Kieffer, Mary Pat MSc*; Nhlabatsi, Bonisile RN†; Mahdi, Mohammed MD, MPH*; Hoffman, Heather J PhD‡; Kudiabor, Kwashie MPH*; Wilfert, Catherine M MD*
Provision of antiretroviral (ARV) prophylaxis for prevention of mother-to-child transmission of HIV (PMTCT) has been one of the major successes in public health. Transmission in developed countries has dropped to <2%,1 and in countries where health resources and access to services are limited, transmission rates below 10% have been documented.2-5 However, worldwide scale-up of services to prevent HIV vertical transmission has faced many challenges. As countries look toward elimination of pediatric AIDS, finding new ways to address these challenges becomes increasingly important.
Swaziland, with a population of approximately 1 million people, has the highest HIV seroprevalence in pregnant women in the world, increasing from 3.2% in 1992 to 42% in 2008.6 Swaziland has made significant progress in scaling up identification of pregnant women with HIV infection and delivery of ARV treatment and prophylaxis during pregnancy. By 2008, PMTCT services were integrated into 80% (137 of 172) health facilities providing antenatal care.7 An estimated 67% of pregnant women with HIV received ARV prophylaxis in 2007.8 With antenatal clinic (ANC) prevalence of 42%, however, any gap in delivery of PMTCT interventions has a large population impact in terms of additional infant HIV infections.
The PEARL study provided comprehensive analysis of the PMTCT cascade in 4 countries, showing multiple points where opportunities were lost.9 Ultimately, only half of the HIV-infected study women and their infants received the nevirapine (NVP) component of their treatment or prophylaxis. Efforts to improve access and uptake of PMTCT have heavily focused on reaching women in ANC clinics. Offering PMTCT services to women during labor is the last opportunity for pregnant women with HIV to access preventive services before delivery, when the largest proportion of perinatal HIV transmission occurs.10
Pregnancy has been shown to be a vulnerable time for new HIV infection. Gray showed a 2-fold increased risk of HIV infection in pregnant and breastfeeding women, primarily attributed to biological and hormonal changes and after controlling for sexual behavior.11 Research in Rwanda demonstrated higher HIV incidence for women up to 6 months postpartum,12 and a study in Zimbabwe showed a high incidence of seroconversion among young mothers during the period of 2 years postpartum.13
Since 2003, the Elizabeth Glaser Pediatric AIDS Foundation (EGPAF) has supported the Swaziland Ministry of Health (MOH) to scale-up PMTCT with opt out HIV testing, HIV treatment for those eligible (zidovudine/lamivudine/NVP), and ARV prophylaxis for treatment ineligible mothers and infants. Maternal prophylactic regimens included either combination ARV prophylaxis with zidovudine/single-dose NVP (sdNVP)/zidovudine-lamivudine tail where possible or sdNVP if not. Swaziland PMTCT guidelines also stipulated that women presenting to maternity units with unknown status and those who tested HIV negative more than 3 months prior should be counseled and offered HIV testing before delivery. An assessment in 2007 found that these activities were not routinely done, leading to several missed opportunities for PMTCT in maternity. Based on the assessment, the team sought ways to improve staff performance to address these missed opportunities.
MATERIALS AND METHODS
EGPAF-Swaziland, developed a targeted training intervention for midwives in maternity wards to improve their performance as the critical pathway to improved PMTCT service delivery. The training focused on specific strategies to maximize PMTCT effectiveness depending on a woman's HIV status (positive, negative, or unknown) at the time of arrival in maternity. The EGPAF research team, in partnership with the Sexual and Reproductive Health Unit of the Swaziland MOH, designed a study to measure the impact of the training intervention on ARV coverage among HIV-infected women at delivery. The protocol was approved by the Scientific and Ethics Committee of the Swaziland MOH.
The study used a quasi-experimental design comparing 3 intervention maternity sites with 3 control maternity sites to test the hypothesis that targeted training for maternity nurses would lead to an increase in identification of women with HIV infection who receive and adhere to ARV prophylaxis before delivery. Secondary objectives included determining the HIV seroincidence rate during late pregnancy and testing whether provision of NVP prophylaxis to women who refused HIV testing would significantly increase ARV coverage in women with HIV infection. Detection of NVP in HIV antibody-positive cord bloods from mother-infant pairs was used as the marker for ARV prophylaxis or combination antiretroviral therapy (ART).
A convenience sample of 6 maternity wards, representing more than 60% of deliveries in the country, was selected from the 7 EGPAF-supported maternity sites across Swaziland. All sites received the same level of supportive supervision and technical assistance from EGPAF. Sites, matched for type of facility and volume of deliveries, included 2 large urban hospitals, 2 rural hospitals, and 2 rural health centers to support applicability across the health care system. One site of each type was designated as an intervention site and one as a control site by a coin flip. The 3 facilities designated as control sites did not receive the study intervention. All pregnant women presenting for delivery at these 6 maternity facilities from October 2008 to January 2009 were eligible for study participation after verbal consent was obtained and documented by health care workers. Women at intervention sites who refused HIV testing but who agreed to take sdNVP signed a separate written consent.
All midwives in the study facilities had previously attended the Swaziland National PMTCT training, which focuses heavily on approaches in ANC but includes minimal information about providing services in maternity settings. A targeted 1-day training course was provided to increase knowledge and skills in provision of PMTCT and to enhance confidence and skills in counseling, with the intended program outcome of increasing uptake of HIV testing and ARV prophylaxis for PMTCT. The training focused on critical activities for maternity personnel beginning with the importance of ascertaining the HIV status of all women. Integrating testing women of unknown status, retesting women who tested negative 3 or more months prior, and ensuring that all HIV-positive women receive and take their ARV prophylaxis or treatment as standard practice in maternity were emphasized. Offering empiric NVP prophylaxis to women of unknown status who refused HIV testing was conducted on a limited basis for the purpose of this study and did not replace the standard of care or the PMTCT package offered in maternity.
In the 6 maternity units, 120 MOH midwives (81% of total maternity staff) were trained by study personnel on the protocol and study procedures. In intervention sites, 69 midwives (88%) received the training intervention before initiation of study data collection. Study data were collected on questionnaires and abstracted from routinely collected data in registers and participant records. HIV status on arrival was determined through examination of the woman's ANC card. If not available, women were categorized as “unknown” HIV status. Standardized questionnaires and cord blood samples were collected anonymously but linked through a unique study identification number issued at the time of consent. Cord blood samples were collected as dried blood spots (DBS) by the trained personnel from discarded placentas after delivery. HIV counseling and rapid HIV testing followed Swaziland National Guidelines (Determine initial test, UniGold confirmation). Extra work was compensated equally at all sites based on the number of completed questionnaires and cord blood samples.
DBS samples were stored at room temperature, batched, and sent weekly to Groote Schuur Hospital's virology and pharmacology laboratories (South African National Accreditation System-certified laboratories affiliated with the University of Cape Town). Samples from women with negative or unknown HIV status on arrival in maternity were tested in the virology laboratory using Abbott Architect HIV Ag/Ab Combo enzyme-linked immunosorbent assay. (Abbott Laboratories, Abbott Park, IL) Samples that tested positive, along with samples from women identified as HIV positive on arrival in maternity, were confirmed with Siemens Enzygnost Anti-HIV 1/2 Plus enzyme immunoassay. (Siemens Healthcare Diagnostic Products, Marburg, Germany) Detection of NVP in cord blood was used as a marker of adherence to ARV prophylaxis or treatment. The NVP assay was performed on all HIV-positive samples in the pharmacology laboratory by liquid chromatography-mass spectrometry (API 3200 triple quadrupole), lower limit of detection: 100 ng/mL.
Data were entered into the study database by trained personnel at EGPAF-Swaziland offices using EpiInfo version 3.5.1. Double date entry was used with comparison and verification.
Based on 2007 program data on ARV uptake from the MOH and findings from the PEARL Study, baseline maternal ARV coverage was estimated at 80%, with the intervention estimated to improve the proportion to 90%. Using a 5% significance level, approximately 199 HIV-positive women were required in each of the control and intervention groups to achieve 80% power to detect a 10% improvement in coverage with the intervention. Assuming that 40% of participants in each group were HIV positive, the total number of participants required in each group was 498. Therefore, total sample size required to test the research hypothesis was 996 pregnant women and was increased for the secondary analyses to 2400 women.
Contingency tables and χ2 tests were used to summarize data and test for associations between frequencies of events. The χ2 test statistics for specified proportions were calculated and reported for 1-way tables, whereas Pearson χ2 test statistics were calculated and reported for 2-way tables. All 2-way test statistics were used to test the null hypothesis of no association between the row and column variables. If more than 20% of expected cell counts were less than 5, Fisher exact test was applied. Test statistics and the corresponding P values are reported along with relative risks (RRs). For determination of HIV incidence rates, women who seroconverted from HIV negative to HIV positive were censored at the midpoint of the time since their last HIV negative test. Data analysis was done with SAS Version 9.1.
Data on HIV testing and self-reported NVP use were available from maternity registers of 3001 women who arrived in maternity during the study period (Table 1). Of these, 2444 (81%) were enrolled in the study and had a study questionnaire completed (Fig. 1). Only 1 woman refused to participate; the remaining women were not approached for enrollment due to lack of time, staff, or personnel trained on the study. Both forms and DBS were collected from 2386 (98%) study participants. There were no significant differences in testing status on arrival between enrolled women in control and intervention sites (Table 1).
Of 2403 cord blood samples collected at the intervention and the control sites, 923 were positive for HIV antibodies in the laboratory and tested for NVP. The sample from 1 woman was excluded from analysis because she was receiving a non-NVP-containing ART regimen. Of the remaining 922 HIV-positive samples, 689 (75%) had detectable NVP levels. NVP coverage was significantly higher (80% = 369 NVP positive/459 HIV positive) at intervention sites compared with 69% (320 of 463) at control sites (RR = 1.16; confidence interval: 1.08 to 1.25; P < 0.0001). Logistic regression in the primary analysis controlling for type of site (urban hospital, rural hospital, and rural health center) found no significant differences (data not shown).
There were 141 HIV positive cord blood results (15% of total positive samples) in women who were not known to be HIV positive on arrival in maternity. NVP was detected in 56% of these women (32 of 57) in intervention sites compared with only 23% of women (19 of 84) in control sites (Table 2).
The cord blood samples from 15 of the 813 women identified as HIV positive at enrollment were HIV negative on retesting. They were excluded from the primary analysis but included in secondary analysis because the staff performance was based on assumed HIV status. These may represent false-positive results from ANC or specimen labeling error either in the ANC or maternity. NVP was detected in 9 of these 15 specimens.
Women Identified as HIV Positive on Arrival at Maternity
Of 813 women identified as HIV positive on arrival at maternity, 206 (25%) were on ART, 419 (52%) said they had swallowed their NVP prophylaxis at home, and 188 (23%) indicated that they had not taken their NVP. Control sites reported more women on ART (29% vs. 22%; P = 0.02) and more women receiving prophylaxis who reported taking their NVP dose before arrival (76% vs. 64%; P = 0.0001) compared with intervention sites. However, this was not associated with an increased proportion of women with NVP detected in cord blood (Table 2).
Cord blood analysis was available for 796 of 813 women identified as HIV positive on arrival (17 were missing cord blood samples). There were no differences between intervention and control sites in the rate of NVP detected in cord bloods of women receiving ART (92% vs. 88%; P = 0.30) or women with known positive status who were not on ART (81% vs. 75%; P = 0.08) (Table 2). Women who reported not taking NVP at home were just as likely to have NVP in their cord blood as those who reported taking it before arrival due to NVP dosing in maternity.
Women Who Previously Tested in ANC and Identified as HIV Negative on Arrival at Maternity
According to maternity registers, 2144 HIV-negative women (including nonenrolled women) arrived at study maternity units during the study period. Significantly more HIV-negative women were retested at intervention sites, 45% (528 of 1185), compared with only 14% (135 of 959) at control sites (RR = 3.17; 95% confidence interval: 2.67 to 3.74; P < 0.0001). For enrolled women who tested HIV negative in ANC, 1377 cord blood results were available, representing a total of 388 person-years of observation. Fifty-eight women had HIV antibody positive results, giving a cumulative incidence of seroconversion during pregnancy of 4% in this population, with an incidence rate of 16.8 new infections per 100 person-years. NVP prophylaxis was detected in 54% (13 of 24) of these newly identified positive women in intervention sites compared with only 26% (9 of 34) of women in control sites (P = 0.03) (Table 2).
Women Arriving at Maternity With Unknown HIV Status
Using register data, the proportion of all women arriving in maternity with unknown status who were tested in intervention sites was 96% (153 of 159) compared with 65% (102 of 156) in control sites (P < 0.0001). Although rates of HIV infection in enrolled women with previously unknown status were similar in intervention and control sites (35% and 42%, respectively), uptake of NVP prophylaxis nearly tripled in intervention sites (Table 2). Of 33 newly identified HIV-positive women in intervention sites, 19 (58%) had NVP detected in cord blood as compared with only 10 of 50 (20%) in control sites (P = 0.0004) (Fig. 2).
NVP Prophylaxis for Women of Unknown HIV Status
Only 1 woman in the intervention sites refused HIV testing in maternity and remained with unknown status. She was offered and accepted sdNVP.
This study demonstrates that a targeted 1-day training intervention for maternity nurses had significant impact on the number of HIV-positive women who received ARV prophylaxis through improved PMTCT service delivery in Swaziland. The intervention significantly increased NVP coverage in HIV antibody-positive cord blood specimens. Participating midwives, who had all previously received the Swaziland national PMTCT training, reported that the targeted training vastly improved their skills and confidence. The study intervention improved adherence to the delivery portion of the overall PMTCT cascade by addressing potential missed opportunities for prevention in maternity.
That 15% of the total HIV-positive cord blood samples were from women not known to be HIV positive on arrival in maternity clearly shows the importance of strengthening PMTCT services at maternity if countries are to reach elimination of mother-to-child transmission. These additional infections can be attributed to those women who arrived with unknown status and were positive (the largest contribution), to those with a previously HIV negative test who had either seroconverted or become newly infected, or to those who had an incorrect test result from ANC or maternity due to specimen mix-up or laboratory error. Ensuring all women receive HIV testing before delivery must be a priority to identify HIV-exposed infants who need postpartum prophylaxis and women who need HIV care.
In intervention sites, women who did not take their labor dose before arrival in maternity were just as likely to have NVP in cord blood as those who did, showing the value of working with maternity staff to improve their skills, both technical and counseling.
The study intervention was not able to document increased reliability of self-reported NVP intake. Other studies have also shown variation in the concordance of NVP presence in cord blood with women's self-reported NVP ingestion.14,15
Despite the relatively high utilization of ART among pregnant women in Swaziland and given the importance of ART for eligible pregnant women and the half-life of NVP, it is concerning that only 89% of pregnant women on NVP-containing ART regimens had detectable NVP levels in cord blood.
Only 1 woman refused HIV testing in maternity, suggesting that offering sdNVP to women of unknown status would have limited value in settings such as Swaziland with high HIV testing rates.
This study has several strengths including the large number of women, different types of facilities studied, and high rates of data and specimen collection. Although 19% of women in the maternities were not enrolled in the study, data on all women were examined from maternity registers (some included here) with no major differences identified in those enrolled and not enrolled. There were no exclusion criteria for enrollment, and the opportunities to “choose” enrollees were extremely limited in both intervention and control sites; thus, it is unlikely that there was a conscious choice of which clients to enroll that would have altered the conclusions of the study. Women who were not enrolled in the study likely still benefited from the enhanced care that resulted from the training at the intervention sites.
There are some limitations to this study. The determination of seroconversion was based on the date and result of a previous HIV negative test recorded on the woman's health card, but the original result was not directly verified. There is a possibility that some women who were actually HIV positive were misclassified as HIV negative on arrival and were subsequently counted as women who seroconverted. Some women whose records show NVP provision in maternity had none detected in their cord blood. Data on the length of time between ingestion of NVP and the delivery of the infant were not recorded; thus, the NVP may have been ingested too close to delivery for detection using the study assay or it was not actually ingested. This limitation applies to both the control and intervention sites, so would not have affected the statistical significance of study findings, but could have lowered the overall rate of NVP coverage reported.
HIV testing is an essential step in PMTCT and is an indicator by which PMTCT program quality is often measured. Focused training of maternity midwives led to significant increases (P < 0.0001) in HIV testing among women who had tested HIV negative ≥3 months before and also among women who had never tested. Twice as many new infections were detected in intervention sites, allowing women and their infants to receive ARV prophylaxis and appropriate follow-on care. Maternity wards remain the last opportunity to establish the presence of HIV before delivery and provide access to critical PMTCT interventions for mother and baby. Recent World Health Organization recommendations that include prophylaxis to prevent HIV transmission through breastfeeding16 makes identification of HIV infection in maternity even more valuable.
The importance of retesting HIV-negative women in high prevalence settings is underscored by the high levels of new HIV infections found in late pregnancy in this study, which are consistent with seroconversion rates of 5%-10% reported from South Africa17 and from 2008 to 2009 in Swaziland.18 Women who seroconvert during pregnancy have acute or recent infection with higher viral loads, with transmission rates to infants reported as high as 67%.19 Detection of incident HIV infections is often delayed or missed due to the window period for rapid tests; however, a large proportion of incident infections during late pregnancy go undetected not because of the limits of technology but due to lack of attention. Based on these results, Swaziland has incorporated third trimester, maternity, and postnatal retesting into its 2010 guidelines.
The level of seroconversion during pregnancy in Swaziland and other HIV prevalence countries is alarmingly high, despite excellent HIV testing coverage and uptake of ARV prophylaxis among HIV-positive pregnant women in ANC. This suggests that programs are giving significant attention to women who test positive in ANC but that women who test negative are not receiving adequate assistance to stay negative. Few health workers understand the increased risk of HIV infection during pregnancy20 and therefore do not provide crucial prevention interventions. Creative strategies that include partner testing and address behavioral issues are needed to keep HIV-negative women from becoming infected during pregnancy and breastfeeding. HIV exposure is ongoing for many women, and the increased vulnerability of pregnant women to HIV needs to be urgently addressed in PMTCT, prevention, ART, and community-based programs.
The results of this study show benefit in reducing missed opportunities to reach women at maternity and taking advantage of this last chance to intervene before delivery. Widespread implementation of the simple inexpensive training program evaluated in this study can significantly increase identification of women with HIV infection and decrease the number of HIV-infected infants.
The successful implementation of the study would not have been possible without the dedication of EGPAF research assistants Edward Groening, Noluthando Bokwe, Mfanimpela Hlatshwako, and Dumisile Simelane. The authors particularly thank Rejoice Nkambule, Deputy Director for Public Health, for her leadership and support of this study; Allison Spensley and Jennifer Ballard Dwan for early work on the protocol; and Laura Guay who provided valuable review and comments for this article. M. P. Kieffer contributed to design, implementation, and analysis and was the primary writer. B. Nhlabatsi and C. M. Wilfert contributed to design and implementation. M. Mahdi and K. Kudiabor contributed to design, implementation, and analysis. H. J. Hoffman provided statistical analysis.
1. Volmink J, Siegfried NL, van der Merwe L, et al. Antiretrovirals for reducing the risk of mother-to-child transmission of HIV infection. Cochrane Database Syst Rev. 2007;24:CD003510.
2. Creek T, Tanuri A, Smith M, et al. Early diagnosis of human immunodeficiency virus in infants using polymerase chain reaction on dried blood spots in Botswana's national program for prevention of mother-to-child transmission. Pediatr Infect Dis J. 2008;27:22-26.
3. Kilewo C, Karlsson K, Ngarina M, et al; Mitra Plus Study Team. Prevention of mother-to-child transmission of HIV-1 through breastfeeding by treating mothers with triple antiretroviral therapy in Dar es Salaam, Tanzania: the Mitra Plus study. J Acquir Immune Defic Syndr. 2009;52:406-416.
4. Palombi L, Marazzi MC, Voetberg A, et al. Treatment acceleration program and the experience of the DREAM program in prevention of mother-to-child transmission of HIV. AIDS. 2007;21(suppl 4):S65-S71.
5. Nyandiko WM, Otieno-Nyunya B, Musick B, et al. Outcomes of HIV-exposed children in western Kenya: efficacy of prevention of mother to child transmission in a resource-constrained setting. J Acquir Immune Defic Syndr. 2010;54:42-50.
6. Sentinel Surveillance Survey 2007. Mbabane, Swaziland: Swaziland Ministry of Health; 2008.
7. Service Availability Mapping. Mbabane, Swaziland: Swaziland Ministry of Health; 2008.
8. UNAIDS. Report on the global HIV epidemic. UNAIDS/08.25E/JC1510E (English original, August 2008).
9. Coetzee D, Stringer EM, Chi BH, et al. Evaluation of PMTCT coverage in four African countries: the PEARL study. Presented at: 5th IAS Conference on HIV Pathogenesis, Treatment, and Prevention; July 19-22, 2009; Cape Town, South Africa. Abstract no. WELBD101.
10. Simonon A, Lepage P, Karita E, et al. An assessment of the timing of mother-to-child transmission of human immunodeficiency virus type 1 by means of polymerase chain reaction. J Acquir Immune Defic Syndr. 1994;7:952-957.
11. Gray RH, Li X, Kigozi G, et al. Increased risk of incident HIV during pregnancy in Rakai, Uganda: a prospective study. Lancet. 2005;366:1182-1188.
12. Leroy V, Van De Perre P, Lepage P, et al. Seroincidence of HIV-1 infection in African women of reproductive age: a prospective cohort study in Kigali, Rwanda, 1998-1992. AIDS. 1994;8:983-986.
13. Mbizvo MT, Kasule J, Mahomed K, et al. HIV-1 seroconversion incidence following pregnancy and delivery among women seronegative at recruitment in Harare, Zimbabwe. Cent Afr J Med. 2001;47:115-118.
14. Jackson JB, Parsons T, Musoke P, et al. Association of cord blood nevirapine concentration with reported timing of dose and HIV-1 transmission. AIDS. 2006;20:217-222.
15. Stringer JS, Sinkala M, Maclean CC, et al. Effectiveness of a city-wide program to prevent mother-to-child HIV transmission in Lusaka, Zambia. AIDS. 2005;19:1309-1315.
16. Antiretroviral Drugs for Treating Pregnant Women and Preventing HIV Infections in Infants: Recommendations for a Public Health Approach, 2010 Version. Geneva, Switzerland: World Health Organization; 2010.
17. Moodley D, Esterhuizen TM, Pather T, et al. High HIV incidence during pregnancy: compelling reason for repeat HIV testing. AIDS. 2009;23:1255-1259.
18. Kieffer MP, Mahdi MA, Nakato-Waligo A, et al. High levels of seroconversion among pregnant women in Swaziland. International AIDS Conference; 2008; Mexico City, Mexico. Abstract no. MOPE0508.
19. Humphrey J, Marinda E, Mutasa K, et al. Mother to child transmission of HIV among Zimbabwean women who seroconverted postnatally: prospective cohort study. BMJ. 2010;341:c6580doi:10.1136/bmj.c6580.
20. Zetolan NM, Pilcher CD. Diagnosis and management of acute HIV infection. Infect Dis Clin North Am. 2007;21:19-48.
© 2011 Lippincott Williams & Wilkins, Inc.