Review Antiretroviral drugs to prevent breastfeeding HIV transmission

Transcription

1 Antiviral Therapy : (doi: /IMP1574) Review Antiretroviral drugs to prevent breastfeeding HIV transmission Lynne M Mofenson 1 * 1 Pediatric, Adolescent and Maternal AIDS Branch, Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, Rockville, MD, USA *Corresponding author LM65D@nih.gov Observational studies and randomized clinical trials demonstrating that antiretroviral prophylaxis of the breastfeeding infant or triple-drug antiretroviral prophylaxis of the lactating mother can significantly affect the risk of postnatal transmission of HIV via breast milk have recently become available. In resource-limited countries, breastfeeding is a cornerstone of infant survival. While shortening the duration of breastfeeding by HIV-infected women reduces postnatal HIV transmission, increasing data suggest this may also decrease overall infant survival. Thus, there is a crucial need for interventions to allow safer and more prolonged breastfeeding. This paper will critically review the results of studies of postnatal antiretroviral prophylaxis to prevent breast milk HIV transmission. Introduction In 2008, an estimated 430,000 children were infected with HIV worldwide [1]. The vast majority of these infections were acquired by mother-to-child HIV transmission occurring in utero, intrapartum or through breastfeeding. Over 90% of these new paediatric infections occurred in sub-saharan Africa, where acquisition of HIV through breast milk accounts for an estimated 40% or more of new infections, but where breastfeeding has long been a cornerstone of child survival programmes. HIV-infected mothers in this setting are faced with an agonizing choice: to breastfeed but risk transmitting HIV to their infant or to not breastfeed and risk their infant dying of malnutrition or other infectious diseases. Exclusive breastfeeding can lower the risk of postnatal HIV transmission compared to mixed feeding, but it does not eliminate risk [2,3]. In well-resourced countries, such as the United States, early identification of HIV infection in pregnant women through universal routine opt-out antenatal HIV testing, provision of antiretroviral therapy for HIV-infected women who need treatment for their own health, provision of antiretroviral prophylaxis if therapy is not yet required, elective caesarean delivery and complete avoidance of breastfeeding has dramatically reduced the risk of mother-to-child HIV transmission to approximately 1 2% [4]. More than a decade ago, clinical trials identified simple, effective and relatively inexpensive antiretroviral interventions to prevent mother-to-child transmission during pregnancy and delivery that are applicable to resource-limited settings; however, implementation of these interventions has been slow. The proportion of pregnant women accessing antenatal HIV testing in midand low-resource countries has increased from 10% to 21% between 2007 and 2009, and the proportion of HIV-infected pregnant women receiving antiretroviral drugs for prevention of transmission has increased from 11% to 45% during the same period; however, almost one-quarter of pregnant women globally do not access any antenatal care and many countries continue to use less effective antiretroviral interventions, such as singledose nevirapine alone, for prevention of transmission [5,6]. Importantly, although the current antepartum and peripartum antiretroviral prophylaxis regimens reduce the risk of in utero and intrapartum transmission, the subsequent risk of HIV transmission through breastfeeding remains high and can result in postnatal transmission rates >15% with prolonged breastfeeding [3]. The only method that can eliminate breastfeedingassociated HIV transmission is to completely avoid breastfeeding. This is recommended in settings in which replacement feeding is affordable and sustainable, clean water is widely available, hygiene and sanitary conditions are good, and deaths caused by infectious diarrhoeal and respiratory diseases are relatively uncommon. This approach, however, is not feasible or safe in many 2010 International Medical Press (print) (online) 537

2 LM Mofenson resource-limited countries, where there are high rates of infant morbidity, mortality and malnutrition, and the high costs of formula, inadequate replacement foods to meet the nutritional needs of a growing infant, unsafe water supply, and stigma associated with not breastfeeding make replacement feeding untenable. In the past year, exciting new results from randomized clinical trials have become available demonstrating that antiretroviral prophylaxis of the breastfeeding infant or the lactating mother can significantly decrease the risk of postnatal acquisition of HIV. Available data suggest that implementation of these postpartum antiretroviral interventions, combined with treatment of HIV-infected pregnant and lactating women who require treatment for their own health and antepartum and intrapartum antiretroviral prophylaxis interventions for those not requiring treatment, could decrease mother-to-child transmission rates in resource-limited countries to <4%. The crucial need to provide antiretroviral therapy to pregnant and lactating women Clinical trials of interventions to prevent mother-to-child transmission have largely focused on use of antiretroviral drugs solely for prevention of transmission, without consideration of what might be optimal treatment for the mother. Given the inextricable link between maternal and infant survival, this approach is short-sighted, and only interventions that address maternal health as well as prevention of transmission are likely to provide maximal benefit. Although this paper is focused on antiretroviral prophylaxis to prevent postnatal HIV transmission, prevention of mother-to-child transmission should begin during pregnancy and must consider maternal health status. A key issue related to choosing an antiretroviral regimen for an HIV- infected pregnant woman (or for a postpartum lactating woman) is whether the antiretroviral drugs are being provided for treatment (in which case combination antiretroviral therapy should be provided and continued for life) or solely for prophylaxis of mother-tochild transmission (in which case less intensive regimens might be equally as effective as combination antiretroviral therapy, and the antiretroviral drug therapy would be stopped when the risk of transmission has ceased). The 2006 WHO guidelines on when to treat pregnant women recommended treating all women with WHO clinical stage 4; for women with WHO clinical stage 3, starting when CD4 + T-cell count is <350 cells/µl; but for women with WHO clinical stage 1 or 2, which constitute the vast majority of pregnant women, therapy was only recommended if CD4 + T-cell count was <200 cells/ µl [7]. In 2009, WHO updated these guidelines to recommend treatment for all HIV-infected adults, pregnant and non-pregnant, with CD4 + T-cell count <350 cells/µl regardless of clinical stage and for those with WHO clinical stage 3 or 4 disease regardless of CD4 + T-cell count [8,9], based on results from a clinical trial from Haiti that demonstrated improved survival and decreased morbidity in non-pregnant adults initiating treatment at a CD4 + T-cell threshold of <350 cells/µl compared with <200 cells/µl [10]. In pregnant and lactating women, both maternal mortality as well as risk of mother-to-child transmission is associated with CD4 + T-cell count, and use of a higher CD4 + T-cell threshold to determine when to initiate lifelong therapy is particularly important as it affects outcome for both mother and child. Looking specifically at postnatal transmission risk, in the absence of antiretroviral prophylaxis, month postnatal infection rates in infants uninfected at 4 6 weeks of age born to mothers with baseline CD4 + T-cell count <350 cells/µl were as high as 17%, whereas infants of mothers with CD4 + T-cell count >350 cells/µl had postnatal infection rates of 2 6% (Table 1) [11 15]. Data from Zambia indicate that approximately 92% of maternal deaths and 88% of perinatal or postnatal infections occur among women who would meet the new WHO criteria for treatment (CD4 + T-cell count <350 cells/µl or WHO clinical stage 3 or 4 disease) [16]. In a study in Kenya, overall maternal mortality at 1 and 2 years postpartum was 2% and 6%, respectively; all deaths occurred in women with CD4 + T-cell count <350 cells/µl at 32 weeks gestation (20/169; 12% died by 2 years postpartum), with no deaths in women with CD4 + T-cell count >350 cell/µl [17]. Because the majority of HIV-infected pregnant women are asymptomatic or have only mild symptoms (unlike adults enrolled in HIV care and treatment programmes who are often symptomatic) [18], it is critical that programmes that provide care to pregnant (and lactating) women provide access to CD4 + T-cell lymphocyte assays to determine which women need to initiate antiretroviral treatment for their own health benefit that would be continued for life. Interventions provided solely for prophylaxis of transmission, which stop after transmission risk has ceased (upon complete cessation of breastfeeding), should therefore theoretically be restricted to women without clinical symptoms and with CD4 + T-cell count >350 cells/µl. In studies from Kenya and Côte d Ivorie, approximately 52 65% of HIV-infected pregnant women seen in antenatal clinics had CD4 + T-cell count >350 cells/µl [17,19]. Prevention of postnatal transmission through breastfeeding using antiretroviral drugs Two types of antiretroviral interventions to prevent postnatal transmission have been evaluated in observational studies and clinical trials in resource-limited settings: provision of antiretroviral drugs to infants International Medical Press

3 ARV drugs and breast milk HIV transmission Table 1. Postnatal infant HIV infection rates in studies with no extended postnatal antiretroviral intervention by baseline maternal CD4 + T-cell count Study details Median BF Maternal CD4 + Total number Postnatal Study Transmission Treatment group duration, months T-cell count, cells/µl of women infection, % Reference HIVNET Month sdnvp arm, 18 < [11] transmission in n= uninfected at 56 days VTS, Ditrame 18-Month n=1,151 4 < [12] transmission in uninfected at 4 weeks PEPI-Malawi 18-Month Control arm 9 < [13] transmission in (sdnvp+1 week uninfected at 6 weeks AZT), n= HIVNET Month n=1,182 >12 < [14] transmission in uninfected at 4 weeks > Breastfeeding and 18-Month n=2, < [15] HIV International transmission in Transmission Study uninfected at 4 weeks >500 1, AZT, zidovudine; BF, breastfeeding; sdnvp, single-dose nevirapine. exposed to HIV during breastfeeding [20 25] and provision of combination antiretroviral prophylaxis to lactating women [25 32]. The design of these studies is shown in Table 2. Detailed comparative information for the infant prophylaxis studies is shown in Table 3, and for the maternal prophylaxis studies is shown in Table 4. Both of these strategies have been predicated on breastfeeding during the period of most benefit, followed by early weaning (for example, at or before age 6 months). Although the benefit of breastfeeding in terms of reducing infant mortality appears to be greatest in the first 6 months of life in resource-limited countries, significant benefit is furthermore observed throughout the first year of life [33]. Cessation of breastfeeding at age 6 months might be associated with increased morbidity and mortality in infants, as will be discussed later. There are major differences between studies, particularly between maternal and infant prophylaxis studies, which make it difficult to compare absolute HIV transmission rates between the studies. Reports often lack a 95% confidence interval to help understand the range of transmission encompassed by the intervention. The patient populations significantly differ; for example, all the infant prophylaxis studies except one (the Breastfeeding, Antiretroviral and Nutrition [BAN] study) enrolled women regardless of CD4 + T-cell count, whereas the three randomized trials of maternal prophylaxis (BAN, Kesho Bora and Mma Bana) restricted enrolment to women with CD4 + T-cell counts cells/µl (Table 2). Antepartum antiretroviral drug administration and duration (when given) differs significantly between the studies, yet is clearly important in terms of reducing in utero infection and comparisons of cumulative infection risk. The three large infant prophylaxis studies (the Six Week Extended Nevirapine [SWEN] study, Post-Exposure Prophylaxis of the Infant [PEPI-Malawi] study and BAN) enrolled women who had not received any antepartum drugs, whereas all the maternal prophylaxis studies except one (BAN) provided antepartum drugs (of different durations, ranging from initiation at weeks gestation, making comparison even among maternal studies difficult; Table 2). The duration of postnatal prophylaxis differs between studies, with two infant prophylaxis studies (SWEN and PEPI-Malawi) providing only 6 14 weeks of postnatal prophylaxis, whereas all of the eight maternal prophylaxis studies provided 6 months of postnatal prophylaxis (Table 2). The duration of breastfeeding, and hence the time at risk for postnatal infection, and rates of exclusive breastfeeding, which can affect postnatal transmission risk, differs between studies and is not specified in several maternal prophylaxis studies (Tables 3 and 4). Several studies do not provide birth infection rates, making it difficult to compare the incremental benefit of interventions during the breastfeeding period because the proportion of infections occurring in utero (and hence not affected by a postnatal intervention) cannot be determined (Tables 3 and 4). For example, the Mma Bana trial of maternal Antiviral Therapy

4 LM Mofenson Table 2. Design and drug regimens for infant and maternal antiretroviral postnatal prophylaxis studies Study name Study design Antepartum Intrapartum Postpartum Postpartum and location and enrolment regimen regimen mother regimen infant regimen Overall results for RCTs Reference Infant prophylaxis Mashi; Botswana Randomized trial; Starting at 34 weeks; First randomization; No ARV Second randomization; Arm 1: BF, At 7 months, HIV infection higher in [20] enrolled women Arm 1: AZT, Arm 2: AZT Arm 1: AZT; Arm 2: sdnvp+azt through age BF than FF (9.1% versus 5.6%), but regardless of CD4 AZT+sdNVP 6 months; Arm 2: FF, sdnvp+azt mortality lower (4.9% versus 9.3%); through age 1 month resulting in no difference in HIV-free survival (12.5% versus 12.9%) Mitra; Tanzania Open-label non- Starting at 36 weeks; AZT/3TC AZT/3TC 1 week AZT/3TC 1 week, then 3TC No comparison group (see Table 3 [21] randomized; enrolled AZT/3TC alone through age 6 months for results) women regardless of CD4 SIMBA; Uganda, Randomized trial; Starting at 36 weeks; AZT/ddI AZT/ddI 1 week Randomized; Arm 1: daily NVP No difference in HIV [22] Rwanda enrolled women AZT/ddI through age 20 weeks; Arm 2: infection between NVP and 3TC arms regardless of CD4 daily 3TC through age 20 weeks (overall 2.4% at 6 months) SWEN; Ethiopia, Randomized trial No ARV Arm 1: sdnvp; No ARV Randomized; Arm 1: sdnvp; Arm 2: In infants uninfected at birth, lower [23] Uganda, India (combined analysis Arm 2: sdnvp sdnvp+ daily NVP HIV infection in extended NVP than of three separate from age 7 days through 6 weeks control arm at 6 weeks (2.5% versus trials); enrolled women 5.3%; P=0.009) but not 6 months regardless of CD4 (6.9% versus 9.0%, P=0.16); Lower rates of death in extended NVP than control arm at 6 months (1.1% versus 3.6%, P=0.016) PEPI-Malawi, Randomized trial; No ARV sdnvp (if No ARV Arm 1: sdnvp+azt 1 week; In infants uninfected at birth, lower HIV [24] Malawi enrolled women presented early Arm 2: sdnvp+azt 1 week + infection in extended NVP and NVP/AZT Infant or maternal prophylaxis regardless of CD4 enough to receive) daily NVP starting at age 7 days than control arm at 14 weeks (2.8%, through 14 weeks; Arm 3: 2.8%, 8.4%, respectively) and 9 months sdnvp+azt 1 week + daily (5.2%, 6.4%, 10.6%, respectively) NVP/AZT from age 7 days but no significant difference between through 14 weeks extended (NVP versus NVP/AZT) arms BAN; Malawi Randomized trial; Arm 1: no ARV; Arm 1: AZT/3TC+ Arm 1: AZT/3TC 1 week; Arm 1: sdnvp+azt/3tc 1 week; Infants uninfected at 2 weeks had [25] enrolled women with Arm 2: no ARV; sdnvp; Arm 2: AZT/ Arm 2: AZT/3TC 1 week, Arm 2: sdnvp+azt/3tc lower HIV infection at 28 weeks than CD4 250 Arm 3: no ARV 3TC+sdNVP; Arm 3: then, from 7 days to 1 week; Arm 3: sdnvp+ control arm (6.4%) with infant AZT/3TC+sdNVP 6 months postpartum AZT/3TC 1 week, then daily NVP prophylaxis (1.8%, P< versus control) ABC, abacavir; ARV, antiretrovirals; AZT, zidovudine; BAN, Breastfeeding, Antiretrovirals and Nutrition study; BF, breastfed; CD4, CD4 + T-cell count, measured in cells/µl; ddi, didanosine; DREAM, Drug Resource Enhancement against AIDS and Malnutrition study; d4t, stavudine; EFV, efavirenz; FF, formula fed; KiBS, Kisumu Breastfeeding Study; LPV/RTV, lopinavir/ritonavir; NFV, nelfinavir; NVP, nevirapine; PEPI-Malawi, Post-Exposure Prophylaxis of the Infant study; RCT, randomized controlled trial; sdnvp, single-dose nevirapine; SIMBA, Stopping Infection from Mother to Child via Breastfeeding in Africa study; SWEN, Six Week Extended Nevirapine study; 3TC, lamivudine International Medical Press

5 ARV drugs and breast milk HIV transmission Table 2. Continued Study name Study design Antepartum Intrapartum Postpartum Postpartum and location and enrolment regimen regimen mother regimen infant regimen Overall results for RCTs Reference AZT/3TC/NVP (n=45), from age 7 days and maternal prophylaxis (3.0% modified 01/2005 to through 6 months P= versus control) AZT/3TC/NFV (n=141) and but no difference between 02/2006 to AZT/3TC/ maternal and infant prophylaxis LPV/RTV (n=664); Arm 3: arms (P=0.12) AZT/3TC 1 week Maternal prophylaxis DREAM; Mozambique Observational; Starting at 25 weeks; Continue CD4<350: continue regimen sdnvp+azt 1 week No comparison group (see Table 3 [26,27] enrolled women AZT/3TC/NVP regimen lifelong; CD4>350: for results) regardless of CD4; continue regimen through women choose mode 6 months postpartum of feeding KiBS; Kenya Observational; Starting at weeks; Continue CD4<200: continue regimen sdnvp No comparison group (see Table 3 [28] enrolled women AZT/3TC/NVP; mid-2005 regimen lifelong; CD4>200: for results) regardless of CD4 modified if CD4>250: continue regimen AZT/3TC/LPV/RTV through 6 months postpartum Mitra-Plus; Observational; Starting at 34 weeks Continue CD4<200 or WHO stage AZT/3TC 1 week No comparison group (see Table 3 [29] Tanzania enrolled women (earlier if CD4<200): regimen 3/4: continue regimen for results) regardless of CD4 AZT/3TC/NVP; 10/2005 lifelong; CD4>200: continue modified if CD4>200: regimen through 6 months AZT/3TC/NFV postpartum Amata; Rwanda Observational; Starting at 28 weeks; Continue CD4<350: continue regimen sdnvp+azt 1 week No comparison group (see Table 3 [30] enrolled women CD4<350 or WHO regimen lifelong; CD4>350: for results) regardless of CD4; stage 4: d4t/3tc/nvp; continue regimen through women choose mode CD4>350 and WHO stage 6 months postpartum of feeding 1/2/3: AZT/3TC/EFV Kesho Randomized trial; Starting at weeks; Arm 1: AZT/3TC/ Arm1: Continue Arm 1: sdnvp+azt HIV infection at birth not different [31] Bora; Kenya, enrolled women with Arm 1: AZT/3TC/LPV/RTV; LPV/RTV; Arm 2: AZT/3TC/LPV/RTV through 1 week; Arm 2: sdnvp+azt between antepartum maternal triple South Africa, CD Arm 2: AZT AZT+sdNVP 6 months postpartum; 1 week (no postnatal prophylaxis versus AZT (1.8% Burkina Faso Arm 2: AZT/3TC prophylaxis) versus 2.2%, respectively); HIV infection 1 week (no postnatal at 12 months lower with maternal prophylaxis) triple prophylaxis than AZT with no postnatal prophylaxis (5.5% versus 9.5%) Mma Bana Randomized trial; Starting at weeks: Arm 1: AZT/3TC/ABC; Arm 1: continue AZT/3TC/ABC Arm 1: sdnvp+azt HIV infection at 6 months not different [32] Study; Botswana enrolled women with Arm 1: AZT/3TC/ABC; Arm 2: through 6 months postpartum; 4 weeks; Arm 2: sdnvp+azt between AZT/3TC/ABC versus AZT/3TC/ CD4>200 Arm 2: AZT/3TC/LPV/RTV AZT/3TC/LPV/RTV Arm 2: continue 4 weeks LPV/RTV (1.8% versus 0.4%; P=0.53) AZT/3TC/LPV/RTV through 6 months postpartum Antiviral Therapy

6 LM Mofenson Table 3. Study results for infant antiretroviral postnatal prophylaxis studies Baseline median maternal Study name, Maternal antepartum and Number CD4, cells/µl and HIV transmission at HIV transmission at location and design infant ARV prophylaxis of infants HIV RNA, copies/ml Infant feeding birth and 4 6 weeks a 6 7 months a HIV or death a Reference Mashi; Botswana; Mother: AZT 34 weeks to FF, 591; BF, 588 BF, CD4: 372; HIV FF and BF; BF median Birth: 3.3% (19/558 BF); Cumulative at 7 months: 6.1% at 4 weeks; [20] Randomized trial delivery (±sdnvp); RNA: 4.35 log duration 5.9 Cumulative at 4 weeks: 4.6% 9.0% (51/541 BF); 12.9% at 12 months Infant: AZT to 6 months months; 51% (27/557 BF); Increment Increment from 4 weeks to if BF (±sdnvp) EBF at 3 months day 1 to 4 weeks: 1.3% 7 months: 4.4% Mitra; Tanzania; Mother: AZT/3TC 36 BF, 398 CD4: 411 (15.4% BF only, median Birth: ND; Cumulative at 6 months: 4.5% ( ) at [21] Observational weeks to 1 week CD4<200); duration 18 weeks Cumulative 4.9% ( ); 6 weeks; 8.5% postpartum; Infant: HIV RNA: at 6 weeks: Increment from 6 weeks ( ) AZT/3TC 1 week, then not specified 3.8% ( ) to 6 months: at 6 months daily 3TC to 6 months 1.2% (0 2.4) SIMBA; Uganda, Mother: AZT/ddI 36 BF, 397 CD4: 427; HIV RNA: BF only, median No difference between arms; Cumulative at 6 months: Not specified [22] Rwanda; weeks to 1 week (199 randomized Not specified at duration Birth: 6.0% (24/397 infants); 7.6% (5 14; 30/397 infants); Randomized trial postpartum; Infant: to 3TC, 198 to NVP baseline days Cumulative at 4 weeks: Increment from 4 weeks randomized at birth to (~3.3 months) 6.8%; Increment from to 6 months: 0.8% daily 3TC or NVP to 1 week to 4 weeks: 0.8% (3/358 infants) 6 months (3/373 infants) SWEN; Ethiopia, Mother: Late presenter, BF, 2,074 (placebo CD4: 397; HIV RNA: BF only; most Extended NVP arm: Extended NVP arm: Extended NVP arm: [23] Uganda, India; no antepartum ARVs; 1,047, extended NVP 16,457 17,400 weaned between Birth: 4.7%; Cumulative Cumulative at 6 months: 3.7% at Randomized trial Infant: sdnvp, and 977); Uninfected at 14 weeks (73% BF) at 6 weeks: 7.2%; 11.6%; Increment from 6 weeks; 8.1% (combined analysis randomized to daily birth and data and 6 months Increment from day 1 6 weeks to at 6 months 3 separate trials) placebo versus at 6 months: (31% BF); to 6 weeks: 2.5% 6 months: 4.4% extended NVP from placebo, 928; extended 54% EBF day 8 to 6 weeks NVP, 831 at 14 weeks PEPI-Malawi, Mother: late presenter, BF, 3,016 (placebo, 989; CD4: ; HIV BF only, most Extended NVP or NVP/AZT: Cumulative at 6 months: Extended NVP arm: [24] Malawi; no antepartum extended NVP, 993; RNA: not specified weaned between Birth: 7.1% both; Cumulative 11.1% extended NVP, 3.3% at 6 weeks; 6.6% Randomized trial ARVs; Infant: sdnvp extended NVP/AZT, 980); 6 months (90% BF) at 6 weeks: 8.8% 12.3% extended NVP/AZT; at 6 months; 13.9% at +1 week AZT, Uninfected at birth and and 9 months extended NVP and 8.7% Increment from 6 weeks 12 months; Extended randomized to daily data at 9 months: (29 32% BF); NVP/AZT; Increment from to 6 months: NVP/AZT arm: 1.8% at placebo versus NVP placebo, 788; extended 61 64% EBF at day 1 to 6 weeks: 2.3% extended 6 weeks; 8.2% versus NVP/AZT from NVP, 800; extended 6 months 1.7% extended NVP and NVP, 3.6% at 6 months; day 7 to 14 weeks NVP/AZT, % NVP/AZT extended NVP/AZT 15.0% at 12 months BAN; Malawi Mother: No antepartum BF: infant NVP arm, CD4: 440; HIV RNA: BF only; duration not Birth: enrolled at Cumulative at 7 months Uninfected at 2 weeks: [25] Infant prophylaxis ARVs; Infant: sdnvp+ 848 enrolled not specified specified (counseled delivery, rates based on (uninfected at 2 weeks): Extended infant arm; Randomized AZT/3TC 1 week, then to wean at 6 months) uninfected at 2 weeks; 1.8%; Increment NVP arm, 2.9% trial for women daily NVP to age Cumulative at 6 weeks from 2 weeks at 7 months with CD4>250 6 months not specified to 6 months: 1.8% a Range values shown in parentheses are 95% confidence intervals. ARV, antiretroviral; AZT, zidovudine; BAN, Breastfeeding, Antiretroviral and Nutrition study; BF, breastfed; CD4, CD4 + T-cell count, measured in cells/µl; ddi, didanosine; EBF, exclusive breastfeeding; FF, formula fed; ND, no data; NVP, nevirapine; PEPI-Malawi, Post-Exposure Prophylaxis of the Infant study; sdnvp, single-dose nevirapine; SIMBA, Stopping Infection from Mother to Child via Breastfeeding in Africa study; SWEN, Six Week Extended Nevirapine study; 3TC, lamivudine International Medical Press

7 ARV drugs and breast milk HIV transmission Table 4. Study results for maternal antiretroviral postnatal prophylaxis studies Baseline median Timing of maternal maternal CD4, Study name, triple ART cells/µl and HIV HIV transmission at HIV transmission at location and design administration Number of infants RNA, copies/ml Infant feeding birth and 4 6 weeks a 6 7 months a HIV or death Reference DREAM; Mozambique; Median 26.8 weeks 985 Mothers enrolled; CD4: 489; HIV FF and BF; BF duration Birth: ND; Cumulative at Cumulative at 6 months: 5.3%; Not specified [26] Observational gestation to 707 Infants tested at RNA: 4.27 log not specified 4 weeks: 3.8% ( ) Increment from 4 weeks to 6 months 1 month, 467 infants (counselled to wean 6 months: 1.5% ( ) postpartum if BF tested at 6 months; Did at 6 months) not specify % FF versus BF DREAM; Mozambique; 25 Weeks gestation FF, 891 (data on 809 Not specified FF and BF; BF duration Birth: ND; Cumulative at Cumulative at 6 months: 2.2% Not specified; [27] Observational to 6 months infants); BF, 341 infants not specified 4 weeks: 1.2% (4/341 BF) ( ; 6/266 BF); Increment Infant mortality at postpartum if BF tested at 1 month, 251 (counselled to wean from 4 weeks to 6 months: 6 months: 2.85 per infants tested at 6 months at 6 months) 0.8% ( ; 2/251 BF) 100 person-years KiBS; Kenya; 34 Weeks gestation BF, 497 infants CD4: 394 (24% BF only; Duration not Birth: 2.4% ( ); Cumulative at 6 months: 5.0% Not specified [28] Observational to 6 months CD4 250); specified (counselled to Cumulative at 6 weeks: ( ); Increment postpartum HIV RNA: 4.5 log wean at 6 months) 3.9% ( ); Increment from 1 week and 6 months: 2.6% from day 1 to 6 weeks: 1.5% Mitra-Plus; Tanzania; 34 Weeks gestation 501 Mothers enrolled, 441 CD4: 415 (17.5% BF only; Median Birth: ND; Cumulative at Cumulative at 6 months: 5.0% 8.6% (6 11.2) at [29] Observational to 6 months with data; BF 441 infants CD4<200); duration 24 weeks 6 weeks: 4.1% ( ; ( ; 22/397 infants); Increment 6 months; 12.8% postpartum HIV RNA: 14,621 18/423 infants) from 6 weeks to 6 months: 1.0% (9.6 16) at 12 months Amata; Rwanda; 28 Weeks gestation 562 Mothers enrolled, 551 CD4: Mean 498 FF (57%) and BF (43%); BF; Birth: 1.3% (3/227 infants); BF: Cumulative at 9 months: BF: 4.9% (3 9; 7/216 [30] Observational to 7 months delivered, 532 infants alive (IQR ); BF duration not Cumulative at 6 weeks: 1.8% ( ; 4/227 infants); infants) at 9 months postpartum if BF after 2 days; BF 227 infants HIV RNA: Not specified; (counselled 1.3% ( ); Increment Increment from 6 weeks to specified to wean at 6 months) from birth to 6 weeks: 0% 6 months: 0.5% (1/227 BF) Mma Bana Study; Weeks BF, 560 mothers enrolled, CD4: ; BF only; 71% BF for Birth: Maternal triple-drug Cumulative at 6 months: 1.1% Not specified; Infant [31] Botswana; Randomized gestation to 553 infants with 6 month HIV RNA: 5 months but <1% prophylaxis arms combined: ( ); Increment from mortality at 6 months: comparative trial for 6 months data 9,100 13,300 after 6 months; 93% 0.7% (4/553 infants); No birth to 6 months: 0.4% 2.5% women with CD4>200 postpartum; reported EBF 6-week data but no intrapartum Median duration to weaning transmission; thus, increment 11 weeks from birth to 4 weeks: 0% Kesho Bora; Kenya, Weeks BF: maternal triple-drug CD4: 335; HIV FF (23%) and BF (77%); Birth: 1.8% ( ); Cumulative at 6 months: 4.9% Maternal triple-drug [32] South Africa, Burkina gestation to prophylaxis arm, 413 RNA: Not median duration Cumulative at 6 weeks: 3.3 ( ); Increment from prophylaxis arm: 8.3% Faso; Randomized trial for 6.5 months mothers enrolled, 402 specified 21.4 weeks; 46% EBF ( ); Increment from 6 weeks to 6 months: 1.6% (6 11.5) at 6 months; 10.4% women with CD postpartum live births at 3 months birth to 6 weeks: 1.5% ( ) at 12 months BAN; Malawi Maternal CD4>250; BF: maternal triple-drug CD4: 428; HIV BF only; duration not Birth: Enrolled at delivery; Cumulative at 7 months Uninfected at 2 weeks: [25] prophylaxis arm; Delivery to prophylaxis arm, RNA: Not specified (counselled to Rates based on uninfected (uninfected at 2 weeks): 3.0%; Maternal triple-drug Randomized trial for 6 months 851 mothers enrolled specified wean at 6 months) at 2 weeks; Cumulative at Increment from 2 weeks to prophylaxis arm: 4.7% women with CD4>250 postpartum 6 weeks: not specified 6 months: 3.0% at 7 months a Range values shown in parentheses are 95% confidence intervals. ART, antiretroviral therapy; BAN, Breastfeeding, Antiretroviral and Nutrition study; BF, breastfed; CD4, CD4 + T-cell count measured in cells/µl; DREAM, Drug Resource Enhancement against AIDS and Malnutrition study; EBF, exclusive breastfeeding; FF, formula fed; IQR, interquartile range; KiBS, Kisumu Breastfeeding Study; ND, no data. Antiviral Therapy

8 LM Mofenson prophylaxis (Tables 2 and 4), which reported the lowest rate of infant infection of all studies to date (cumulative infection 1% at age 6 months) had a median maternal CD4 + T-cell count at enrolment of approximately 400 cells/µl, baseline HIV RNA levels prior to initiation of prophylaxis of 9,100 13,300 copies/ml, median duration of antepartum prophylaxis of 11 weeks, a proportion of infants exclusively breastfeed through age 6 months of 93%, and an adherence rate to maternal prophylaxis of 93% [32]. By contrast, in the maternal prophylaxis arm of the Kesho Bora trial (Tables 2 and 4), where the 6-month cumulative HIV infection rate was 4.9%, the median maternal CD4 + T-cell count at enrolment was 335 cell/µl, median duration of antepartum prophylaxis was 6 weeks, and only 48% of infants exclusively breastfed to 3 months [31]. Thus, it is difficult to make direct comparisons even between studies of similar interventions. Given these caveats, the currently available data suggest that provision of antiretroviral drugs to the breastfeeding infant could have comparable efficacy to provision of maternal combination antiretroviral prophylaxis to the lactating mother. Tables 3 and 4 provide data on six infant prophylaxis studies and eight maternal prophylaxis studies aimed at reducing postnatal transmission that have been published or presented at meetings as of February 2010, and includes data on the timing and type of prophylaxis, numbers enrolled, maternal CD4 + T-cell count and HIV RNA level (when available), infant feeding and duration of feeding, transmission rates at birth, 4 6 weeks and 6 7 months, the incremental risk of early (before 4 6 weeks) and late (between 4 6 weeks and 6 7 months) postnatal infection (when available) and rates of HIV or death (HIVfree survival, when available) [20 32]. Because of differences among the studies in administration of maternal antepartum antiretroviral drugs, comparison of cumulative rates of transmission is misleading when trying to compare the effect of the interventions to reduce breast milk transmission. This is because the infection rate at birth, reflecting in utero infection, will be lower if the mother has received drugs during pregnancy than if she received no drugs, and further will be lower with longer than with shorter duration of antepartum drug administration. Therefore, when considering the effect of the postpartum intervention, the better comparison is the rate of infection at 4 6 weeks or 6 7 months in infants who are uninfected at birth; however, many of the maternal prophylaxis studies do not provide information on infection rates at birth, and, as a result, only the comparison of late postnatal infection occurring between 4 6 weeks and 6 7 months was possible between maternal and infant strategies. Another problem is that the duration of the actual postnatal intervention also differs between the studies. Specifically, two large infant prophylaxis randomized trials provided 6 and 14 weeks of prophylaxis, whereas all the maternal studies provided 6 months of prophylaxis; thus, comparisons of late transmission might also be misleading. In the four maternal prophylaxis and the four infant prophylaxis studies with adequate information for a meaningful comparison of early postnatal infection rates (that is, studies that provided transmission rates at birth to allow assessment of in utero infection and 4 6 week data to allow description of the increment in infection between birth and 4 6 weeks of age), the rate of infection at age 4 6 weeks in infants uninfected at birth with maternal prophylaxis was 0% in Amata and Mma Bana studies, 1.5% in the Kisumu Breastfeeding Study (KiBS), and 1.5% in the Kesho Bora study (Table 4) [28,30 32], and with infant prophylaxis was 0.8% in the Stopping Infection from Mother to Child via Breastfeeding in Africa (SIMBA) study, 1.3% in the Mashi study, 1.7% in the PEPI-Malawi study and 2.5% in the SWEN study (Table 3) [20,22 24]. Thus, the early (between birth and age 4 6 weeks) postnatal infection rates appear relatively similar with either maternal (range 0 1.5%) or infant (range %) interventions. Although the ability to evaluate late postnatal infection between 4 6 weeks and 6 7 months of age is possible for most of the studies, it is important to note that in some of the infant prophylaxis studies the intervention stops at 6 14 weeks. In the maternal prophylaxis studies, the rates of late postnatal infection are 0.4% (Mma Bana study), 0.5% (Amata study), 0.8% (Drug Resource Enhancement against AIDS and Malnutrition [DREAM] study [27]), 1.0% (Mitra-Plus study), 1.5% (DREAM study [26]), 1.6% (Kesho Bora study), and 2.6% (KiBS study; Table 4) [25 32]. In the infant prophylaxis studies, in which infant prophylaxis was given for 6 months as in the maternal studies allowing a comparison of prophylaxis over similar time periods, the rates of late postnatal infection were 0.8% (SIMBA study), 1.2% (Mitra study) and 4.4% (Mashi study; Table 3) [20 22]. Of note, the one infant prophylaxis study with the highest rate of late infection (4.4%) gave infant zidovudine prophylaxis, whereas all the others used nevirapine or lamivudine [20]. The late infection rate in the PEPI-Malawi study, where infant prophylaxis stopped at 14 weeks, was 2.3% [24]. The SWEN study only administered 6 weeks of infant prophylaxis and therefore no prophylaxis was being received during the period of late transmission risk (after age 6 weeks) [23]. Thus, in the evaluable studies, the late (4 6 weeks to 6 months) postnatal infection rates also appear relatively similar with either maternal (range %) or infant (range %) interventions. In studies that provided data on the end point of HIV International Medical Press

9 ARV drugs and breast milk HIV transmission or death, comparisons at age 6 7 months ranged from % in the maternal prophylaxis studies and % in infant prophylaxis studies. Taken together, the early and late postnatal infection rates appear relatively low and similar with either maternal or infant interventions if being compared during similar periods of prophylaxis. The Mitra study of infant prophylaxis and Mitra-Plus study of maternal prophylaxis provide a non-randomized comparison of interventions, as both studies were conducted sequentially in the same clinics, both provided some maternal antepartum antiretroviral prophylaxis, and both provided the same duration (6 months) of postnatal prophylaxis (Table 2) [21,29]. The cumulative transmission risk at 6 months was 4.9% with infant prophylaxis in Mitra and 5.0% with maternal prophylaxis in Mitra-Plus, and the risk of late transmission between 6 weeks and 6 months was 1.2% with infant prophylaxis and 1.0% with maternal prophylaxis (Tables 3 and 4). Data from a randomized comparison of maternal and infant interventions for prevention of postnatal transmission is available from the BAN study, which compared 6 months of maternal prophylaxis or infant nevirapine prophylaxis to a control short-course arm with no maternal or infant prophylaxis during breastfeeding (Table 2) [25]; this study evaluated an intrapartum and postpartum intervention (no maternal antenatal drugs were received) and enrolled women with CD4 + T-cell counts >250 cells/µl. Postnatal transmission rates at age 28 weeks (7 months) in infants uninfected at age 2 weeks were 6.4% in the control arm compared with 3.0% in the maternal prophylaxis arm (P= versus control) and 1.8% in the infant nevirapine arm (P< versus control; Table 2). Although the transmission rate in the infant nevirapine arm appeared lower than in the maternal prophylaxis arm, there was no significant difference between maternal and infant prophylaxis arms (P=0.12) and the study was not powered to detect a difference between the two experimental arms. The data from these studies also indicate the importance of providing antiretroviral drugs during the antepartum period, and that longer antepartum prophylaxis is more effective than shorter prophylaxis. In the maternal prophylaxis studies where initiation of antiretroviral prophylaxis started at weeks gestation, overall 6 7 month transmission rates were 1 5% (Table 4) [26 32]. In the infant prophylaxis studies in which maternal antepartum prophylaxis was given but started significantly later, at weeks gestation, overall 6 7 month transmission rates were 5 9% [20 22], whereas in the infant prophylaxis studies where no maternal antepartum drugs were received, overall 6 month transmission rates were approximately 11 12% [23,24] (Table 3). Thus, for optimal prevention of mother-to-child transmission, it is crucially important to identify HIV-infected women early in pregnancy and initiate prophylaxis by at least 28 weeks gestation, if not earlier. Given two presumably similarly effective interventions, the choice of intervention to prevent motherto-child transmission for women who do not require treatment for their own health will involve weighing a number of different considerations, including relative costs, feasibility, and risks and benefits of the interventions (Table 5). Few studies have evaluated the cost-effectiveness of these two interventions [34]. When administering maternal triple-drug combination therapy solely for prevention of transmission, the risks of maternal drug toxicities, treatment interruption following prolonged exposure to three drug combination prophylaxis during pregnancy and breastfeeding (presuming treatment stops after cessation of breastfeeding), and of fetal exposure to multiple drugs, need to be weighed against the incremental benefit in preventing transmission compared to less complex regimens in the population of women who do not yet need treatment for their own health. Infant prophylaxis is less expensive but might be programmatically more difficult to implement if combined with use of zidovudine plus single-dose nevirapine plus a tail to reduce nevirapine resistance and prevent in utero/intrapartum infection. None of the studies to date have been adequately powered to address the comparative efficacy and safety of maternal versus infant prophylaxis, and although there are studies planned (Table 6), data will not be available for several years. Safety of infant and maternal antiretroviral prophylaxis Table 7 summarizes toxicity reports from the infant and maternal prophylaxis studies. Evaluation of safety can be best assessed in the randomized clinical trials. The SWEN and PEPI-Malawi large randomized clinical trials of 6 and 14 weeks of extended daily infant nevirapine prophylaxis (Table 2), respectively, together enrolled 5,090 infants [23,24]. There were no significant differences in infant adverse events (including rash and hepatic events) between the experimental and control arms (control arm was single-dose nevirapine in SWEN and single-dose nevirapine plus 1 week of infant zidovudine in PEPI-Malawi). The SWEN trial found significantly lower infant mortality at 6 months in the extended nevirapine group compared with the control arm, whereas the PEPI-Malawi trial found similar rates of mortality at 9 months between the two extended infant prophylaxis arms and the control arm (Table 7). In the PEPI-Malawi trial, infants in the extended dual nevirapine/zidovudine prophylaxis arm had a Antiviral Therapy

10 LM Mofenson significantly higher number of infant adverse events deemed possibly related to study drug (primarily neutropaenia) compared with the control group (Table 7) [24]. Thus, the addition of zidovudine potentially increased toxicity of the extended infant prophylaxis, without increasing efficacy, as there was no significant difference in efficacy between the two extended prophylaxis arms [24]. The Mashi trial randomized infants to formula feeding with 1 month of infant zidovudine or to infant breastfeeding with 6 months of infant zidovudine. There were significantly higher rates of grade 3 or 4 signs or symptoms in infants receiving 6 months versus 1 month of zidovudine and higher rates of grade 3 or 4 laboratory abnormalities, particularly neutropaenia, similar to the PEPI-Malawi trial (Table 7) [20]. Finally, the SIMBA trial compared extended infant prophylaxis with daily lamivudine or daily nevirapine; there were no significant differences in serious adverse events or mortality between the two study arms [22]. The BAN trial compared 6 months of maternal tripledrug combination prophylaxis and 6 months of infant daily nevirapine prophylaxis to a control intervention of single-dose nevirapine with 1 week of zidovudine/ lamivudine (Table 2). Mothers in the triple-drug combination prophylaxis arm had significantly higher rates of grade 3 or 4 neutropaenia than those in the infant Table 5. Issues relating to choice of infant versus maternal postnatal antiretroviral prophylaxis for preventing breast milk HIV transmission for women with CD4>350 cells/µl who do not require treatment for their own health Infant antiretroviral prophylaxis (combined with AZT/sdNVP Maternal antiretroviral prophylaxis (given during Factor with 1-week antiretroviral tail to prevent resistance) pregnancy and through breastfeeding postpartum) Comparative Similar effectiveness to maternal triple-drug combination Same as adjacent. efficacy prophylaxis for prevention of in utero infection in women with high CD4 (Kesho Bora trial in women with CD , transmission at birth 2.2% with AZT/sdNVP regimen versus 1.8% with maternal triple-drug regimen). No study with adequate power to provide comparative efficacy of infant versus maternal postnatal prophylaxis in women with CD4>350, although available data suggest at minimum similar efficacy. Cost Inexpensive. More expensive. Pregnancy No indication of adverse effect on pregnancy outcome. Potential increase in preterm delivery or low birth weight outcome safety with maternal triple-drug regimen seen in some studies in resource-limited and resource-rich countries. Maternal safety Use of intrapartum sdnvp can be associated with NVP Possible higher rates of haematological toxicity. resistance in mother, hence need for AZT/3TC (or other Effect on maternal health of interruption of drugs drug) tail to prevent resistance. No maternal drugs postpartum. after prolonged use during pregnancy and breastfeeding and of repeated courses with subsequent pregnancies is a concern. No data on prophylaxis extended past 6 months. Infant safety Up to 6 months of daily NVP prophylaxis is safe. No data Infant exposed to subtherapeutic levels of antiretroviral on more prolonged prophylaxis yet. drugs in maternal milk, although BAN trials suggests similar safety between infant and maternal prophylaxis. Resistance If infant is receiving prophylaxis during breastfeeding and If mother is receiving triple-drug during breastfeeding and becomes infected, drug resistance likely. infant becomes infected, drug resistance is likely. Choice of drug Extended prophylaxis with infant NVP as effective as More complicated: NVP hepatotoxicity a concern; in NVP/AZT (PEPI-Malawi study) resource-rich countries, protease-inhibitor-based regimens used for prophylaxis; use of EFV for prolonged periods postpartum a concern regarding risk for teratogenicity should woman become pregnant again. Complexity Use of AZT/sdNVP and tail is more complex. Use of single regimen during pregnancy and postpartum might be easier to implement. Assessment of CD4 prior to initiation important to determine which CD4 prior to initiation important as maternal drugs are mother for women require treatment for own health. CD4 could be stopped after infection risk ceases (should not stop treatment need obtained and antepartum AZT (or postnatal NVP) initiated drugs if maternal CD4<350). Regimen likely to be (for example, while waiting for result; women with CD4<350 can then different for women with CD4>350 where NVP cannot be CD4 count) be switched to cart for own health which also provides safely used. postnatal prophylaxis. AZT, zidovudine; BAN, Breastfeeding, Antiretroviral and Nutrition study; cart, combination antiretroviral therapy; CD4, CD4 + T-cell count, measured in cells/µl; EFV, efavirenz; NVP, nevirapine; PEPI-Malawi, Post-Exposure Prophylaxis of the Infant study; sdnvp, single-dose nevirapine; 3TC, lamivudine International Medical Press

11 ARV drugs and breast milk HIV transmission Table 6. Planned studies on prevention of mother-to-child HIV transmission Study name Infant feeding Antepartum Intrapartum Postpartum Postpartum and location Study design and enrolment regimen regimen mother regimen infant regimen Comments ANRS-PEP; Burkina Postpartum intervention BF; n=1,500 Infant enrolled Infant enrolled Infant enrolled postpartum Arm 1: sdnvp at birth; To start in 2010 Faso, South Africa, comparing infant prophylaxis (in planning) postpartum postpartum 3TC from day 7 to Uganda, Zambia during BF for 9 months to 9 months; Arm 2: sdnvp no prophylaxis at birth; placebo from day 7 to 9 months Promoting Maternal For women who do not require BF and FF; Resource CD4>350 starting Arm 1: sdnvp+tdf/ftc CD4>350; Postpartum Postpartum Start: US/South Infant Survival treatment for own health limited: approximately at 14 weeks; (for 7 days randomization at day 7: Arm 3: randomization at day 7: America, late fall Everywhere (CD4>350); Will address: 8,000 mother infant Antepartum postpartum); Arm 2: infant NVP prophylaxis to Arm 3: sdnvp+azt ; Resource- (PROMISE- What is optimal antepartum/ pairs; randomization; Arm 1: AZT/3TC/LPV/RTV (for cessation of BF (up to days; NVP from day 7 to limited countries, IMPAACT 1077); postpartum regimen to US/South America: AZT; Arm 2: maternal 7 days postpartum) 18 months) with no maternal cessation of BF (up to summer 2010 Multiple countries prevent MTCT? Is it safe for 2,000 mothers triple-drug prophylaxis ARV; Arm 4: maternal 18 months); Arm 4: in Africa, Asia, mothers to stop triple-drug AZT/3TC/LPV/RTV triple-drug prophylaxis: TDF/ sdnvp+azt 7 days; South America, US prophylaxis after receiving it FTC/LPV/RTV to cessation Infant health only for PMTCT? How can we of BF (up to 18 months); randomization: BF, infant maintain health of uninfected Maternal Health randomization uninfected and infants after weaning? to stop or continue maternal <12 months at time of BF Resource-limited countries: triple-drug ARV: If FF, cessation; Arm 1: Antepartum/intrapartum/ on maternal triple-drug Cotrimoxazole to postpartum intervention in prophylaxis and CD4>350, 18 months: Arm 2: women with CD4>350, randomize at day 7 12; Cotrimoxazole placebo sequential randomizations; if BF, on maternal triple- to 18 months US/South America: only drug prophylaxis, randomize participate in the Maternal at cessation of BF; US/South Health randomization America: If triple-drug prophylaxis during pregnancy and CD4>400, randomize after delivery ANRS 12200; Antepartum/intrapartum/ BF and FF; n=unknown Starting at 20 weeks; Arm 1: TDF/FTC/EFV; If BF, continue 6 9 months Arm 1: AZT 1 week? In planning, to Côte d Ivoire postpartum intervention Arm 1: Maternal Arm 2: postpartum; Arm 1: Arm 2: AZT 1 week? start 2010 comparing two triple-drug regimen: TDF/FTC/EFV; AZT/3TC/LPV/RTV TDF/FTC/EFV; Arm 2: regimens to prevent MTCT; Arm 2: Maternal AZT/3TC/LPV/RTV Randomized trial regimen: AZT/3TC/LPV/RTV (non-inferiority design) ANRS, Agence Nationale de Recherches sur le Sida; ARV, antiretrovirals; AZT, zidovudine; BF, breastfed; CD4, CD4 + T-cell count, measured in cells/µl; EFV, efavirenz; FF, formula fed; FTC: emtricitabine; LPV/RTV, lopinavir/ritonavir; MTCT, mother-to-child transmission; NVP, nevirapine; PEP, post-exposure prophylaxis; PMTCT, prevention of mother-to-child transmission; sdnvp, single-dose nevirapine; TDF, tenofovir; 3TC, lamivudine. Antiviral Therapy