Postpartum

A randomized controlled trial from 2005-2008 in Burkina Faso, Kenya and South Africa assessed both mothers’ health and mother-to-child transmission among HIV-positive women whose CD4 count was between 200 and 500. The study found that triple-antiretroviral treatment initiated during pregnancy and continued until six months postpartum reduced the risk of transmission to infants and improved HIV-free survival of infants compared to standard short-course antiretroviral therapy. At 12 months, 6.7% of the 402 infants whose mothers received triple-course antiretroviral treatment had died compared to 10.2% of the 403 infants whose mothers received short-course antiretroviral treatment. This effect was especially strong in women with CD4 counts between 200 and 350. At 12 months, the rate of transmission from mother to infant for triple-antiretroviral therapy was 5.5% compared to 9.5% for those who received short-course antiretroviral treatment. The infants whose mothers received triple-course antiretroviral therapy experienced a 42% risk reduction in HIV infections and a 37% risk reduction of death at 12 months, for a combined 36% risk reduction of either HIV infection or death. The study also found that triple-antiretroviral therapy had low toxicity for mothers and infants. All infants received single-dose nevirapine plus zidovudine in the first 72 hours and all mothers received counseling on replacement feeding or exclusive breastfeeding with weaning by six months. Formula was provided free of cost.

A study in Botswana (no date given) found a positive association between maternal viral load (in both plasma and breast milk) and mother-to-child transmission after one month in breastfed infants. 1,200 HIV-positive women at 4 sites were enrolled in the study and randomized to either exclusively breastfeed for 6 months in combination with infant zidovudine treatment or to exclusively formula feed. Mothers received antenatal zidovudine starting at 34 weeks of pregnancy along with intrapartum zidovudine and either single-dose nevirapine or a placebo at delivery. During the study HAART became available and women with CD4 counts below 200 cells/mm3 or AIDS defining illnesses were eligible for treatment either antenatally or postnatally. Infants received single-dose of nevirapine or a placebo at birth along with one month of zidovudine prophylaxis for formula fed infants and six months for breastfed infants. After 17 months the study protocol was changed so that all infants received single-dose nevirapine at birth. Of 1,116 infants alive and HIV-negative at birth, 1.1% of formula fed and 1.3% of breastfed infants tested HIV-positive after one month. Of 547 breastfed infants HIV-negative at one month, 4.4% tested HIV-positive before 2 years of age. Only 4 formula-fed infants tested HIV-positive after one month but before 2 years of age. For breastfed infants, the only predictors of mother-to-child transmission after one month of age were high maternal viral load and low maternal CD4 count. Infant prophylaxis with zidovudine was not a significant predictor of transmission. No transmission was observed in breastfeeding mothers who had started treatment with HAART before delivery. Similarly, no transmission was observed in breastfeeding mothers who had viral loads of less than 3,500 copies/mL. Maternal treatment with single-dose nevirapine at delivery did not predict mother-to-child transmission.

The Post-Exposure Prophylaxis of Infants (PEPI) trial in Malawi found that extended infant prophylaxis with nevirapine or with nevirapine and zidovudine for the first 14 weeks of life significantly reduced breast-feeding acquired HIV-1 infection in 9-month-old infants. Between 2004 and 2007, 3016 breastfeeding infants were randomly assigned to one of three different drug regimens. The control group received single-dose nevirapine plus one week of zidovudine, the second group received the control regimen plus daily extended prophylaxis with nevirapine (extended nevirapine group) and the third group received the control regime plus nevirapine and zidovudine (extended dual prophylaxis group). At nine months (the primary end point in the study), the estimated rate of HIV-1 infection in the control group was 10.6%. The extended nevirapine group had an infection rate of 5.2% and the extended dual prophylaxis group had a rate of 6.4%. There were no significant differences between the two extended prophylaxis groups although the extended dual prophylaxis group had a significant increase in the number of adverse events which were thought related to a study drug. This study demonstrated a protective efficacy of more than 60% for the two extended prophylaxis groups at 14 weeks. Cumulative risk of postnatal infection between birth and 14 weeks was 8.4% in the control group and 2.8% in the extended prophylaxis groups. This net difference of approximately 5% continued at 24 months.

The Six Week Extended-Dose Nevirapine (SWEN) study combined study data from sites in Ethiopia, India and Uganda to assess whether daily nevirapine given to breastfed infants through six weeks of age would decrease HIV transmission from breastfeeding. HIV-positive women who were breastfeeding their infants were randomized to receive either single-dose nevirapine (during labor for the mother and after birth for the baby), or six week extended dose nevirapine (during labor for the mother and after birth for the baby) plus daily nevirapine doses for the baby from day 8 to 42. The primary goal of the study was to assess HIV-infection rates at six months of age for infants who were HIV PCR negative at birth. The study concluded that a six week regimen of daily nevirapine might be associated with a reduction in the risk of HIV transmission at six weeks of age but the lack of a significant reduction of HIV transmission at the study end point of six months of age suggests that a longer course of daily infant nevirapine to prevent HIV transmission via breast milk might be more effective. 

Note: The three co-principal investigators of this study from India published a critique of this study write up in Lancet in the same publication issue. These investigators disagree with the statistical analyses used in this study, express concern about the 40% of infants who experienced grade III or IV side effects during treatment and conclude that the recommendation to continue nevirapine beyond six weeks is “inappropriate.” The investigators suggest that a more prudent strategy is to “follow WHO/UNICEF guidelines for developed countries and to make formula feeding safe, sustainable, acceptable, and affordable for mothers in developing countries” (SWEN, 2008: 287).

A study in Ethiopia, India and Uganda found that at 12 months of age infants who were on extended-dose nevirapine had reduced mortality and HIV transmission compared to infants who received single-dose nevirapine. Two thousand and seventeen HIV-positive pregnant women were recruited for the study from 2001 to 2004 in the three countries. A total of 1,890 infants were randomized in to single-dose nevirapine (given at birth once) and extended-dose nevirapine (given at birth and for 6 weeks daily) group. At the end of the 12 months follow-up period in 2007, data was available for 902 infants in the single-dose nevirapine group and 803 infants in the extended-dose group. The number of infants who received HAART during infancy was similar between the groups. HIV transmission was 8.9% in the extended-dose group compared to 10.4% in the single-dose nevirapine group. At 12 months, the overall mortality was 5% in the single-dose group compared to 1.9% in the extended-dose group. The impact of extended-dose nevirapine was highest in mothers whose CD4 counts were above 350 at birth (HIV transmission 8.4%; death 7.8%) compared to infants born to mothers whose CD4 counts were under 200 (HIV transmission 15.4%; death 27.6%). The extended-dose group had significantly fewer deaths (62% reduction) compared to the single-dose group among the uninfected infants.

A follow-up study of the PEPI trial conducted from 2004 to 2007 in Malawi assessed the impact of maternal HAART on vertical transmission from 14 weeks until 24 months postpartum among 2,188 HIV-negative and breastfed infants, who had received 14 weeks of extended antiretroviral prophylaxis. Among eligible mothers (CD4 cell count below 250), HAART provided an 82% reduced risk of vertical transmission compared to those who did not receive HAART, after adjusting for infant antiretroviral prophylaxis. In other words, mothers who were eligible for HAART and did not receive treatment were at 5.89 times increased risk of transmitting HIV to their infants compared to mothers who were eligible and received treatment. Per maternal treatment category, the rate of HIV transmission to infants for the eligible but untreated mothers was 10.56 per 100 person-years, for the eligible and treated mothers was 1.79 per 100 person-years and for the ineligible mothers was 3.66 per 100 person-years. In the PEPI trial, infants were randomized into 3 arms with varying lengths of infant antiretroviral prophylaxis (single-dose nevirapine and one week of zidovudine, nevirapine for 14 weeks or nevirapine plus zidovudine for 14 weeks) and followed for 24 months. The trial initially reported 5.6% (130 in 2,318) mother-to-child-transmission from birth to week 14. Mothers were counseled to exclusively breastfeed for 6 months and then wean. About 80% of infants were weaned between 6 and 9 months. Infants were provided with antibiotics from age 6 weeks until 3 months after weaning. At the start of the study, maternal HAART was not available in Malawi. Once the government program introduced HAART in 2006, most mothers enrolled in the study began treatment at 14 weeks postpartum, after infants in the intervention arm had finished their treatment. Maternal HAART status was categorized as eligible but untreated, eligible and treated or ineligible, based on a CD4 cell count threshold of 250 and HAART treatment status determined by records.

A 2001-2003 study in Tanzania assessing 398 infants of HIV-positive women intending to breastfeed who were treated with zidovudine and lamivudine at antenatal clinics, found a 3.8% transmission rate of HIV from mother-to-child at week six and a 4.9% transmission rate after six months. The cumulative rate of HIV infection or death for infants was 8.5% at six months. Women were treated with zidovudine or lamivudine from 36 weeks gestation to one week post-delivery. Infants were treated with zidovudine and lamivudine for the first week of life and then lamivudine throughout six months of breastfeeding. Follow-up appointments included infant feeding counseling and occurred at weeks 1, 3 and 6 and months 3, 6, 9, 12, 15, 18, 21 and 24. Women were counseled to breastfeed exclusively and wean by six months. The infants were breastfed for a median of 18 weeks. Mothers reported 95% breastfeeding at six weeks, 86% after 12 weeks and 18% after 26 weeks. A total of 19 children became HIV-positive, 15 were considered early transmissions and 4 were considered late transmissions. CD4 count and viral load were significantly associated with mother-to-child transmission. No infants suffered serious adverse outcomes due to antiretroviral treatment. The comparison group for this study was a historical study of the same cohort where mothers received the same antiretroviral regimen but infants were not treated throughout breastfeeding. In this earlier study, mothers reported 85% breastfeeding at six weeks, 77% at 12 weeks and 64% at 26 weeks. This study revealed a 5.4% transmission rate at six weeks and 11.9% transmission rate at six months. The cumulative risk for HIV acquisition or death was 8.7% at six weeks and 15.5% at six months, about 50% higher than the current study.

“A recent analysis undertaken for WHO for southern African countries found that the cost per 10,000 HIV-positive mothers would be US$522,542 with the option of breastfeeding plus maternal HAART for women with a CD4 count under 350. In comparison, it would cost US $2,063,100 per 10,000 HIV-positive mothers provided with maternal HAART and six months of formula milk for women with a CD4 count under 350. The study concluded ‘any feeding strategy that includes free provision of infant formula to HIV-infected mothers, even for a limited six months, is between two and six times more costly than a strategy that provides ARVs as prophylaxis to reduce postnatal transmission'" (WHO, 2010 cited in Doherty et al., 2010).

The BAN (Breastfeeding, Antiretrovirals and Nutrition) study in Malawi randomized 2,369 mother-infant pairs. HIV-positive pregnant women and their infants were randomized to three groups: 1) HAART for the mother (849); 2) twenty-eight weeks of nevirapine for the infant (852); and 3) neither (668). Enrollment to the control group, which received neither of the interventions, was halted by the Data and Safety Monitoring Board by NIH’s Division of AIDS in 2008 and ethically could not be conducted today. All mothers received single dose nevirapine, which WHO no longer recommends. All women were counseled to wean rapidly between 24 and 28 weeks (also no longer recommended by WHO). By twenty-eight weeks, complete cessation of breastfeeding was reported by 67% to 68% of mothers in each study group. Among infants who were HIV-negative at two weeks, the estimated risk of HIV-1 infection by 28 weeks for infants was 10.9% in the control group, 8.2% in the maternal regimen group and 6% in the infant regimen group. All mothers and infants received perinatal prophylaxis with single-dose nevirapine and 1 week of zidovudine plus lamivudine.

A nonrandomized interventional cohort study from four public health centers in Rwanda from 2005 to 2007 found that maternal HAART while breastfeeding resulted in a minimal risk of postnatal transmission. Women could choose to feed their infants either formula (305 infants) or breastfeeding (227 infants) while on HAART. All women regardless of CD4 count, were given HAART. Only one infant who was breastfed became HIV-positive while the mother was on HAART; none of the infants fed by formula became HIV-positive, with this difference not being statistically significant. Adherence to HAART by the HIV-positive mother whose infant acquired HIV was uncertain, due to vomiting from gastritis. Among breastfeeding mothers on HAART, the cumulative probability of vertical transmission by nine months was 1.8%. Formula was provided at no cost. After birth, prophylactic HAART to prevent vertical transmission rather than treatment for the mother, was stopped among those women who chose to formula feed their infants. For women who choose to breastfeed, prophylactic HAART was given until the infant was 7 months, i.e. one month after weaning. Following WHO recommendations, all newborn infants exposed to HIV received nevirapine at birth and ZDV twice daily for 7 days. Follow up visits were scheduled at 15 days, 6 weeks and 3, 6, 7, and 9 months postpartum where adherence to HAART was assessed through pill counts and mothers received feeding counseling.

A pooled analysis of data from Kenya, Burkina Faso, Côte d’Ivoire, Rwanda, Mozambique and Tanzania found that “HAART in breastfeeding women results in transmission rates generally under 5%" (Becquet et al., 2009a: 1942).

A study in Mozambique from 2005-2007 followed 313 HIV-positive mothers on HAART, who were counseled to breastfeed exclusively for six months and found that HAART reduced the risk of mother-to-child transmission by 93%. There were a total of 8 cases of HIV transmission, 4 of which were considered late postnatal transmission. Women with repeat pregnancies, who had previously received antenatal care and HAART through six months of breastfeeding, did not transmit HIV to their infants. HIV testing of infants was performed at 1, 6 and 12 months. Antiretroviral treatment began at 15 weeks of gestation and continued until six months after delivery. HAART was continued beyond six months if the mothers had CD4 cell counts that remained below 350. In combination with HAART, nutritional supplements to mother and infant, patient counseling to increase adherence to breastfeeding and a strong network of support within the community led to the marked reduction of maternal and infant deaths.

A study enrolling HIV-positive pregnant women receiving a single dose of nevirapine for preventing perinatal transmission of HIV during labor from 2003-2004 in Uganda found that nevirapine was detectable in breast milk, maternal plasma, and infant plasma for 2-3 weeks after a single dose of maternal nevirapine. Overall, 62 women were included in the study. Sixty-one women received a single dose of nevirapine at least 1.5 hours before delivery, and 53 women chose to breastfeed. All infants received a single dose of nevirapine syrup within 72 hours of birth. Samples of breast milk and plasma from both mothers and infants were taken 1, 2, and 6 weeks after maternal nevirapine treatment. Infant plasma levels of nevirapine at delivery were correlated with the timing of maternal nevirapine intake. Infant nevirapine levels were the highest approximately 4 hours after maternal nevirapine intake, after which infant treatment with nevirapine only slightly increased infant nevirapine plasma concentrations. Furthermore, nevirapine transferred from maternal plasma to breast milk rapidly, and nevirapine in breast milk was detectable before infants initiated breastfeeding. The long-term duration of nevirapine in breast milk was determined to be protective against postnatal transmission due to the effective suppression of HIV in breast milk for up to 3 weeks after maternal single dose nevirapine intake. However, the long-term duration of nevirapine also increases the risk for nevirapine resistance mutation development, and the acquisition of a resistant virus for infants. Because the risk of nevirapine resistance decreases over time, infants are most at risk for acquiring a resistant virus during the initial breastfeeding period. Extended antiretroviral treatment with zidovudine/lamivudine should therefore be considered to reduce the risk of nevirapine resistance. [See introduction of Antenatal Care - Treatment for discussion of nevirapine resistance]

Lower maternal CD4 count was associated with a significantly higher risk of transmission through breastfeeding (Mofeson et al., 1999 cited in Abrams et al., 2007), therefore HAART, by increasing CD4 counts, can reduce transmission of HIV during breastfeeding.

A prospective study of 354 HIV-positive pregnant women who attended two urban clinics in Rwanda found that providing long-acting reversible contraceptives (Implants and IUDs) on site increased the number of women starting hormonal implants in the postpartum period. In Rwanda in 2009, among women currently not using contraceptives but intending to use a modern method in the future one in five would prefer a long-acting reversible contraceptive or sterilization (Ministry of Health, Rwanda, 2009 cited in Dhont et al., 2009). However, access to long acting reversible contraceptives has been limited. Between 2005 and 2007, HIV-positive pregnant women entering the PMTCT program from 28 weeks of gestation were invited to a non-randomized prospective cohort study where 179 women at one site were referred for public health family planning services and at the other site, 175 women were offered hormonal implants free of charge (Norplant or Jadelle) and IUDs on site and were referred to the adjacent public family planning services for short-acting contraceptives. At the public family planning services, long acting reversible contraceptives were only available occasionally due to stockouts, lack of qualified staff, and women were charged a small fee. The overall postpartum uptake of modern contraception, measured at the nine month visit, was 84% at both sites, significantly higher than that in a routine PMTCT setting where the rate was 45%. Only three out of 242 women without desire for future children were not using contraception. Uptake of implants among the women at the site where implants were offered on site was 38%, significantly higher than the 6% at the site where women were referred for public family planning services. When adjusting for maternal age and having at least one living child at enrolment, women with onsite access to implants were 11.8 times more likely to use implants than women referred for public family planning services.

A cross-sectional analysis from 2007 to 2009 derived from baseline visits of 435 HIV-positive women in St. Petersburg, Russia, of whom 120 were postpartum, who were counseled by clinicians on a range of contraceptive options for which they were eligible based on medical criteria found that 35% used a highly effective method of oral contraceptives together with condoms and 51.7% used DMPA along with condoms as their preferred choice. Family planning services were integrated into HIV clinical care. DPMA could be started immediately postpartum, while oral contraceptives could not be started until 21 days after delivery.

A quasi-experimental pre-post test study conducted from 2006 through 2007 of maternal health care interventions in Swaziland that provided care for all pregnant women, including HIV-positive women at several intervals (within the first six hours after delivery; an exam once per day postpartum while the woman was in the health facility; providing assessment, care and counseling, along with a specific appointment for the first postnatal visit upon being discharged from the facility and providing a postnatal visit at one week postpartum and a second visit at four to six weeks postpartum) increased contraceptive use and counseling on condom use. Over 60% of maternal deaths occur within 48 hours after childbirth (Lewis, 2004 cited in Mazia et al., 2009), yet in Swaziland, mothers are usually discharged within 12 hours of delivery. The conventional recommendation for the first postnatal visit is at four to six weeks, by which time most of the postpartum deaths have already taken place. The study collected data on 114 HIV-positive women at the start of the study and from 136 HIV-positive women to evaluate the impact a year later. The intervention increased early postnatal visits by twenty-fold. Providers increased counseling of HIV-positive women on the need to regularly monitor CD4 counts for the mother from 41% to 74%. Following the intervention, 93% of mothers were assured of privacy. While at baseline, the provider asked the woman her preferred family planning only 32% of the time, by the end of the intervention, 82% did so. While at baseline, only 28% of clients received their preferred family planning method, at the end of the intervention, 70% did so. While at baseline, providers only counseled on condom use 16% of the time, by the end of the intervention, 25% did so. The percent of women on HAART increased from 4% to 15% and the mother tested for her CD4 count since giving birth increased from 4% to 26%. There was also a statistically significant increase in the proportion of postpartum women (88 to 98%) and their partners (from 28% to 56%) getting tested for HIV. Since the postnatal visit within one week of delivery did not exist anywhere in the country at the pre-intervention phase, conclusions following the introduction of the new timing of postnatal care could be assessed. Actual condom use was not measured.

A study in Côte d’Ivoire with 546 HIV-positive women and 393 HIV-negative women who were tested for HIV prenatally and followed up for two years following delivery and were provided contraception as desired at each postpartum visit, resulting in high rates of contraception use after delivery and low pregnancy incidence. HIV-positive women had fewer unwanted pregnancies than HIV-negative women. At each postpartum visit, women received family planning counseling and free contraception. Between 6 and 24 months, proportion of women using modern contraception varied from 52 to 65% among HIV-positive women. Among HIV-positive women, pregnancy incidence for 100 women years at risk was 5.70 and unwanted pregnancy incidence was 1.07.

A pre-post test design with 356 postpartum women and 53 health care workers that instituted a one week post-delivery postpartum visit along with provider training in Swaziland from 2006 to 2007 found that the proportion of HIV-positive postpartum women not wanting another child increased from 77% to 83%. Provider training increased the woman being asked about her preferred contraceptive method, from 32% to 82% and receiving her preferred method, from 28% to 70%. Male partners who tested for HIV increased from 28% to 56%.

A study of 319 HIV-positive pregnant women who were followed postpartum for one year in a perinatal HIV transmission study in Kenya and were referred to local clinics for contraceptive counseling and management resulted in high rates of contraceptive use and dual method use, with 72% initiating hormonal contraceptive use and 61% of 231 hormonal contraceptive users reporting condom use in additional to hormonal contraceptives. Prior to this project, which had linked antenatal care with family planning, only 50% of the currently using 231 hormonal contraceptive users had a history of previous hormonal contraceptive use. Prior to this project, only 6 or 3% had used condoms. Of those using contraception, 44% used DMPA, 31% used oral contraception and 25% switching methods at follow up. Women were counseled antenatally to initiate contraception postpartum and dual contraception was encouraged. No particular method of contraception was given priority. Hormonal methods were the most popular contraceptive method, possibly because they are female controlled and available. Women who opted of formula feed their infants were counseled to initiate contraception four weeks after delivery, whereas those who opted to breastfeed were counseled to initiate contraception six weeks after delivery. Breastfeeding women who wanted oral contraception received progesterone only pills and non-breastfeeding women received combined oral contraceptive pills. DPMA was available for both breastfeeding and non-breastfeeding women. Median time to initiation of sexual activity was two months following delivery, ranging between one and 11 months, with 77% of women resuming sexual activity within three months of delivery. Partner notification and condom use were similar between those using and not using other forms of contraception besides condoms (Balkus et al., 2007). Other studies that did not provide contraceptive counseling in antenatal care found much lower rates of contraceptive use post partum (Nebie et al., 2001; Desgrees-du-Lou et al., 2002 cited in Balkus et al., 2007).

A study in Zimbabwe from 1997 to 2000 of 2,060 infants born to HIV-positive mothers found that solid foods or animal milks given to infants prior to three months of age was associated with a fourfold greater risk of postnatal transmission of HIV at six months compared with exclusive breastfeeding. The protective effects of early exclusive breastfeeding were still significant at 18 months with a 61% reduction in postnatal transmission compared with mixed breastfeeding. Thus, the more strictly HIV-positive mothers are able to breastfeed exclusively, the lower the risks of HIV or death for their infants. More than two-thirds of all postnatal transmission of HIV occurred after six months. This is consistent with other studies from West Africa, South Africa and Tanzania and supports early cessation of breastfeeding among HIV-positive women. Lastly, women with CD4 counts less than 200 cells/ul were five times more likely to transmit HIV during breastfeeding compared with women with CD4 cell counts over 500 cells/ul, confirming the findings of other studies that postnatal transmission of HIV is highly correlated with immune suppression in the mother.

A 2001-2004 study in Zambia followed 958 HIV-positive mothers, who intended to breastfeed their infants, for 24 months and found that weaning by five months was associated with a 2-fold increase in infant mortality and weaning by five to 18 months was associated with a 4-fold increase when compared to weaning after 18 months. The mothers were recruited from antenatal clinics, offered voluntary counseling and testing and single-dose nevirapine at birth. All mothers were counseled to breastfeed exclusively. Mothers were randomized into two groups: the first to abruptly wean at four months, and the second to continue breastfeeding and wean whenever they chose. For infants in the early weaning group, the study offered three-months supply of formula and fortified cereal. For 24 months, mothers and infants were followed every two weeks by either home visits or site visits. Among 749 uninfected infants who were breastfed for 18 months, mortality was 9.7% compared to 17.4% for infants who were weaned by 5 months. Infants weaned by 4-5 months experienced a 2-fold increase in mortality, by 6-11 months experienced a 3.54-fold increase in mortality, by 12-18 months experienced a 4.22-fold increase in mortality compared to infants who were weaned after 18 months. The leading causes of infant death were diarrhea (among infants aged 6-24 months) and pneumonia (all ages).

A retrospective study (year unspecified) analyzed data from two randomized trials (HIVNET 012 1997-2001 and HIVIGLOB/NVP 2004-2007) conducted at the same hospital in Uganda and found that early cessation of breastfeeding was associated with increased risk of serious gastroenteritis among HIV-exposed uninfected infants when compared to later breastfeeding cessation. Infants in the HIVIGLOB/NVP trial were weaned at a median age of four months, with complete cessation of breastfeeding by six months. Infants in the HIVNET trial were weaned at a median age of 9.3 months, with some infants continuing to breastfeed after 12 months. Study analyses included a total of 623 infants from the HIVNET trial and 698 infants from the HIVIGLOB/NVP trial. The rates of serious gastroenteritis events were higher among infants in the HIVIGLOB/NVP trial compared to the HIVNET trial reaching statistical significance at 3-4 months (16.2 events per 1000 child-months versus 0 events per 1000 child-months). The overall rates of serious gastroenteritis events were highest in the HIVIGLOB/NVP trial at 8 events per 1000 child-months compared to 3.1 events per child-months in the HIVNET trial, which was also statistically significant. Both trials weaned infants at much earlier ages than the national median age of 19.9 months according to the Uganda Demographic Health Survey in 2000. The HIVNET trial counseled mothers to exclusively breastfeed for at least six months and then to wean, though no specific time was indicated. The HIVIGLOB/NVP trial counseled to exclusively breastfeed for 3-6 months and then abruptly wean before six months. All infants in this trial received co-trimoxazole from six weeks until HIV-negative following cessation of breastfeeding. HIV-positive children continued to receive co-trimoxazole indefinitely. Gastroenteritis was defined as an episode of diarrhea (3 or more loose or watery stools within 24 hours), regardless of vomiting. Serious gastroenteritis was defined as diarrheal events that resulted in hospitalization or death of the infant.

A study from Zambia (2001 to 2004) enrolling HIV-positive pregnant women from PMTCT programs, found that infants born to HIV-positive mothers who were exclusively breastfed up until at least 4 months were at least 50 percent less likely to acquire HIV through breastfeeding than infants fed any non-breast milk substances in addition to breast milk. Furthermore, the study found no difference in the rates of HIV transmission between infants weaned at 4 months and those who continued breastfeeding past 6 months. Overall, 734 infants who tested HIV-negative at 6 weeks of age and were still breastfeeding at 6 months of age were included in the study. Mothers were randomized into an intervention group in which women were counseled to exclusively breastfeed for 4 months and then wean abruptly, and a control group in which women were counseled to breastfeed for at least 6 months and then introduce complimentary foods while maintaining breastfeeding. At 4 months, 83.5 percent of mothers reported exclusively breastfeeding. The risk of acquiring HIV before 4 months of age was over 3 times higher for infants who were non-exclusively breastfed compared to those who received only breast milk. A maternal CD4 count of below 350 was a strong predictor of HIV transmission before 4 months of age, but a significant reduction in HIV transmission related to exclusive breastfeeding remained after controlling for CD4 count. For exclusively breastfed infants, the risk of acquiring HIV was greatest in the first 4 months and then declined thereafter. The rate of HIV transmission for non-exclusively breastfed infants was 2.4 percent per month compared to less than 1 percent per month for exclusively breastfed infants.

A sub-study followed 118 infants born to 102 HIV-positive women receiving HAART for their own health from 2003-2007 in Uganda. After adjusting for maternal CD4 count, maternal marital status or maternal death, the study found that infants who were breastfed for less than six months had a 6-fold increased risk of death compared to infants who were breastfed longer than six months. Infants were followed for a median of 18 months and mothers received HAART prior to delivery for a median of 20 months. By six months of age, 25% (29 in 118) of infants were exclusively breastfed, 20% (23 in 118) received mixed feeding and 48% (57 in 118) had been weaned. No infants in this study were confirmed as HIV-positive at the end of follow-up or death, though only 86% had a final HIV test. About 19% (23 in 118) of infants died during follow-up and these infants were breastfed (exclusively or mixed) for a median of three months. Of infant deaths, about 65% were preceded by severe diarrhea and vomiting, possibly caused by gastroenteritis as a result from weaning. The infants included in this sub-study were part of a randomized clinical trial aimed to evaluate monitoring strategies for 1,100 HIV-positive adults (aged 18-49 with CD4 cell counts below 250 or WHO stage III/IV) receiving HAART in two rural districts in Uganda. All women who gave birth to a live infant during this trial were recruited into the sub-study. Mothers were given single-dose nevirapine at labor and, after 2005 infants were given daily zidovudine treatment for one week in addition to maternal treatment. All mothers were counseled to exclusively breastfeed for six months, then wean and begin replacement feeding. Health workers determined breastfeeding status at weekly home visits. HIV and CD4 tests were collected at 3-month intervals.

A sub-analysis (year unspecified) reported on gastroenteritis-related morbidity and mortality among infants from two randomized trials (one trial from 2003 to 2003 and another trial from 2004 to 2009) on prevention of mother-to-child transmission of HIV in Malawi. The analysis found that after age six months, hospitalization and mortality related to gastroenteritis were significantly higher among infants who were weaned early (PEPI) compared to infants who were breastfed long-term (NVAZ). Frequency of hospitalization related to gastroenteritis was consistently higher among infants in infants weaned before six months compared to infants breastfed for longer than six months: 2.9% versus 0.1% at 7-9 months and 1.6% versus 0.2% at 10-12 months. Gastroenteritis-related deaths were also consistently higher among infants in weaned before six months (PEPI) compared to infants breastfed for longer than six months (NVAZ): 19 versus 7 per 1,000 infants at nine months and 24 versus 12 per 1,000 infants at twelve months. The study included 2,035 mother-infant pairs in the PEPI trial from 2004-2007, when women were encouraged to breastfeed exclusively for six months and then wean, and 1,810 mother-infant pairs in the NVAZ trial from 2000-2003, when breastfeeding lasted as long as desired. Among women who weaned their infants before six months, only 23.3% had running water at home; among those who breastfed longer than six months, 26.2% had running water at home.  “…Mothers …had poor access to safe water” (Kafulafula et al., 2010: 12). However, only 3.5% of infants who were weaned before six months were HIV-positive compared to 5.7% of HIV-positive infants who were breastfed longer than six months. In the PEPI trial, infants were randomized into 3 arms with varying lengths of infant antiretroviral prophylaxis (single-dose nevirapine and one week of zidovudine, nevirapine for 14 weeks or nevirapine plus zidovudine for 14 weeks) and followed for 24 months. Mothers were counseled to exclusively breastfeed for 6 months and then wean. Infants were provided with antibiotics from age 6 weeks until 3 months after weaning. About 80% of infants were weaned between 6 and 9 months. In the NVAZ trial, infants were randomized to receive single-dose nevirapine or single-dose nevirapine plus one week of twice daily zidovudine. Mothers received single-dose nevirapine if they presented early during labor. Infants were followed for 24 months. About 89% were still breastfeeding at 9 months and 60% by 24 months.

A sub-study (year unspecified) analyzed 1,761 infants born to HIV-positive women, who participated in the PEPI trial in Malawi, and found that morbidity and mortality were higher among the infants who were weaned prior to 6 months of age regardless of treatment arm compared to infants who continued breastfeeding. At 15 months after adjusting for infant antiretroviral prophylaxis, pneumonia prophylaxis and mother’s HIV disease stage, the cumulative mortality for infants who were weaned by 6 months was 6.4% compared to 3.5% among infants who continued breastfeeding. In terms of morbidity, after adjusting for the same factors, infants who were not breastfed had approximately 1.7 times higher rate of illness and/or hospital admission than infants who were breastfed. The PEPI trial was conducted from 2004 to 2009. It was a randomized controlled trial aiming to measure the efficacy of extended infant antiretroviral prophylaxis to prevent mother-to-child HIV transmission. The trial recommended weaning by 6 months. All infants included in this analysis were breastfed for a minimum duration of 14 weeks (through treatment). The treatment arms were composed of the following regimens: 1) single-dose nevirapine given at birth and one week of twice-daily zidovudine (control); 2) control regimen plus daily nevirapine until 14 weeks of age; and 3) control regimen plus daily nevirapine and zidovudine until 14 weeks of age. Infants were followed-up regularly from birth until 24 months of age. All infants were provided with pneumonia prophylaxis from age 6 weeks through 3 months after weaning. Analysis was done by 3-month age intervals: 6-9 months, 9-12 months and 12-15 months.

A retrospective cohort study analyzed data on 1,261 infants of HIV-positive mothers and 1,061 infants born to HIV-negative mothers who attended 9 antenatal clinics in South Africa from 2001 to 2004. The study found that infants born to HIV-positive women grew as well as the children born to HIV-negative women, regardless of feeding method though the median duration of exclusive breastfeeding was six months in both groups. In adjusted analysis, the factors most strongly associated with poor growth were maternal mid-upper arm circumference (proxy for BMI or maternal nutrition), CD4 count, infant birth weight and HIV status. All HIV-positive mothers and infants received single-dose nevirapine during labor. HAART was not widely available during the study period. HIV-positive mothers were included in the study if they had known HIV status and CD4 count below 250. Infants of HIV-negative and HIV-positive mothers born between 19-44 weeks gestation with at least one growth measurement were included in the analysis. The study provided counseling support on infant feeding prior to birth, which continued through weekly home visits conducted by lay health workers until infants reached nine months of age to support breastfeeding. At these visits data on infant feeding and morbidity were collected. High rates of exclusive breastfeeding were achieved, with a median duration of 175 days, or a little less than five months. Infants given breast milk had an average weight of .048 higher z-score (standard deviations above or below the mean) compared with children with no breast milk, after allowing for birth weight, maternal mid-upper arm circumference, HIV infection status and maternal CD4 count. Mother-infant pairs also visited the clinic monthly from week 6 to month 9 and then quarterly until 2 years.

A study in Nigeria, which screened pregnant women for HIV-1 between 2004 and 2006, found that risk factors for mother-to-child transmission of HIV-1 differed by infant age. Infants at highest risk of acquiring HIV were those who had mothers with CD4 counts less than 200 and who received mixed feeding. The study analyzed 391 mothers and 371 infants, using follow-up visits 1 week after delivery and 1, 3, 6, and 12 months after delivery. A single-dose of nevirapine was given to each mother during delivery and to her infant within 48 hours of delivery. Women who chose replacement feeding were provided a 6-month supply of formula free of charge, as well as training and counseling on formula preparation, sterilization, and storage processes. Mothers who chose exclusive breastfeeding were provided counseling on the importance of weaning before 4-6 months. Exclusive breastfeeding was defined as only breast milk up until 6 months, with no other liquids or solids; replacement feeding as the use of formula only with no breast milk; and mixed feeding as a combination of breast milk and nonhuman milk or other solids before 6 months of age. For infants who were exclusively breastfed, 8.1% tested HIV-positive by 6-months of age compared to 9.5% of infants exclusively formula fed, and 29.2% of infants who received mixed feeding. After delivery, 71.7% of mothers chose replacement feeding while 28.3% chose to exclusively breastfeed. At 6-month follow-up, 71.1% of mothers who initially chose to breastfeed reported maintaining exclusive breastfeeding, 80.2% of mothers who initially chose formula feeding reported exclusive replacement feeding, and 82 mothers reported using mixed feeding. During the study period, 50 infants became infected with HIV-1, 34% in utero, 30% intrapartum or early postnatally, and 36% postnatally, with an overall transmission rate of 13.5%. For infants infected in utero, risk factors included maternal CD4 count of less than 200 and high maternal viral load. For infants infected during the intrapartum or early postnatal period, risk factors included high maternal viral load, gestational age of less than 37 weeks, and prolonged membrane rupture during delivery. Infants infected during the intrapartum or early postnatal period were at higher risk if they received mixed feeding compared to infants who were exclusively formula or breast-fed (12% compared to 2.2%). For infants infected during the postnatal period, mixed feeding and low birth weight increased the risk of HIV transmission. The risk of transmission for infants who were exclusively breastfed increased from 1.4% during the intrapartum/early postnatal period to 4.2% postnatally. The rate of transmission during all three infant-age periods for infants who were exclusively formula fed was similar. For mothers who initially chose to replacement feed but then switch to mixed feeding, stigma, pressure from family members, and no partner support were reported as reasons for not maintaining exclusive formula feeding.

A 2001-2005 South African intervention cohort study of 1,372 women and infants which examined the effect of breastfeeding by HIV-positive mothers found that exclusive breastfeeding leads to significantly lower rates of HIV transmission and higher rates of survival than does mixed feeding. “Infants who received formula milk in addition to breast milk, before or after 14 weeks of age, were nearly twice as likely” and “infants who were breastfed but also received solids were nearly 11 times” as likely to become infected than infants who were exclusively breastfed (Coovadia et al., 2007: 1113). HIV-positive women were provided during antenatal care, nevirapine, infant feeding counseling, and no cost commercial infant formula. After delivery, clinic nurses and counselors provided mothers with breastfeeding and replacement feeding support, with infant-feeding counselors visiting mothers three to four times within the first two weeks after birth and once every two weeks until six months after birth. Independent field monitors who visited mothers once a week assessed infant feeding practices. The study defined “exclusive breastfeeding” as feeding a child with breast milk, providing no solid food, and not giving non-human milk or water for more than three days total. After delivery, 1,132 mothers began exclusive breastfeeding, and the median duration of breastfeeding of infants for whom HIV test results were available was 159 days. Of the mothers who decided to exclusively breastfeed, 82% exclusively breastfed for at least 6 weeks, 67% exclusively breastfed for at least three months, and 40% exclusively breastfed for 6 months. The study found that 22% of exclusively breastfed infants died or became HIV-infected, resulting in an overall Kaplan-Meier estimated HIV-free survival rate of 75.4% at six months. The risk of HIV transmission was associated with low maternal CD4-cell counts. The study found that the health of mothers was strongly correlated with PMTCT. “Infants exclusively breastfed by women with CD4-cell counts less than 200 µL were twice as likely to become infected and almost four times more likely to die before 6 months of age than were infants exclusively breastfed by women with CD4-cell counts above 500 µL" (Coovadia et al., 2007: 1115).

A pooled analysis from 1,115 infants born to women living with HIV in Côte d’Ivoire and South Africa found that the overall risk of vertical transmission was twice as high among infants who breastfed for more than six months than among children who were breastfed for less than six months (Becquet et al., 2009b cited in Becquet et al., 2009a).

A study done in Zambia found that post-weaning breast milk HIV RNA was higher in women who stopped breastfeeding suddenly or used mixed breastfeeding. A total of 958 women living with HIV in the pre-ART era that breastfed were randomized into two groups: half of the women abruptly weaned at 4 months and the other half continued breast-feeding. Breast milk HIV RNA was measured at 1 and 4 months; and 2 weeks after weaning abruptly or at 4 months for those who continued breast-feeding. Seventy seven percent of the 154 women who stopped abruptly had viral RNA above 50 copies versus 40% of 394 women who exclusively breast- feeding at 4 months. The study also found that women who used mixed breast-feeding had significantly higher viral RNA levels. At four and half months 69% of women who used mixed feeding had viral RNA levels above 50 copies.

A 2001-2003 study that followed HIV-positive and HIV-negative pregnant women attending antenatal clinics in South Africa found that postnatal home visits offering infant feeding counseling significantly improved adherence to either exclusive breastfeeding or exclusive replacement feeding. The study followed 1,253 HIV-positive and 1,238 HIV-negative pregnant women who attended nine different clinics. Adherence was significantly associated with the number of antenatal feeding counseling home visits for both options. A breastfeeding counselor performed one antenatal home visit for every woman to discuss feeding options and three additional visits were available to those who chose to breastfeed. For women who chose to replacement feed, a specialist visited the home to teach methods of safe replacement feeding. The study also collected data on access to clean water, a refrigerator, fuel for boiling water and regular income for the mother, and found that only 3% of HIV-positive pregnant women had access to all four resources and 32.1% had access to all but regular income. “Of those who intended to replacement feed… few had the necessary resources to prepare infant formula safely” (Bland et al., 2007: 292). Infant formula became available in 2002 for HIV-positive pregnant women.

A study from 1999 to 2000 that provided VCT for women attending maternal and child health clinics for their first postpartum or well-baby visit in Botswana found that 937 or 54% of 1,735 postpartum women accepted VCT. 30% of those who accepted VCT were HIV-positive.

In a project in South Africa, maternal CD4 cell count was determined every six months during the infant’s immunization visit, with rapid referral for HAART for mothers with CD4 cell counts of less than 200/mm3.

One study found that "ANC as the only entry point in preventing MTCT will not capture all pregnant women" (Beltman et al., 2010: 1371). The study authors went on to suggest that "maternity and delivery wards could serve as alternative entry points in capturing [HIV-positive] pregnant women for the PMTCT programme" (Beltman et al., 2010: 1371), as well as infant and child health programs, particularly in high prevalence settings.

A randomized, double blind, placebo controlled trial in South Africa, Tanzania and Uganda from 2008 to 2010 found that the use of daily nevirapine for infants for six months of HIV-positive mothers who breastfed, had CD4 counts above 350 and did not receive HAART reduced the risk of vertical transmission by 48%. At the end of twelve months, for the 759 infants who were HIV-negative at randomization, the HIV status of infants at six months was 729 with known HIV status, 30 infants with HIV status not known, 9 infants deaths and 21 infants lost to follow up by the study. For the 763 infants who received placebo, at six months 733 had a known HIV status, 30 infants had unknown HIV status, 9 infants died and 21 infants were lost to follow up by the study. More than 95% of infants in both groups were no longer breastfed by the nine-month study visit. Among infants with 6 months of daily Nevirapine, 1.1% acquired HIV; in the placebo group, 2 to 4% of infants acquired HIV, equating to a 54% reduction in HIV transmission. When stratified by mothers who had HAART prior to randomization, infants did not have different rates of vertical transmission between the two groups. For infants born to mothers with CD4 counts above 350 who did not receive HAART found that the rate of vertical transmission for infants receiving daily nevirapine for six months, infants had a six months HIV infection rate of 0.7% compared to 2.8% in infants of mothers in the placebo group who did not receive daily Nevirapine for six months, a 75% reduction in vertically acquired HIV transmission. For women on HAART, there is no additional benefit to extend daily nevirapine use in infants from six weeks to six months. In addition, infants suffered high rates of adverse events. Loss to follow up by pregnant HIV positive women with CD4 counts above 350 in accessing HAART was not assessed. Infants whose mothers die are at high risk for mortality and infants who seroconvert to HIV positive who have had single dose Nevirapine have had problems with drug resistance. [See Antenatal Care - Treatment] A cost analysis of providing clean water as compared to NVP for infants for six months was not included as part of the analysis.

A study of HIV-positive infants in South Africa randomly assigned HIV-positive infants aged six to 12 weeks to HAART when CD4 was reduced to less than 20% (or 25% for infants less than one year of age) or clinical criteria were met or to immediate initiation of HAART until two years of age. At a median age of 74 weeks and a CD4 percentage of 35.2%, 125 infants were randomly assigned to receive deferred therapy and 252 infants were randomly assigned to receive early therapy. After a median follow up of 40 weeks, HAART was initiated in 66% of infants in the deferred therapy group. Twenty infants in the deferred therapy group (16%) died versus 10 infants in the early therapy groups (4%). Early HIV diagnosis and early HAART reduced infant mortality by 76%. 

Note: While What Works does not cover the vast topic of pediatric HIV treatment, this study, which took place in a resource-limited setting, is of great importance to parents living with HIV, and therefore has been included.