d e c e m b e r 2 0 0 9 T H E S O U T H E R N A F R I C A N J O U R N A L O F H I V M E D I C I N E HIV infection has a high prevalence in antenatal at- tendees in South Africa. In the annual seroprevalence survey conducted through the National Department of Health from 2005, the prevalence in the Western Cape province was 15.7%.1 Here, a pilot zidovudine (ZDV)- based prevention of mother-to-child transmission (PMTCT) programme began in 1999,2 and has gradually been expanded since January 2001. Since April 2003 the PMTCT interventions have been available at all public sector antenatal service facilities in the prov- ince (300 antenatal clinics and 53 delivery centres and hospitals) and 350 primary health care clinics where infant follow-up occurs. Attendees are offered volun- tary, confidential counselling and testing (VCT) and if HIV positive, antiretrovirals (ARVs) for the mother and infant. Uptake was reported as 97% in 2006 (Status Report – Prevention of Mother-to-Child Transmission Programme, 14 July 2006, HIV/AIDS/STI Directorate, A WINDOW INTO A PUBLIC PROGRAMME FOR PREVENTION OF MOTHER-TO-CHILD TRANSMISSION OF HIV: EVIDENCE FROM A PROSPECTIVE CLINICAL TRIAL O R I G I N A L A R T I C L E M Cotton1, FCPaed (SA), PhD S Kim2, ScD (Biostats) H Rabie1, FCPaed (SA) J Coetzee1, CPN, PN S Nachman3, MD for the PACTG 1041 team 1Centre for Infectious Diseases, Department of Paediatrics and Child Health and KID-CRU, Tygerberg Children’s Hospital and Faculty of Health Sciences, Stellenbosch University, Tygerberg, W Cape 2Department of Biostatistics, Harvard School of Public Health, Boston, Mass, USA 3Division of Infectious Diseases, State University of New York, NY, USA Objectives. To evaluate efficacy of the antenatal, intrapartum and postnatal antiretroviral components of a public service prevention of mother-to-child (PMTCT) programme in infants. Design. Analysis of prospectively collected screening data of demographic and MTCT-related interventions and HIV infection status of infants identified through HIV-specific DNA polymerase chain reaction. Setting. Tygerberg Children’s Hospital, Western Cape, South Africa. Subjects. HIV-infected women and their infants identified through participation in a public service PMTCT pro- gramme were referred for possible participation in a prospective study of isoniazid prophylaxis. Interventions. Key components of the programme include voluntary counselling and testing, administration of zidovudine to the mother from between 28 and 34 weeks’ gestation and to the newborn infant for the first week, single-dose nevirapine to the mother in labour and to the newborn shortly after birth, and free formula for 6 months. Main outcome measures. Number and percentage of HIV-infected infants and extent of exposure to antenatal, intrapartum and postnatal antiretrovirals. Results. Of 656 infants with a median age of 12.6 weeks, screened between 1 April 2005 through May 2006, 39 were HIV-infected, giving a transmission rate of 5.9% (95% confidence interval (CI) 4.4 - 8.0%). Antenatal prophylaxis was significantly associated with reduced transmission (odds ratio (OR) 0.43 (95% CI 0.21 - 0.94)) as opposed to intrapartum and postpartum components (p=0.85 and p=0.84, respectively). In multivariable analysis the antenatal component remained significant (OR=0.40 (95% CI 0.19 - 0.90)). Conclusions. The antenatal phase is the most important antiretroviral component of the PMTCT programme, allowing most opportunity for intervention. 16 T H E S O U T H E R N A F R I C A N J O U R N A L O F H I V M E D I C I N E d e c e m b e r 2 0 0 9 Western Cape). Follow-up of mother and infant, co- trimoxazole from 6 weeks of age and modified infant feeding practices are also important components. The majority of women (95%) choose formula feeding, which is provided free for the first 6 months. The initial ARV intervention was single-dose nevirap- ine (sd-NVP) to mother and infant, introduced after the success of the HIVNET 012 study.3 Since mid-2003, ZDV was added from 34 weeks’ gestation for the mother and for a week for the neonate.4 In early 2006, an- tenatal ZDV from 28 weeks was gradually introduced. With the advent of the national antiretroviral rollout in 2004, all pregnant women with a CD4 count below 200 cells/µl were offered highly active antiretroviral therapy (HAART). PACTG 1041 is a prospective phase III clinical trial eval- uating the efficacy of isoniazid (INH) primary prophy- laxis in HIV-exposed infected and uninfected infants. Through screening for this trial at Tygerberg Children’s Hospital (TCH), we had the opportunity to evaluate the PMTCT programme in referred infants. HIV-exposed infants between 3 and 4 months of age were referred for study participation from health care facilities in the urban and semi-rural areas close to TCH. Referring clinics were requested not to refer in- fants exposed to TB. Infants were pre-screened by lay counsellors and nurses for eligibility to enrol in PACTG 1041. Exclusion criteria included exposure to tubercu- losis and not receiving bacille Calmette-Guérin (BCG) immunisation within the first week of life. Eligible sub- jects were entered onto a screening log, comprising the dataset for the present report. A medical doctor undertook formal screening. Note was taken of the extent of participation in the PMTCT programme and whether the mother received HAART in pregnancy. Those receiving either antenatal HAART or ZDV were categorised as having received antenatal prophylaxis. Intrapartum prophylaxis was either ZDV or NVP or both. Postnatal intervention to the neonate was either ZDV or NVP or both. HIV DNA polymerase chain reaction (PCR) was per- formed on all exposed infants eligible for the trial. All samples were tested in duplicate. For discordant re- sults, the test was repeated in duplicate. Simple percentages were used to estimate rates of transmission and 95% confidence intervals (CIs) were based on the score method. Medians and interquar- tile ranges (IQR) were used to summarise continuous data. Logistic regression was used to evaluate the ef- fectiveness of PMTCT components. Odds ratios (ORs) were calculated through the logistic regression model and 95% CIs are based on the profile likelihood. For antenatal HAART, Fisher's exact test was used because of the small cell size. In multivariable logistic regres- sion analysis, we evaluated all two-way and three-way interactions between the three PMTCT programme components and found no statistically significant in- teractions, and therefore present a multivariable model that includes main effects for the three components. All tests are two-sided at the 5% significance level and are not adjusted for multiple comparisons. Analyses were done using SAS 9.1 (Cary, NC, USA). Permission to conduct P1041 and to report on ante- natal interventions was obtained from the Committee for Pharmaceutical Trials, Stellenbosch University, and the Medicines Control Council of South Africa. The trial was approved by the National Institute of Allergy and Infectious Diseases (NIAID) according to the Office of Human Rights Protection, National Institutes of Health guidelines. Between 1 April 2005 and 31 May 2006, 773 infants were referred for pre-screening. Seven infants were excluded because their mothers were HIV-negative and had been referred in error. One hundred and ten HIV-exposed infants were excluded at pre-screening, of whom 52 (47.3%) had known exposure to tubercu- losis.5 Other common reasons included the infant being too old for participation in the INH trial (15), BCG given after 7 days of life (15) and family relocating (7). Six hundred and fifty-six infants were entered onto the screening log and are reported on here. The median age (IQR) of infants was 12.6 (11.0 - 13.6) weeks and that of mothers 26 (23 - 30) years. Thirty-nine of 656 infants had a positive HIV DNA PCR, giving a transmis- sion rate of 5.9% (95% CI 4.4 - 8.0%). Two hundred and seventy-eight (85.8%) of 324 mothers who were asked decided on exclusive formula feeding. Data on transmission and extent of participation in the PMTCT programme are shown in Table I. We evaluated the antenatal, intrapartum and postnatal components of the programme separately, using logistic regression (Table II). We found that antenatal prophylaxis was significantly associated with a reduction in rate of transmission (OR 0.43 (95% CI 0.21 - 0.94), p=0.035) as opposed to intrapartum and postnatal components (p=0.85 and p=0.84, respectively). The results of fitting a multivariable logistic regression model to the data, which included the three compo- nents of the PMTCT programme, are shown in Table III. 17 METHODS RESULTS d e c e m b e r 2 0 0 9 T H E S O U T H E R N A F R I C A N J O U R N A L O F H I V M E D I C I N E 18 The antenatal component of the PMTCT regimen re- mained significant in the multivariable model, indicat- ing that it is an independent predictor of decreased transmission (OR 0.40 (95% CI 0.19 - 0.90), p=0.027). Of 57 mothers with CD4 cell counts below 200/µl receiving HAART, only 1 (1.8% (95% CI 0.3 - 9.3%)) transmitted HIV to her infant versus 26 of 453 with CD4 counts >200/µl not receiving HAART (5.7% (95% CI 3.9 - 8.3%), p=0.34). Fifty-one (7.8%) mother/infant pairs with missing PMTCT information were excluded from the above analyses. Of these, only 1 woman (2.0% (95% CI 3.5 - 10.3%)) transmitted HIV to her infant. There is little information on vertical transmission in the absence of intervention in South Africa. Transmission rates vary between 15% and 34%.6,7 In the ZDV-based pilot programme in Khayelitsha, Western Cape, the transmission rate was 11%.2 In a study evaluating sd- NVP in different South African settings, the transmission rate in Paarl, a recruitment site for P1041, was 8.3% at 3 weeks of age.8 The combination of antenatal ZDV from 28 weeks and sd-NVP under optimal circumstances is associated with a transmission rate as low as 1.1%.4 Our data confirm a relatively effective PMTCT pro- gramme despite only 53% actually participating in all components of the programme. Importantly, an addi- tional 25% received antenatal ARVs. A multi-faceted intervention programme means that there are many opportunities for intervention, as opposed to one rely- ing on only a single intervention such as sd-NVP, which, if missed, severely compromises efficacy. The World Health Organization has endorsed the programme as practised in the Western Cape.9 Our data confirm the relative importance of antenatal as opposed to perina- tal or postnatal intervention. The transmission rate of 5.9% was achieved due to concerted efforts to facili- tate success of the programme despite widespread per- ceived obstacles to initial implementation.10 Although we did not record duration of antenatal ARVs in each case, 75% of mothers were reported to have received ≥2 weeks of therapy, defined as adequate by the pro- gramme (personal communication – Pauline Pieters, PMTCT Co-ordinator, 20 September 2006). An important preliminary finding in our study is that among women with CD4 cell counts <200/µl, only 1 of 57 mothers on HAART (1.8%) transmitted HIV, as op- posed to 26 (5.7%) of 453 mothers receiving ZDV. Al- though this was not statistically significant, we expect- PMTCT component Adjusted OR (95% CI) p-value* Antenatal 0.40 (0.19 - 0.90) 0.027 Intrapartum 1.26 (0.57 - 2.89) 0.57 Postnatal 1.05 (0.46 - 2.60) 0.91 *Likelihood ratio test. TABLE III. MULTIVARIABLE ANALYSIS Received component/total (%) Infant not Infant OR Programme component HIV infected HIV infected (95% CI) p-value† Antenatal PMTCT 483/568 (85.0) 27/38 (71.1) 0.43 (0.21 - 0.94) 0.035 Intrapartum PMTCT 380/568 (66.9) 26/38 (68.4) 1.07 (0.54 - 2.25) 0.85 Postnatal PMTCT 425/568 (74.8) 29/38 (76.3) 1.08 (0.52 - 2.48) 0.84 Antenatal HAART* 56/564 (9.9) 1/38 (2.6) 0.25 (0.01 - 1.52) 0.24 *Exact CI and Fisher’s exact test p-value provided for antenatal HAART due to the small number of infections among those who received antenatal HAART. † Likelihood ratio test from logistic regression model, except for antenatal HAART which is a Fisher’s exact test. TABLE II. UNIVARIATE ANALYSES DISCUSSION Infants HIV-infected Screened (N (%))* (N (% of screened)) PMTCT 656 39 (5.9) Full participation 348 (53.0) 21 (6.0) Antenatal only 105 (16.0) 4 (3.8) Antenatal + postnatal (no intrapartum) 42 (6.4) 1 (2.4) Antenatal + intrapartum 15 (2.3) 1 (6.7) Postnatal only 39 (5.9) 4 (10.3) Intrapartum only 18 (2.7) 1 (5.6) Intrapartum + postnatal 25 (3.8) 3 (12.0) None 13 (2.0) 3 (23.1) No data 51 (7.8) 1 (2.0) *Percentages do not add to 100.0 due to rounding. Antenatal = either ZDV or HAART; intrapartum and postnatal = sd-NVP, ZDV. TABLE I. EXTENT OF PARTICIPATION OF MOTHERS AND THEIR INFANTS IN THE PMTCT PROGRAMME AND HIV TRANSMISSION RATES T H E S O U T H E R N A F R I C A N J O U R N A L O F H I V M E D I C I N E d e c e m b e r 2 0 0 9 ed more transmission among these mothers because of their low CD4 counts before initiation of HAART. There are a number of limitations to our study. We did not record the mothers’ CD4 counts; rather, we as- sumed that they were appropriately managed accord- ing to the PMTCT guidelines. Also, we only screened in- fants whose mothers expressed interest in their infants participating in the INH study. Nevertheless, our data are similar to those of the Department of Health, West- ern Cape, which reports a transmission rate of 6.2% (personal communication – Pauline Pieters, PMTCT Co- ordinator, 20 September 2006). There were missing PMTCT data on 7.8% of screened mother/infant pairs, and the transmission rate was low in this group (2.0%). If we assume that none of these women received any component of PMTCT or that these women received all components of PMTCT, the conclusions drawn here remain unchanged. The antenatal ARV component is extremely important for reduction of MTCT and reduces intra-uterine infec- tion, For example, Lallemant et al. showed that initiat- ing ZDV at 28 weeks was far more effective than at 35 weeks and was not compensated for by extending postnatal ZDV for 6 weeks.11 Nevertheless, the intra- partum and postnatal components also have an im- portant role. For example, Wade et al. showed that in the absence of ZDV, perinatal transmission of HIV was 26.6% (95% CI 21.1 - 32.7%).12 When ZDV was be- gun antenatally, the transmission rate was 6.1% (95% CI 4.1 - 8.9%). Administration of postnatal ZDV alone within 48 hours of life had a transmission rate of 9.3% (95% CI 4.1 - 17.5%), indicating the importance of the postnatal component.12 Gray et al. found postnatal sd- NVP to be slightly more effective than ZDV for 6 weeks when neither antenatal nor intrapartum ARVs could be given.13 The public service PMTCT programme in the Western Cape has successfully reduced the vertical transmis- sion of HIV. The antenatal ARV component is critical for success. REFERENCES 1. National Department of Health. National HIV and Syphilis Seroprevalence Survey South Africa 2005. Pretoria: National Department of Health, 2005. 2. Abdullah MF, Young T, Bitalo L, Coetzee N, Myers JE. Public health lessons from a pilot programme to reduce mother-to-child transmission of HIV-1 in Khayalitsha. S Afr Med J 2001; 81: 579-583. 3. Guay LA, Musoke P, Fleming T, et al. Intrapartum and neonatal single- dose nevirapine compared with zidovudine for prevention of mother-to-child transmission of HIV-1 in Kampala, Uganda: HIVNET 012 randomised trial. Lancet 1999; 354: 795-802. 4. Lallemant M, Jourdain G, Le Coeur S, et al. Single-dose perinatal nevirapine plus standard zidovudine to prevent mother-to-child transmission of HIV-1 in Thailand. N Engl J Med 2004; 351(3): 217-228. 5. Cotton MF, Schaaf HS, Lottering G, Weber HL, Coetzee J, Nachman S. Tuberculosis exposure in HIV-exposed infants in a high-prevalence setting. Int J Tuberc Lung Dis 2008; 12(2): 225-227. 6. Bobat R, Coovadia H, Coutsoudis A, Moodley D. Determinants of mother-to-child transmission of human immunodeficiency virus type 1 infection in a cohort from Durban, South Africa. Pediatr Infect Dis J 1996; 15(7): 604-610. 7. The PETRA study team. Efficacy of three short-course regimens of zidovudine and lamivudine in preventing early and late transmission of HIV-1 from mother to child in Tanzania, South Africa, and Uganda (Petra study): a randomised, double- blind, placebo-controlled trial. Lancet 2002; 359: 1178-1186. 8. Good Start Study Team. The Good Start Study: Early Perinatal Transmission, Infant Feeding and HIV-Free Survival. Paper presented at the 25th Conference on Priorities in Perinatal Care, 7 - 10 March 2006, Drakensberg, KwaZulu-Natal. 9. World Health Organization. Antiretroviral Drugs for Treating Pregnant Women and Preventing HIV Infection in Infants in Resource-limited Settings: Towards Universal Access. Recommendations for a Public Health Approach. Geneva: WHO, 2006. 10. Delva W, Draper B, Temmerman M. Implementation of single-dose nevirapine for prevention of MTCT of HIV – lessons from Cape Town. S Afr Med J 2006; 96: 706-709. 11. Lallemant M, Jourdain G, Le Coeur S, et al. A trial of shortened zidovudine regimens to prevent mother-to-child transmission of human immunodeficiency virus type 1. Perinatal HIV Prevention Trial (Thailand) Investigators. N Engl J Med 2000; 343(14): 982-991. 12. Wade NA, Birkhead GS, Warren BL, et al. Abbreviated regimens of zidovudine prophylaxis and perinatal transmission of the human immunodeficiency virus. N Engl J Med 1998; 339(20): 1409-1414. 13. Gray GE, Urban M, Chersich MF, et al. A randomized trial of two postexposure prophylaxis regimens to reduce mother-to-child HIV-1 transmission in infants of untreated mothers. AIDS 2005; 19(12): 1289-1297. 19 Acknowledgements We thank the KID-CRU PACTG team: G Lottering, H S Schaaf, H C Weber, J Karpakis, H L Weber, M Louw, V Ntlokondala, E Thompson, C Janse van Rensburg, P Ketelo, G Boswell, M Tizora, K Smith, N Dlaku, L Hoorn, N Mpotololo and L Kwane. We also thank Najmaar Shaikh, senior public health specialist and infectious disease epidemiologist, Provincial Gov- ernment of the Western Cape, for helpful advice. Support for this study was provided by the NIAID of the US National Institutes of Health (NIH), grant number UO1 A1 41809. The content of this pub- lication does not necessarily reflect the views or policies of NIAID, nor does mention of trade names, commercial projects, or organisations imply en- dorsement by the US Government. The authors de- clare no conflicts of interest. CONCLUSIONS