Mother-to-child transmission (MTCT) of HIV-1 is a major public health problem in developing countries where approximately 420,000 children were newly infected in 2007.1 MTCT can occur prenatally, peripartum, and during breastfeeding. The monthly probability of late HIV transmission through breastfeeding after the age of 1 month is estimated to be 6.5-8.5 transmissions per 1000 exposed infants.2 Prolonged breastfeeding into the second year of life presents a cumulative risk of HIV infection from breastfeeding of 12%-14%, accounting for as much as 42% of overall HIV transmission in breastfed infants.2-5 Conversely, weaning by 6 months may prevent more than 85% of transmission due to breastfeeding.6
Although the most effective method to eliminate breastfeeding-associated HIV transmission is not to breastfeed, this approach is not feasible or safe for most HIV-infected women in resource-constrained settings because of cost, unsafe water supply, and low acceptability due to stigma associated with not breastfeeding. Not breastfeeding also deprives the baby of essential nutrition and immunologic defenses in breast milk, which are crucial to infant survival. Based on general infant and child survival literature, not breastfeeding in low-resource settings where water supplies may not be safe and sustainable replacement feeding is not possible, places the infant at increased risk of mortality [primarily related to malnutrition and infections such as severe gastroenteritis (GE)]-with a 2.6- to 5.8-fold and a 1.4- to 1.8-fold increased risk of mortality in the first 6 and, in the second 6 months of life, respectively.7 Therefore, intervention strategies to make breastfeeding safer in HIV-infected women are being considered, and several trials using antiretroviral drugs to interrupt transmission during breastfeeding are now being conducted to reduce breastfeeding-associated HIV transmission.8 These trials are predicated on providing interventions, such as maternal or infant antiretroviral prophylaxis, to allow breastfeeding during the critical first 6 months of life, followed by early weaning after the intervention is stopped.
A previous randomized trial in Malawi (the nevirapine/AZT [NVAZ] trial) showed that short-course (1 week) infant antiretroviral prophylaxis can reduce MTCT of HIV.9,10 Based on the positive results of the NVAZ trial, a second randomized trial was undertaken in Malawi extending infant antiretroviral prophylaxis to 14 weeks of age. In this trial [Postexposure Prophylaxis to the Infant (PEPI)], based on 2003 guidelines from the Malawi Ministry of Health, all women were encouraged to exclusively breastfeed for 6 months, and then cease breastfeeding at 6 months, rather than continue breastfeeding out to age 2 years as is typically recommended for uninfected women.11 This strategy of exclusive breastfeeding followed by early breastfeeding cessation was recommended by World Health Organization (WHO) in 2000 to minimize the risk of HIV transmission through breastfeeding while maximizing the benefits of breast milk during the first 6 months of life.12
However, emerging data from several ongoing HIV MTCT prevention trials13,14 suggest that early cessation of breastfeeding may place an HIV-exposed but uninfected infant at increased risk for severe GE and mortality. Given this concern and the high background mortality rates in infants and children under 5 years of age in Malawi,15 we compared morbidity and mortality data in the ongoing PEPI trial, which included counseling for early breastfeeding cessation at 6 months, with data from the earlier NVAZ trial conducted in 2000-2003 at the same site clinics in Blantyre, Malawi. The option of cessation of breastfeeding by 6 months was part of the study design in PEPI based on implementing the national and WHO recommendations12 as standard of care in a study setting. In contrast, the NVAZ trial was conducted before the Malawi Ministry of Health implemented the WHO recommendations for early breastfeeding cessation at 6 months by HIV-infected mothers. Thus, the HIV-infected women in NVAZ were not counseled to stop breastfeeding early; therefore they followed the usual infant feeding practices in Malawi which includes prolonged breastfeeding of greater than a year in most cases.
In this analysis, we compare rates of GE-related serious adverse events and mortality among HIV-exposed, uninfected infants in the PEPI and NVAZ trials.
Study Design, Follow-up and Procedures
PEPI and NVAZ were both randomized, controlled, open-label phase III clinical trials to prevent MTCT of HIV. Both were conducted at the Queen Elizabeth Central Hospital and health centers within Blantyre, Malawi. Each trial was reviewed and approved by the College of Medicine Research and Ethics Committee, University of Malawi, and the Johns Hopkins Bloomberg School of Public Health Committee on Human Research. The PEPI trial was also approved by the Institutional Review Board at the Centers for Disease Control and Prevention, one of the sponsoring organizations. Written informed consent for enrollment of infants was obtained from their mothers in both trials when they came for delivery or up to 24-hours postpartum.
In both trials, data were collected on all clinical reports of GE, hospitalizations, and mortality. GE was defined as the acute onset of diarrhea, vomiting, and associated features of fever, poor feeding, lethargy, and dehydration. Serious GE events were defined as those requiring hospitalization and/or intravenous hydration. Infant deaths were considered GE-related if they were immediately preceded by a GE event or a GE-related hospitalization.
The PEPI trial,16 with a total sample size of 3500, began enrollment in April 2004; the 24-month follow-up of all participants will be completed in September 2009. HIV-exposed infants were randomized to 3 study arms at birth. Infants in the control arm received oral single-dose nevirapine (NVP, 2 mg/kg body weight once) plus oral zidovudine (AZT, 4 mg/kg body weight twice a day) for 1 week (the regimen shown to be effective in the NVAZ study). Infants in the second arm received the same regimen as those in the control arm for 1 week, followed by daily oral NVP to age 14 weeks (dosed as 2 mg/kg body weight once daily from weeks 1 through 2 and 4 mg/kg body weight once daily for weeks 3 through 14). Infants in the third arm received the same regimen as those in the control arm and followed by daily oral NVP (dosed as above) and oral AZT (dosed at 4 mg/kg body weight twice daily from weeks 1 through 5, 4 mg/kg body weight 3 times daily from weeks 6 through 8, and 6 mg/kg body weight 3 times daily from weeks 9 through 14) to age 14 weeks.
Infants' blood samples for HIV DNA PCR testing were obtained at the time of randomization (at birth). At the 1-week visit, infants found to have a positive birth DNA PCR had study drug discontinued if in one of the extended prophylaxis arms; both HIV-infected and uninfected infants were followed-up through the duration of the study. Consistent with the Malawi Ministry of Health and WHO recommendations, all mothers were encouraged to exclusively breastfeed for 6 months followed by early weaning at 6 months of age. Mothers of infants with confirmed HIV infection before 6 months of age were counseled against early weaning.
In the PEPI trial, infants and their mothers were seen for study visits within 24 hours of delivery and at ages 1, 3, 6, 9, and 14 weeks, and 6, 9, 12, 15, 18, and 24 months. Infant blood samples were collected by heel-stick or venous puncture for HIV testing and dried blood spot storage at each visit except week 3; complete blood count and alanine aminotransferase were collected at each visit through 6 months; and plasma was stored at birth, 6 and 14 weeks and 6, 12, and 18 or 24 months. Infant HIV diagnosis was based on DNA PCR HIV testing on whole blood (Roche Amplicor 1.5 DNA PCR, Roche Molecular Systems, Pleasanton, CA); HIV serology was also used if infant was 15 months or older. HIV infection was confirmed if an infant had at least 2 positive HIV test results on separate visits. Mothers were encouraged to report to the study clinic for any interim illnesses. Using standard questionnaires (similar to those used in NVAZ) detailed information on breastfeeding (exclusive or combined with other liquids/foods), health status of the infant, adverse events (since the last visit) was collected at each visit.
During each study visit, women received individual counseling and education by trained research nurses about exclusive breastfeeding for the first 6 months of life and replacement and complementary feeding options (commercial infant formula, locally available fresh cow's milk) after weaning of the baby, based on Malawi Ministry of Health guidelines. The nurses provided mothers, at each visit, information on potential benefits and risks of breastfeeding by HIV-infected women, advantages and potential risks of replacement feeding; and proper breastfeeding techniques and breast and nipple care. Mothers were encouraged to choose an infant feeding option most suitable to the family.
Three steps were followed to support mothers and infants if they chose early weaning. Firstly, mothers were instructed to cup-feed the infant with expressed breast milk from the age of 5 months. Secondly, at 6 months postpartum mothers were instructed to change from expressed breast milk to commercial formula or full cream powder milk to transition from breast milk to semisolid food (porridge). Thirdly, they were instructed to eventually feed the infants with soy and peanut based locally available enriched porridge called Likuni Phala as the replacement food. Mothers were encouraged to introduce locally available and home-prepared fruit (eg, mango, banana) juices to complement the replacement feeding. The study provided both the full cream milk as transition feeds and the Likuni Phala for replacement feeding. Mothers who did not have the appropriate set of clean cutlery to accomplish hygienically safe replacement feeding were provided with the cutlery through the study and instructed on the hygienic care of the cutlery to minimize the risk of GE. In June 2006, information regarding increased risk of GE-associated mortality after early weaning was incorporated in the informed consent form as recommended by the Data Safety and Monitoring Board after the first interim analysis data review.
In PEPI, women who require antiretroviral therapy for their own health, based on Malawi treatment guidelines, were referred to a government-sponsored antiretroviral clinic for treatment. Infant study antiretroviral prophylaxis was discontinued if an infant was confirmed to be HIV infected before the completion of prophylaxis regimen. HIV-exposed infants were provided with cotrimoxazole from age 6 weeks through 3 months after weaning. In contrast, in NVAZ, cotrimoxazole was not provided to HIV-exposed infants because at the time of the conduct of NVAZ this policy was not yet in the national guidelines for the care of the HIV-infected or exposed infant. Other clinical care was provided at the study clinics at no cost to the participants or through referral to specialized clinics at the Queen Elizabeth Central Hospital.
The details and results of the NVAZ trial were reported previously.9,10 Infants enrolled in the NVAZ trial (N = 2000) were randomized to receive either oral single-dose NVP or single-dose NVP plus AZT orally twice daily for 1 week, administered in the same dosing schedule as in PEPI. The mothers could have received intrapartum single oral dose NVP if they presented early (more than 4 hours before anticipated delivery) or did not receive NVP if they presented later for delivery. Infants and their mothers returned for follow-up visits at ages 1 and 6 weeks and 3, 6, 9, 12, 15, 18, and 24 months. Mothers were also encouraged to bring their infants to the study clinic for any interim illnesses. At each follow-up visit, using standard questionnaires, information was collected about maternal and infant health, breastfeeding, and adverse events (including GE events and hospitalizations) since the last visit.
The case finding for GE events (reported by the mother and/or documented in the infants health passport for ambulatory and inpatient care) and definition of GE as stated above were the same in the 2 trials. Standardized toxicity evaluations, based on DAIDS Toxicity Tables, were carried out in both studies. During the period of the NVAZ trial antiretroviral therapy for infected mothers and infants and infant cotrimoxazole prophylaxis to HIV exposed but uninfected infants were not available in Malawi.
Infants who were identified to have HIV infection at birth or at a subsequent visit (ie, had a positive virologic test at birth or at a later visit) were excluded from the current analysis. This analysis was restricted to children enrolled in the PEPI and NVAZ trials who were HIV-uninfected at the time of the specific study visit under consideration. The analysis included 1810 mother-infant pairs from the NVAZ trial, which has completed follow-up. Data up to August 7, 2007 were included for the PEPI trial, which is still ongoing. The PEPI trial subset included a total of 2671 infants enrolled before August 7, 2006 and therefore have had at least 1 year of follow-up data through August 7, 2007, which was the data freeze date for this analysis. Six hundred and thirty-six infants were excluded (182 were HIV-infected, 300 were lost to follow-up, and 154 had died HIV-uninfected before the last visit in the specific interval under consideration), therefore only 2035 PEPI infants are included in this analysis. We combine data from the 3 treatment arms in the PEPI trial and from the 2 treatment arms in the NVAZ trial.
We examined age-specific frequency of overall GE and serious GE-associated events reported or identified as happening since the last scheduled visit at each of the scheduled follow-up visits in both the NVAZ and PEPI trials among infants who were HIV-uninfected at the particular study visit. The GE and GE-associated hospitalization frequencies were evaluated as interval probabilities with intervals including ages 0-6 weeks, 7 weeks-3 months and 4-6, 7-9, and 10-12 months. An infant was counted only once in each age interval, and the cross-sectional data only include infants with a negative HIV test (HIV DNA PCR if age <15 months or HIV antibody EIA/western blot if ≥15 months) at or after the last visit in the interval. Infants are excluded from the analysis if they failed to attend the last visit for the interval unless a GE adverse event was identified and documented to have occurred in the interval between study visits. In the mortality analyses, probability of death was analyzed using the Kaplan-Meier method, where the time to death is the response variable, with censoring at the time of HIV infection, loss to follow-up, or exit from the trial. All calculations were done in SAS version 9.1 (Gary, NC).
Maternal Demographic Characteristics
Table 1 compares selected maternal demographic characteristics at baseline of the NVAZ and PEPI trials. This analysis included 1810 mother-infant pairs from the NVAZ trial and 2035 mother-infant pairs from the PEPI trial. There were some statistically significant baseline differences between the 2 groups of women. For example, women who enrolled in the NVAZ trial, on average, were younger, had fewer living children and were less educated than women enrolled in the PEPI trial. Women who enrolled in the PEPI trial, compared with those who enrolled in the NVAZ trial, were more likely to have electricity at home but less likely to have running water. A larger proportion of mothers in the NVAZ trial had medical problems (eg, respiratory tract infections, anemia) during pregnancy compared to mothers in the PEPI trial.
Duration of Breastfeeding
Median duration of breastfeeding among surviving infants not infected with HIV was longer than 24 months in the NVAZ trial, with 89% of infants still breastfeeding at 9 months and 60% were still breastfeeding at 24 months. By contrast, in the PEPI trial, 80% of HIV-uninfected infants were weaned between 6 and 9 months (90% breastfeeding at 6 months and 10% breastfeeding at 9 months); only 3% were still breastfeeding at 24 months.
Frequency of Reported GE
In the PEPI trial, among the HIV-uninfected infants, the probability of experiencing one or more GE events of any severity was 0.6% (11/1750) from birth to 6 weeks, 1.6% (26/1635) from 7 weeks to 3 months, 4.1% (63/1546) from 4 to 6 months, 9.0% (129/1427) from 7 to 9 months, and 5.6% (70/1246) from 10 to 12 months (Table 2). In the NVAZ trial, among HIV-uninfected infants the probability of experiencing one or more GE events was 0.3% (5/1469) from birth to 6 weeks, 1.9% (25/1302) from 7 weeks to 3 months, 6.4% (79/1239) from 4 to 6 months, 9.2% (107/1168) from 7 to 9 months, and 8.2% (88/1097) from 10 to 12 months (Table 2). Seven to 9 months was the peak interval for GE events in both trials. The number of overall GE events of any severity in the PEPI trial during the 4-6 and 10-12 months intervals was lower than in the NVAZ trial (P = 0.01 and P = 0.02, respectively); otherwise there were no significant differences in overall GE events between the trials at other time intervals.
Frequency of GE Hospitalization
In contrast to overall rates of GE-associated events, the rates of serious GE events (hospitalizations, deaths) were significantly increased in the PEPI trial compared with the NVAZ trial. The probability of GE-associated hospitalization was significantly greater in the PEPI trial than NVAZ trial after 6 months of age (Table 2). The frequency of GE-associated hospitalizations during the age interval of 7-9 months was 2.9% (41/1427) in the PEPI trial compared with 0.1% (1/1158) in the NVAZ trial (P < 0.001) and for the age interval 10-12 months, it was 1.6% (20/1246) in PEPI compared with 0.2% (2/1097) in NVAZ (P < 0.001).
Mortality and GE Mortality
In the PEPI trial, of the 2035 infants who were HIV-uninfected at birth and born before August 7, 2006, 251 uninfected infants died during the follow-up period. Seventy-two (28.7%) of these deaths were GE associated. In the NVAZ trial, of the 1810 infants who were HIV uninfected at birth,106 uninfected infants died during follow-up. Seventeen (16%) of these deaths were GE associated. Figure 1 (panels A and B) shows Kaplan-Meier curves for cumulative overall and GE-associated infant mortality of HIV-uninfected infants (ie, censored at HIV infection if found to be infected) in the 2 trials. These figures show increasing probabilities of overall mortality and GE-associated mortality in the PEPI trial compared with the NVAZ trial after age 6 months, when breastfeeding ceased in the vast majority of infants in the PEPI trial.
Through age 6 months, the cumulative overall mortality probabilities in uninfected children were not significantly different between the PEPI and NVAZ trials, 35 and 40 per 1000 uninfected infants, respectively. However, by 9 months (the 3-month period during which most weaning occurred in the PEPI trial), cumulative overall mortality probabilities in uninfected infants were 62 and 54 per 1000 uninfected infants for PEPI and NVAZ trials, respectively. The cumulative overall mortality at age 12 months was significantly higher in uninfected infants in the PEPI trial than the NVAZ trial (79 vs. 66 per 1000 uninfected infants, respectively, P = 0.03).
Through age 6 months, the cumulative GE-associated mortality probabilities in uninfected children were 5 and 3 per 1000 uninfected children for the PEPI and NVAZ trials, respectively. By age 9 months, cumulative GE-associated mortality rose to 19 and 7 per 1000 uninfected children for the PEPI and NVAZ trials, respectively, and the cumulative GE-associated mortality at age 12 months was significantly higher in uninfected infants in the PEPI trial compared with the NVAZ trial (24 vs. 12 per 1000 uninfected children, P = 0.0002).
We compared the probabilities of overall GE events, GE-associated hospitalization events, and overall mortality and GE-associated mortality for HIV-uninfected infants in Blantyre, Malawi, in 2 randomized clinical trials. The trials were conducted at the same clinics at 2 different periods when different recommendations related to infant feeding and weaning were followed. In one trial, NVAZ, the vast majority of infants were breastfed into the second year of life; in the other trial, PEPI, the majority of infants was breastfed for the first 6 months of life and then ceased breastfeeding early, between ages 6 and 9 months, as per Malawi and WHO guidelines. In PEPI, mothers received intensive counseling at each visit about exclusive breastfeeding for the first 6 months of life, optimal breast care, and instructions how to wean the infant, and were provided weaning food by the study if needed (see “Methods” section above). The GE outcomes under the age of 6 months were low and similar in the 2 studies most likely is related to protective effects of a predominant breast milk diet and limited introduction of other foods into the infant diet in the first 6 months of life. Although the overall frequency of GE was not increased in the PEPI trial compared with NVAZ trial, we found that after age 6 months, HIV-uninfected infants in the PEPI trial were at significantly increased risk of serious GE events including GE-associated hospitalization; and also had significantly higher overall and GE-associated mortality when compared with HIV-uninfected infants in the NVAZ trial. GE-associated hospitalizations were significantly increased among the PEPI compared with NVAZ infants after the cessation of breastfeeding between 6 and 9 months of age in the PEPI trial.
Although the HIV late postnatal transmission rate was lower in the PEPI16 than in the NVAZ6 (3.5% and 5.7%, respectively, at age 9 months), these findings are highly concerning given the advances in general care available to PEPI trial participants compared to those in NVAZ trial. For example, the HIV-exposed infants in the PEPI trial received cotrimoxazole prophylaxis beyond the first 6 months of life, whereas this was not provided for uninfected infants in the NVAZ trial. All HIV-uninfected infants in the PEPI trial received the prophylaxis up to 3 months after cessation of breastfeeding. Given a recent study from Uganda17 suggesting that cotrimoxazole is protective against diarrhea, hospitalizations, malaria and overall mortality among uninfected, HIV-exposed household family members, a lower frequency of overall GE events in infants in the PEPI trial than NVAZ trial might be expected. However, despite cotrimoxazole prophylaxis, serious GE-associated events were more frequent in the PEPI trial than in the NVAZ trial.
We hypothesize that the higher probability of GE-associated hospitalizations and mortality in the PEPI trial may in part be related to early breastfeeding cessation at 6 months of age. Implementation of this recommendation, which follows the Malawi Ministry of Health guidelines in place at the time of the study, may have advanced the cessation of breastfeeding by as much as 12 months compared with traditional practices. Women in the PEPI trial received counseling at each study visit regarding proper use of replacement feeding, as described in the Methods section. However, the absence of protection against enteric pathogens associated with GE which would have been afforded by continued breastfeeding combined with introduction of weaning foods in poor hygienic conditions in many of the households of the study participants may have contributed to the higher serious GE-associated morbidity and mortality in the PEPI trial noted in the immediate 3-month period postweaning.
The challenges of providing safe, uncontaminated foods that fully meet infant nutritional needs in low-resource settings should not be underestimated. Although the PEPI trial mothers have somewhat of a better socioeconomic profile than the NVAZ trial mothers do, the PEPI mothers still had poor access to safe water and challenges with hygiene in preparation of weaning foods (Table 1). This means that as mothers in the ongoing PEPI study complied with advice to stop breastfeeding at 6 months, they were faced with the daily challenge of finding safe water for preparation of breast milk substitutes and complementary foods. The importance of safe water and potential new or ongoing risks of contamination was recently highlighted by an outbreak of severe diarrhea and high infant mortality in the Francistown area of Botswana.18 Botswana promotes replacement feeding with formula from birth for HIV-exposed infants, and the population generally has access to a clean water supply. However, during a heavy rainy season, the ground water became contaminated leading to a 24-fold increase in infant deaths due to diarrhea, primarily among formula-fed infants.
There are certain caveats regarding the findings presented in this comparative analysis of PEPI and NVAZ GE-associated morbidity and mortality data. First, the comparison of early breastfeeding cessation in the PEPI trial was to a historic control, the NVAZ trial. Therefore, there may be temporal changes or other unknown factors that may have contributed to the differences in GE-associated hospitalizations and mortality in the 2 trials. However, the known risk factors we were aware of would have contributed to a lessening of risk of morbidity and mortality for the PEPI trial compared with NVAZ based on both the higher socioeconomic status (though with less access to running water) of the mothers and the use of cotrimoxazole prophylaxis by uninfected infants in the PEPI trial.
Early weaning could significantly reduce postnatal MTCT of HIV and theoretically improve HIV-free survival. However, a recent randomized trial in Zambia (Zambia Exclusive Breastfeeding Study) comparing exclusive breastfeeding with early weaning at age 4 months with exclusive breastfeeding with traditional weaning (median duration of breastfeeding of 16 months) found no overall HIV-free survival benefit with weaning the infant at 4 months.19 The Zambia Exclusive Breastfeeding Study also found that early weaning was associated with decreased survival of HIV-infected infants. For this reason, we have chosen to focus only on those infants documented to be HIV uninfected in our analysis.
The PEPI and NVAZ trials have very different distributions of infants by breastfeeding mode after age 6 months (with the vast majority of NVAZ trial infants continuing to breastfeed past 12 months, whereas most PEPI trial infants were weaned early between 6 and 9 months). The exact breastfeeding status at the time of death for infants who died of GE between study visits was often unknown. For these reasons, it was not possible to make direct comparisons between the trials by actual mode of infant feeding at the time of death.
Strengths of the analyses are that data for both studies are from randomized clinical trials conducted at the same clinical sites with the same staff. In addition, standardized data collection and toxicity evaluations based on DAIDS Toxicity Tables were used in both studies. The observed increase in serious GE events associated temporally with early breastfeeding cessation between 6 and 9 months postpartum in the PEPI trial is also consistent with reports from 2 other ongoing studies in Uganda in which early cessation of breastfeeding by age 6 months was also part of the national policy on infant feeding for HIV-infected women and included in the study research design and for which a historic control without early weaning was available for comparison.20,21
The findings we report from Malawi demonstrating a significant increase in serious GE-associated events (hospitalization, death) in HIV-exposed but uninfected infants which was temporally associated with early breastfeeding cessation are concerning. The findings add to accumulating data from other African countries13,14,19,21 that in some low-resource settings the presumed reduction in postnatal HIV transmission that might be expected by early weaning may be countered by increases in serious GE-associated morbidity and mortality that may occur with early cessation of breastfeeding.19,22 The maternal and infant antiretroviral prophylaxis intervention studies currently under evaluation for prevention of postnatal transmission still presume early weaning for maximal efficacy of these interventions. It is critical to develop new strategies, such as approaches that allow safer longer breastfeeding among HIV-infected mothers in low-resource settings such as Malawi, where background infant morbidity and mortality due to enteric pathogens is high and the ability to provide safe replacement feeding is limited.
We are grateful to the mothers and children who participated in the NVAZ and PEPI studies. We are also grateful to the nursing and technical staff in Malawi for their effort and for the excellent research collaborating in both Malawi and the United States.
2. The Breastfeeding and HIV International Transmission Study Group. Late postnatal transmission of HIV-1 in breast-fed children: an individual patient data meta-analysis. J Infect Dis
3. Bertolli J, St. Louis ME, Simonds RJ, et al. Estimating the timing of mother-to-child transmission of Human immunodeficiency virus in a breast-feeding population in Kinshasa, Zaire. J Infect Dis
4. Dunn DT, Newell ML, Ades AE, et al. Risk of Human immunodeficiency virus type 1 transmission through breastfeeding. Lancet
5. De Cock KM, Fowler MG. Prevention of mother-to-child HIV transmission in resource-poor countries: translating research into policy and practice. JAMA
6. Taha TE, Hoover DR, Kumwenda NI, et al. Late postnatal transmission of HIV-1 and associated factors. J Infect Dis
7. WHO Collaborative Study Team on the Role of Breastfeeding on the Prevention of Infant Mortality. Effect of breastfeeding on infant and child mortality due to infectious diseases in less developed countries: a pooled analysis. Lancet
8. Gaillard P, Fowler MG, Dabis F, et al. Use of antiretroviral drugs to prevent HIV-1 transmission through breastfeeding: from animal studies to randomized clinical trials. J Acquir Immune Defic Syndr
9. Taha TE, Kumwenda NI, Gibbons A, et al. Short post exposure prophylaxis in newborn babies to reduce mother-to-child transmission of HIV-1: NVAZ randomized clinical trial. Lancet
10. Taha TE, Kumwenda NI, Hoover DR, et al. Nevirapine and zidovudine at birth to reduce perinatal transmission of HIV in an African Setting. JAMA
11. Ministry Of Health and Population. Prevention of Mother-to-Child Transmission of HIV in Malawi: Handbook for Health Workers. Chapter 7
. Lilongwe, Malawi: Ministry of Health; 2003:25-32.
12. WHO. New data on the prevention of mother-to-child transmission of HIV and their policy implications. Conclusions and recommendations. WHO technical consultation on behalf of the UNFPA/UNICEF/WHO/UNAIDS Inter-agency Task Team on mother-to-child transmission of HIV. Geneva, October 11-13, 2000. Geneva: WHO 2001, WHO/RHR/01.28.
13. National Statistical Office (NSO) [Malawi], and ORC Macro. Malawi Demographic and Health Survey 2004
. Calverton, MD: NSO and ORC Macro; 2005.
14. Mermin J, Lule J, Ekwaru JP, et al. Effect of co-trimoxazole prophylaxis on morbidity, mortality, CD4-cell count, and viral load in HIV infection in rural Uganda. Lancet
15. National Statistical Office, ORC Macro. Malawi Demographic and Health Survey 2004
. Calverton, MD: NSO and ORC Macro; 2005.
16. Kumwenda NI, Hoover DR, Mofenson LM, et al. Extended antiretroviral prophylaxis to reduce breast-milk HIV-1 transmission. N Engl J Med
17. Creek T, Arvelo W, Kim A, et al. Role of infant feeding and HIV in a severe outbreak of diarrhea and malnutrition among young children, Botswana, 2006. Presented at: 14th Conference on Retroviruses and Opportunistic Infections; February 25-28, 2007; Los Angeles, CA. Abstract 770.
18. Kourtis A, Fitzgerald D, Hyde L, et al. Diarrhea in uninfected infants of HIV-infected mothers who stop breastfeeding at 6 months: the BAN study experience. Presented at: 14th Conference on Retroviruses and Opportunistic Infections; February 25-28, 2007; Los Angeles, CA. Abstract 772.
19. Kuhn L, Aldrovandi GM, Sinkala M, et al. Effects of early, abrupt weaning on HIV-free survival of children in Zambia. N Engl J Med
20. Thomas T, Masaba R, van Eijk A, et al. Rates of diarrhea associated with early weaning among infants in Kisumu, Kenya. Presented at: 14th Conference on Retroviruses and Opportunistic Infections; February 25-28, 2007; Los Angeles, CA. Abstract 774.
21. Onyango C, Mmiro F, Bagenda D, et al. Early breastfeeding cessation among HIV-exposed negative infants and risk of serious gastroenteritis: findings from a perinatal prevention trial in Kampala, Uganda. Presented at: 14th Conference on Retroviruses and Opportunistic Infections; February 25-28, 2007; Los Angeles, CA. Abstract 775.
22. Atashili J, Kalilani L, Seksaria V, et al. Shorter duration of breastfeeding in infants of HIV-infected women in Africa may substantially reduce infant HIV infection but not mortality: a simulation study. Presented at: 14th Conference on Retroviruses and Opportunistic Infections; February 25-28, 2007; Los Angeles, CA. Abstract 771.