Share this article on:

Postexposure Prophylaxis of Breastfeeding HIV-Exposed Infants With Antiretroviral Drugs to Age 14 Weeks: Updated Efficacy Results of the PEPI-Malawi Trial

Taha, Taha E MBBS, PhD*; Li, Qing PhD; Hoover, Donald R PhD, MPH; Mipando, Linda MN-CH§; Nkanaunena, Kondwani MSc§; Thigpen, Michael C MD; Taylor, Allan MD, MPH; Kumwenda, Johnstone FRCP; Fowler, Mary Glenn MD#; Mofenson, Lynne M MD**; Kumwenda, Newton I PhD, MPH*

JAIDS Journal of Acquired Immune Deficiency Syndromes: August 1st, 2011 - Volume 57 - Issue 4 - p 319-325
doi: 10.1097/QAI.0b013e318217877a
Epidemiology and Prevention

Background: This analysis updates and extends efficacy estimates of the PEPI-Malawi trial through age 24 months at study completion in September 2009.

Methods: Infants of breastfeeding HIV-infected women were randomized at birth to the following: (1) single-dose nevirapine (NVP) + 1-week zidovudine (ZDV) (control); (2) control + extended daily NVP (ExtNVP) through 14 weeks; (3) control + extended daily NVP + ZDV (ExtNVP/ZDV) through 14 weeks. We estimated rates of HIV infection, death and HIV infection, or death using Kaplan-Meier analysis.

Results: This analysis includes 3126 infants uninfected at birth as follows: 1004 control, 1071 ExtNVP, and 1051 ExtNVP/ZDV. By 9 months, HIV infection rates were 5.0% in ExtNVP, 6.0% in ExtNVP/ZDV, and 11.1% in control (P < 0.001 comparing extended regimens with control). At age 24 months, HIV infection rates had risen to ∼11% in the extended arms compared with 15.6% in the controls (P < 0.05). The rates of HIV infection or death were also significantly lower in extended arms. There were no differences in severe adverse events with the exception of higher possibly related events in the ExtNVP/ZDV arm.

Conclusions: Daily infant antiretroviral prophylaxis reduces postnatal HIV infection by ∼70% during the period of prophylaxis. But continued HIV transmission after prophylaxis stops suggests more prolonged infant prophylaxis is needed.

From the *Department of Epidemiology, Johns Hopkins University Bloomberg School of Public Health, Baltimore, MD; †Inherited Disease Branch, National Human Genome Research Institute, National Institutes of Health, Baltimore, MD; ‡Department of Statistics and biostatistics and Institute for Health, Health Care Policy and Aging Research, Rutgers University, Piscataway, NJ; §Center for Research for Mothers and Children, College of Medicine-Johns Hopkins University Research Project, Blantyre, Malawi; ‖Division of HIV/AIDS Prevention, National Center for HIV, Viral Hepatitis, STD and TB Prevention, Centers for Disease Control and Prevention, Atlanta, GA; ¶Department of Internal Medicine, College of Medicine, University of Malawi, Blantyre, Malawi; #Johns Hopkins University School of Medicine, Baltimore, MD; and **Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD.

Received for publication December 14, 2010; accepted February 18, 2011.

Supported by a Cooperative Agreement (# 5 U50 PS022061-05; Award # U50/CC0222061) from the Centers for Disease Control and Prevention and the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health. Qing Li was supported in part by the Intramural Research Program of the National Human Genome Research Institute, National Institutes of Health.

Presented in part at the XVIII International AIDS Conference; Vienna, Austria; July 18-23, 2010. Abstract # THPE0161; and at the 2nd International Workshop on HIV Pediatrics; Vienna, Austria; July 16-17, 2010. Abstract # O_04.

The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the US Centers for Disease Control and Prevention or the National Institutes of Health. Use of trade names is for identification purposes only and does not constitute endorsement by the US Centers for Disease Control and Prevention, National Institutes of Health, or the Department of Health and Human Services.

Clinical Trial Registration number: NCT00115648.

Correspondences to: Dr Taha E. Taha, MBBS, PhD, Rm E7138, Department of Epidemiology, Bloomberg School of Public Health, Johns Hopkins University, 615 N. Wolfe Street, Baltimore, MD 21205 (e-mail

Back to Top | Article Outline


We previously reported that extended antiretroviral drugs to age 14 weeks substantially decreased HIV-1 transmission from infected mothers to their infants through breastfeeding in the postexposure prophylaxis of infants in Malawi (PEPI-Malawi) trial.1 At the primary endpoint of 9 months, postnatal infection occurred in 5.2% and 6.4% of infants not infected at birth in the 2 extended arms of nevirapine (NVP) and NVP plus zidovudine (ZDV), respectively, compared with 10.6% in the control arm that received single-dose NVP plus 1 week of ZDV. We also showed that transmission continued after cessation of prophylaxis among breastfed infants with approximately 1%-2% increase every 3 months to age 24 months with no difference in the increase among the study arms.2 We update and extend these previous results (based on infants followed to 2007) to 2009 when the PEPI study follow-up was completed.

Back to Top | Article Outline


Study Population, Enrollment and Follow-Up

PEPI-Malawi study details were previously reported.1 Briefly, PEPI was a randomized, open-label, clinical trial that started enrolling in April 2004 and completed follow-up in September 2009 in Blantyre, Malawi. Study inclusion criteria included maternal age 18 years or older, able and willing to give informed consent for HIV testing and enrollment, willing to receive HIV test results, HIV infected, planning to deliver or gave birth at one of the study clinics, willing to return for follow-up visits for up to 2 years postnatally, and resident of Blantyre city or its suburbs. Exclusion criteria included being HIV uninfected, not planning to breastfeed at the time of delivery, inability or unwillingness to follow any of the inclusion requirements, newborn had a life-threatening condition, and being previously enrolled in this trial with an earlier pregnancy.

A signed informed consent for HIV testing and screening for the study was sought from eligible mothers, and a separate signed consent was obtained for enrollment. Infants of HIV-infected breastfeeding mothers were randomized at birth to receive 1 of 3 regimens as follows: single-dose oral NVP plus 1 week of daily oral ZDV (control); control plus extended oral daily NVP (ExtNVP) from day 8 to 14 weeks of age; and control plus extended oral daily NVP plus ZDV (ExtNVP/ZDV) from day 8 to 14 weeks of age. Infant dosing was based on weight at each visit.1 Permutated block algorithms with a 1:1:1 allocation ratio were employed with block sizes of 9 and 12 and blocked within study site. Mothers received intrapartum NVP single dose (orally 200 mg) depending on the time of HIV counseling and testing. Mothers were counseled to stop breastfeeding by 6 months based on the World Health Organization recommendations at the time when the trial started. Postnatal follow-up visits were at 1, 3, 6, 9, and 14 weeks, and 6, 9, 12, 15, 18, and 24 months of infant age. Demographic, obstetric, and medical history were obtained at enrollment, and delivery/newborn information was collected before discharge from study clinics. Breastfeeding history, clinical history, assessment of prophylaxis adherence, monitoring of adverse events, and physical examination reports were obtained at follow-up visits. Severity of adverse experiences was graded using standard DAIDS Toxicity Tables.3 Further details of the study conduct and monitoring procedures of the trial are included in an earlier report.1 The NIAID/NIH Vaccine and Prevention Data and Safety Monitoring Board monitored this trial for safety data and critical efficacy endpoints and made recommendations as appropriate. The PEPI study was approved by institutional review boards at the University of Malawi, Johns Hopkins University, and the Centers for Disease Control and Prevention.

Back to Top | Article Outline

Study Specimen Collection and Laboratory Tests

Details of sample collection and procedures for specific tests were described in an earlier report.1 All women were tested for HIV at screening before being consented for enrollment into the study using established procedures. Maternal blood samples for complete blood count, CD4 cell count, and plasma viral load measurement was collected at scheduled visits. Infant blood samples for safety assessments and HIV testing were collected at each visit. Longitudinal infant HIV testing was performed at the College of Medicine-Johns Hopkins University Research Laboratory in Blantyre (Virology Quality Assurance program monitored laboratory). Positive specimens were confirmed by a second test on a separate specimen no later than the next scheduled visit and sooner whenever possible. Discrepant test results at the local laboratory were retested in a reference laboratory in the United States. An infant was considered HIV infected based on positive HIV DNA polymerase chain reaction assay at any visit, and/or positive enzyme-linked immunosorbent assay and Western blot tests at 15 months or later. An infant with at least 2 positive HIV test results on separate visits was classified as definitively infected. An infant was classified as confirmed HIV uninfected if none of the HIV test results was positive up to the last visit available. The PEPI analysis was restricted to infants not HIV infected at birth. Therefore, infants found HIV infected or with undetermined HIV status at birth were excluded.1

Back to Top | Article Outline

Statistical Analyses

This analysis includes all data collected to the end of study follow-up in September 2009. The primary study outcome is infant HIV infection at 9 months among those who were not infected at birth. In an intent-to-treat analysis, rates of HIV infection from Kaplan-Meier curves were compared in each of the extended prophylaxis arms (ExtNVP or ExtNVP/ZDV) versus the control arm. We also compared rates of death and of either HIV infection or death. Cox proportional hazard models examined associations of extended treatment regimens with HIV infection and either HIV infection or death after adjusting for other factors. Treatment arms were compared with each other by levels of baseline covariates using t tests, exact tests, Analysis of Variance and χ2 tests as appropriate. The estimated protective efficacy of the extended treatment arms at each visit was calculated using the formula 100 × [Proportion infected in the control arm)-(Proportion infected in the extended treatment arm)]/(Proportion infected in the control arm)].1

Back to Top | Article Outline


Figure 1 shows the PEPI-Malawi study profile. Of 46,186 women consented and screened for HIV in participating clinics, 9818 were HIV infected (prevalence 21%). Overall, 3397 infants (born to 3334 mothers) were enrolled in the study and randomized to control (n = 1090), ExtNVP (n = 1160), or ExtNVP/ZDV (n = 1147) arms. Excluding infants found HIV infected or with undetermined HIV infection at birth, 3126 (1004 controls, 1071 extended NVP, 1051 extended NVP + ZDV) had data included in the primary analysis. Of these, 215 control, 253 extended NVP and 237 extended NVP + ZDV were lost before the last visit. There were no baseline differences by study arm in maternal age, CD4 count or presentation time/delivery mode or in infant gender, birth weight, adherence to regimen or breastfeeding before discharge.1,2



Among infants who were HIV uninfected at the visit before the last available visit, the frequency of breastfeeding was high in all visits up to 6 months—ranging from ∼100% to 88% in all study arms. Approximately 96% of the women in each study arm reported exclusively breastfeeding up to 6 months postpartum. After 6 months, there were substantial reductions in the frequency of breastfeeding in the 3 arms (eg, at 18 months, the frequency of breastfeeding in control arm was 15.0%, 11.4% in ExtNVP arm, and 12.6% in ExtNVP/ZDV arm, P = 0.14).

Of the 3126 infants uninfected at birth and included in the primary analysis, 136 in the control arm, 95 in the ExtNVP arm, and 98 in the ExtNVP/ZDV arm were found HIV infected at the completion of follow-up. Rates of HIV infection, death, and either HIV infection or death are shown in Table 1; and Kaplan-Meier curves for each of these outcomes stratified by the intervention arm (control, ExtNVP, ExtNVP/ZDV) are in Figures 2A-C. At the primary endpoint of age 9 months, HIV infection had occurred in 5.0% [95% confidence interval (CI): 3.8 to 6.6] of the ExtNVP and 6.0% (95% CI: 4.7 to 7.7) of the ExtNVP/ZDV arm infants; both were significantly (P < 0.001) lower compared with 11.1% (95% CI: 9.3 to 13.3) of control arm infants (Fig. 2A). At age 24 months, there was ∼4%-5% absolute increase in overall HIV infection for each extended study arm over the infection rate at 9 months, but the rates were still significantly lower in the extended prophylaxis arms than in the control arm (10.8% and 11.2% in ExtNVP and ExtNVP/ZDV arms, respectively, compared with 15.6% in control; P = 0.003 comparing ExtNVP with control and P = 0.008 comparing ExtNVP/ZDV with control). With the exception of 1 week, HIV infection rates in the extended prophylaxis arms were significantly lower than in control arm at all ages (Table 1). There were no statistical differences between the ExtNVP and ExtNVP/ZDV arms although the study was not powered to detect differences between these arms. The rates of either HIV infection or death were also significantly lower in the extended prophylaxis arms compared with control arm at all ages beyond the first week (Table 1 and Fig. 2B). The rates of death were consistently lower in the extended prophylaxis arms than in control arm; however, the differences were not statistically significant at each visit throughout the study (Fig. 2C).





The estimated protective efficacy of ExtNVP and ExtNVP/ZDV at the end of the period of the intervention (age 14 weeks) was ∼70% compared with the control. After cessation of the intervention, the protective efficacy declined in both extended prophylactic arms as follows: at 6, 9, 12, 15, 18, and 24 months the respective estimates in the ExtNVP were 62%, 55%, 47%, 45%, 37%, and 31% and the respective estimates in the ExtNVP/ZDV were 54%, 46%, 39%, 39%, 35%, and 20%.

The reductions in HIV infection associated with ExtNVP and extNVP/ZDV regimens compared with control regimen remained statistically significant after controlling for other factors (Table 2). Other factors associated with HIV transmission were baseline (enrollment) maternal CD4 cell count and presentation of mother at delivery. Decrease in CD4 cell count and late presentation (arrival) at time of delivery (which prevented the mother from receiving single-dose NVP prophylaxis during labor) increased the risk of transmission of HIV from mother to infant after controlling for prophylaxis regimen and other factors. Extended prophylaxis of the infant was also associated with lower risk of HIV infection or death (Table 2). In addition, increase in birth weight was associated with lower risk of HIV or death (ie, better survival). As with HIV infection, low maternal CD4 cell count was associated with increased risk of HIV infection or death.



Reported infant adverse events were not significantly (P > 0.05) different among the 3 study arms. There were no differences in grades 3 or 4 neutropenia toxicity or grades 2 or 3 hemoglobin toxicity levels by treatment arm from birth to 24 months. Table 3 shows there were no statistically significant differences in number of not related or probably related serious adverse events (SAEs). However, the frequency of possibly related SAEs were significantly (P = 0.01) higher in the ExtNVP/ZDV arm than in the ExtNVP or control arms.



Back to Top | Article Outline


The updated results of the PEPI-Malawi trial confirm earlier findings when the data and safety monitoring board recommended stopping of enrollment after the second interim analysis in December 2007 because of statistical difference between the extended NVP arm and the control arm.1 Since then data on 110 infants were added to the primary analysis (at age 9 months) and the 24 months duration from date of birth to end of the study for all infants was reached in September 2009. The updated results continue to demonstrate that infant extended antiretroviral prophylaxis significantly reduces the risk of HIV infection and HIV infection or death. The absolute reduction in HIV transmission of 5%-6% observed in the extended NVP and NVP + ZDV arms at the primary endpoint roughly continued out to 24 months indicating that there was no catch up in HIV infection after the interventions had been withdrawn.2 The protective efficacy was thus ∼70% during the period of prophylaxis, but then gradually declined demonstrating the continued risk of transmission as long as the child is breastfed. The effectiveness by 24 months (end of the study) was 20%-31%.

The safety of extended NVP prophylaxis compared with the short prophylactic control regimen is reassuring. With extended NVP/ZDV regimen, however, the frequency of possibly related SAEs was higher than in other arms. Further discussion of safety data from this study is included in an earlier report.1 Preliminary data from the PEPI-Malawi study on acquisition of NVP resistance shows that the dual regimen of NVP/ZDV can reduce development of resistance if the infection is diagnosed and the NVP/ZDV regimen stopped before 6 weeks of age.4 Further laboratory work on resistance is currently in progress.

In the African setting where replacement feeding is not safe or feasible, breastfeeding is critical to infant survival. Early cessation of breastfeeding can expose the infant to severe morbidity and may lead to infant death, as has been reported during the conduct of the PEPI trial in Malawi5 and elsewhere in Africa.6-8 Therefore, an extended infant antiretroviral prophylaxis strategy provides a safe and effective option to reduce mother-to-child transmission of HIV to breastfed babies. The efficacy of extended antiretrovirals to prevent breastfeeding transmission is further confirmed by another randomized trial in Malawi, The Breastfeeding, Antiretrovirals and Nutrition (BAN) study.9 This trial enrolled HIV-infected pregnant women with CD4 >250 cells/ per microliter; the breastfeeding mother/infant pairs were randomized to a control regimen of single-dose NVP plus 1 week of ZDV/3TC, extended infant NVP prophylaxis through age 6 months, or maternal triple antiretroviral drug prophylaxis through age 6 months. The rate of postnatal infection at 28 weeks in infants uninfected at age 2 weeks was 1.7% in the infant extended NVP arm and 2.9% in the maternal prophylaxis arm; both rates were significantly lower than in the control arm (5.7%). There was not a statistically significant difference between the transmission rates in the 2 extended arms, although the study was not designed to compare the 2 experimental arms. In Botswana, a randomized trial (the Mma Bana trial) further showed that providing antiretrovirals during pregnancy and continued postpartum to 6 months for women with CD4 ≥200 cells per cubic millimeter can reduce overall transmission to 1.1% at age 6 months.10 The BAN study results suggest that in settings such as Malawi, where women present late for delivery and do not receive antepartum antiretrovirals (for the viral load to be adequately reduced by the time breastfeeding starts), the infant NVP prophylaxis may provide a more immediate effect.11

The World Health Organization recently (November 2009) recommended 2 alternative options for HIV-infected women who continue breastfeeding and are not taking antiretroviral treatment.12 These are either (1) infant daily NVP prophylaxis for women who received ZDV during pregnancy or (2) continue a maternal regimen of 3-drug antiretroviral prophylaxis postpartum until the end of breastfeeding for women who received a 3-drug regimen during pregnancy. The effectiveness of these regimens is not yet assessed in a clinical trial but is based on the findings from PEPI, BAN, and other trials.1,9,10,13 A large multisite clinical trial, the PROMISE study (Clinical Trial # NCT01061151), is scheduled to start soon to compare these 2 regimens.

The analysis of the efficacy data from the PEPI-Malawi trial suggests some areas that need further assessment pertaining to HIV transmission through breastfeeding. For example, HIV infection rates in Figure 2A show an initial substantial reduction before 14 weeks of age: at 9 weeks, the rates of postnatal HIV infection are 7.6% and 2.4% in control and extended NVP arms, respectively. This is consistent with data from previous studies conducted in Malawi and Kenya before the introduction of antiretroviral prophylaxis which showed high rates of transmission in early compared with late postnatal periods.14,15 This early benefit seems to be maintained throughout the study. Additionally, the rate of transmission after age 6 months when reported breastfeeding frequency becomes low needs further investigation. It is likely that misclassification of reported breastfeeding is high or there are other factors contributing to transmission that are not well understood, including mastitis or the feeding of premasticated food as reported in other studies.16,17 In this study, most of the loss to follow-up occurred late after the primary endpoint of the trial at 9 months, and the overall losses were comparable among the 3 study arms.

Survival of infants in resource-constrained settings is a major concern, especially in areas where HIV prevalence continues to be high among women of reproductive age. Both prevention of HIV transmission and maintenance of adequate infant nutrition should be maximized. Breastfeeding, however, increases the risk in one and protects against morbidity and mortality for the other. The PEPI-Malawi trial results show that infant extended antiretroviral prophylaxis is a practical effective approach to reduce HIV transmission and improve HIV-free survival. Further assessment to monitor effectiveness and safety throughout the period of breastfeeding is urgently needed.

Back to Top | Article Outline


We are indebted to the mothers and children who participated in the PEPI-Malawi study. We are grateful to the nursing and technical staff in Malawi for their excellent collaboration throughout this study.

Back to Top | Article Outline


1. Kumwenda NI, Hoover DR, Mofenson LM, et al. Extended antiretroviral prophylaxis to reduce breast-milk HIV infection. N Engl J Med. 2008;359:119-129.
2. Taha TE, Kumwenda J, Cole SR, et al. Postnatal HIV-1 transmission after cessation of extended infant antiretroviral prophylaxis and effect of maternal HAART. J Infect Dis. 2009;200:1490-1497.
3. National Institutes of Health. Division of AIDS table for grading the severity of adult and pediatric adverse events. Version 1.0, December 2004; Clarification August 2009. Available at: Accessed August 2, 2010.
4. Lidstrom J, Li Q, Hoover DR, et al. Addition of extended zidovudine to extended nevirapine prophylaxis reduces nevirapine resistance in infants who were HIV-infected in utero. AIDS. 2010;24:381-386.
5. Kafulafula G, Hoover DR, Taha TE, et al. Frequency of gastroenteritis and gastroenteritis-associated mortality with early weaning in HIV-1-uninfected children born to HIV-infected women in Malawi. J Acquir Immune Defic Syndr. 2010;53:6-13.
6. Onyango-Makumbi C, Bagenda D, Mwatha A, et al. Early weaning of HIV-exposed uninfected infants and risk of serious gastroenteritis: findings from two perinatal HIV prevention trials in Kampala, Uganda. J Acquir Immune Defic Syndr. 2010;53:20-27.
7. Homsy J, Moore D, Barasa A, et al. Breastfeeding, mother-to-child transmission, and mortality among infants born to HIV-infected women on highly active antiretroviral therapy in rural Uganda. J Acquir Immune Defic Syndr. 2010;53:28-35.
8. Creek TL, Kim A, Lu L, et al. Hospitalization and mortality among primarily non-breastfed children during a large outbreak of diarrhea and malnutrition in Botswana, 2006. J Acquir Immune Defic Syndr. 2010;53:14-19.
9. Chasela CS, Hudgens MG, Jamieson, et al. Maternal or infant antiretroviral drugs to reduce HIV-1 transmission. N Engl J Med. 2010;362:2271-2281.
10. Shapiro RL, Hughes MD, Ogwu A, et al. Antiretroviral regimens in pregnancy and breast-feeding in Botswana. N Engl J Med. 2010;362:2282-2294.
11. Mofenson LM. Protecting the next generation—eliminating perinatal HIV-1 infection. N Engl J Med. 2010;362:2316-2318.
12. World Health Organization. New recommendations: preventing mother-to-child transmission. Available at: Accessed April 8, 2011. Accepted 2009.
13. de Vincenzi, Kesho Bora Study Group. Triple-antiretroviral (ARV) prophylaxis during pregnancy and breastfeeding compared to short-ARV prophylaxis to prevent, other-to-child transmission of HIV-1 (MTCT): the Kesho Bora randomized controlled clinical trial in five sites in Burkina Faso, Kenya and South Africa. Presented at: 5th International Conference on HIV Pathogenesis Treatment and Prevention; July 19-22, 2009; Cape Town, South Africa. Abstract # LBPEC01.
14. Miotti PM, Taha TET, Kumwenda NI, et al. HIV transmission through breastfeeding, a study in Malawi. JAMA. 1999;282:744-749.
15. Nduati R, John G, Mbori-Ngacha D, et al. Breastfeeding and formula feeding on transmission of HIV-1. A randomized clinical trial. JAMA. 2000;283:1167-1174.
16. Semba RD, Kumwenda N, Hoover DR, et al. Human immunodeficiency virus load and breastmilk, mastitis, and mother-to-child transmission of human immunodeficiency virus type 1. J Infect Dis. 1999;180:93-98.
17. Gaur AH, Dominguez KL, Kalish ML, et al. Practice of feeding premasticated food to infants: a potential risk factor for HIV transmission. Pediatr. 2009;124:658-666.

breastfeeding; infant postexposure prophylaxis; mother-to-child transmission of HIV; PEPI-Malawi

© 2011 Lippincott Williams & Wilkins, Inc.