroviral therapy.
An initial cohort of 100 women had alanine aminotransferase levels determined 2 weeks after sdNVP dosing, for safety assurance. None of the women had abnormal results, and the study proceeded.
Neonatal HIV status was determined from heel prick dried blood spots collected at birth, 2 weeks, and 6 weeks, which were tested for HIV RNA using the Nuclisens HIV-1 (BioMerieux). Those neonates who did not have a 6-week sample obtained but did have subsequent negative results were recorded as being HIV negative at 6 weeks. Study nurses attempted outreach visits to women and their children if they did not attend a follow-up visit. At each postnatal visit, a history was obtained from the mother, the infant was examined, and the results were recorded on postnatal data entry sheets. The fathers were encouraged to attend all appointments. Counseling about safe sex and condom use was also provided. All infants were provided with routine postpartum care, including vaccinations and treatment, as per Zimbabwean national guidelines. At 3 months after delivery, the mother was offered free cotrimoxazole prophylaxis of opportunistic infections, as per WHO recommendations for women with HIV infection in sub-Saharan Africa. Infants were also provided with cotrimoxazole prophylaxis if their HIV status was unknown or if they were HIV positive. Issues surrounding breast-feeding and HIV infection were addressed in counseling sessions. Given the risks of breast-feeding, especially of mixed breast-feeding [15, 16], we followed the Zimbabwe national guidelines and counselled women to consider rapid weaning at 5-6 months of age, to reduce the risk of vertical transmission inherent in breast-feeding. The possible use of formula was discussed with all women.
The maternal HIV serological test result was determined using the Dipstick HIV 1 and 2 (Immuno Chemical Lab). All samples with positive results were retested using the Recombigen test kit (Cambridge Diagnostics Ireland), which is manufactured from genetically engineered gp41 and gp120 (HIV-1 and HIV-2 env) and p24 (HIV-1 gag) gene products. Only those mothers with positive results for both tests were eligible for study enrollment.
The determination of neonatal infection using Nuclisens HIV-1 (qualitative assay) was performed at the Canadian National HIV and Retrovirology Laboratory (Ottawa, Canada) from infant heel prick dried blood spot samples. Dried blood spot samples were cut into strips using sterile, disposable scissors. Nucleic acids were eluted from dried blood spot samples into Nuclisens Lysis Buffer containing internal controls for 1 h with mixing, and the lysis buffer was transferred to a clean tube to separate from filters. All further steps were performed according to the manufacturer's protocol. In brief, nucleic acids were purified by binding to silica, washing the silica-bound nucleic acids extensively, and eluting them into a low-salt buffer [16, 17]. Target wild-type and control nucleic acids were amplified using nucleic acid sequence-based amplification technology, hybridized to ruthenium-labeled probes and magnetic particles, and detected by electrochemiluminescence.
CD4+ cell assays and viral load assessments were not available in this rural region at the time of the study.
SAS software, version 8 (SAS), was used in the analysis. For the bivariate analysis, independent Student's t tests were used for the continuous data, and the 2 test or Fisher's exact test were used for the categorical data. The primary outcome was infant HIV infection determined by Nuclisens HIV-1 RNA at 6 weeks or infant death by 6 weeks.
A sample size of 580 women in each study arm was calculated to be required to detect a vertical transmission rate of 15% in the sdNVP arm and 10% in the usZDV/sdNVP arm. The power of the study was set at 0.8, and a 1-tailed test with = 0.05 was calculated to enable us to accept or reject the null hypothesis that combination therapy is no more effective in reducing vertical transmission than is perinatal sdNVP therapy alone.
Computer-generated block randomization was used in patient allocation to the 2 study groups. HIV-seropositive women who declined study enrollment were offered sdNVP therapy, which became the standard of care in Zimbabwe in 2001.
To adjust for potential maternal prognostic differences in the 2 groups, an adjusted OR was calculated to determine whether there were any differences in the odds of either death or HIV-positive status at 6 weeks, compared with survival and HIV-negative status at 6 weeks, between the 2 treatment arms.
The procedures followed were in accordance with the ethical standards of the Helsinki Declaration of 1975, as revised in 1983. Ethical approval for the study was obtained both from the Medical Research Council of Zimbabwe in Harare and from the Research Ethics Board of Lakeridge Health Center in Canada.
|
|
| |
| |
|
|
|
