Introduction of Rapid Syphilis Testing Within Prevention of Mother-to-Child Transmission of HIV Programs in Uganda and Zambia

A Field Acceptability and Feasibility Study

Abstract

Background: 

Given that integration of syphilis testing into prevention of mother-to-child transmission of HIV (PMTCT) programs can prevent adverse pregnancy outcomes, this study assessed feasibility and acceptability of introducing rapid syphilis testing (RST) into PMTCT services.

Methods: 

RST was introduced into PMTCT programs in Zambia and Uganda. Using a pre–post intervention design, HIV and syphilis testing and treatment rates during the intervention were compared with baseline.

Results: 

In Zambia, comparing baseline and intervention, 12,761 of 15,967 (79.9%) and 11,460 of 11,985 (95.6%) first-time antenatal care (ANC) attendees were tested for syphilis (P < 0.0001), 523 of 12,761 (4.1%) and 1050 of 11,460 (9.2%) women tested positive (P < 0.0001); and 267 of 523 (51.1%) and 1000 of 1050 (95.2%) syphilis-positive women were treated (P < 0.0001), respectively. Comparing baseline and intervention, 7479 of 7830 (95.5%) and 11,151 of 11,409 (97.7%) of ANC attendees were tested for HIV (P < 0.0001) and 1303 of 1326 (98.3%) and 2036 of 2034 (100.1%) of those testing positive received combination antiretroviral drugs or single-dose nevirapine prophylaxis (P < 0.0001). In Uganda, 13,131 of 14,540 (90.3%) women were tested for syphilis during intervention, with 690 of 13,131 (5.3%) positive and 715 of 690 (103.6%) treated. Syphilis baseline data were collected, but not included in analysis, as ANC syphilis testing before the study was not consistently practiced. Comparing baseline and intervention, 6479 of 6776 (95.6%) and 11,192 of 11,610 (96.4%) ANC attendees were tested for HIV (P = 0.0009) and 570 of 726 (78.5%) and 964 of 1153 (83.6%) received combination or single-dose prophylaxis (P = 0.007). In Zambia, 254 of 1050 (24.2%) syphilis-positive pregnant women were HIV-positive and 99 of 690 (14.3%) in Uganda.

Conclusions: 

Integrating RST in PMTCT programs increases screening and treatment for syphilis among HIV-positive pregnant women and does not compromise HIV services.

INTRODUCTION

Integration of maternal and child health (MCH) and HIV services promotes efficient and effective service delivery and can reduce the risk of HIV mother-to-child transmission (MTCT), while improving pregnancy outcomes.¹ A 2010 study of antiretroviral therapy in MCH showed significant increases in enrollment and treatment initiation compared with sites where referral for antiretroviral therapy initiation was required.²

HIV and syphilis coinfection is common. A review of global data estimated on average 9.5% of HIV-positive adults are coinfected with syphilis.³ Pregnant coinfected women are twice as likely to transmit HIV in utero compared with those infected with HIV alone.^4–6 Assuring HIV-positive women are screened and treated for syphilis will reduce risk of MTCT of HIV and prevent adverse outcomes of untreated syphilis. Recognizing this benefit and in pursuit of joint elimination, the World Health Organization is supporting prevention of mother-to-child transmission of HIV (PMTCT) and syphilis integrated programs.

High rates of MTCT and congenital syphilis indicate missed opportunities for disease prevention and co-manage-ment. Globally, only 53% (40%–79%) of HIV-positive pregnant women receive antiretroviral drugs (ARVs) to prevent MTCT.⁷ Although effective inexpensive treatment for syphilis exists, a single dose of benzathine penicillin given before 28 weeks gestation,^8,9 congenital syphilis remains a major cause of adverse pregnancy outcomes.^10–12 Screening and management of syphilis and HIV have been poorly coordinated, and syphilis testing required laboratory processing and revisits for results and treatment.

In Zambia and Uganda, PMTCT programs in antenatal care (ANC) settings, including a range of MCH services, have scaled-up successfully; more than 530,000 and 960,000 women, respectively, received PMTCT services in 2009.¹³ PMTCT sites are important venues for quality-assured point-of-care (POC) tests, such as low-cost rapid syphilis tests (RST) that eliminate referrals and return visits.^14,15 RSTs are treponemal-specific and use serum, plasma, or whole blood, but do not distinguish current versus past infections.¹⁶ Four RSTs evaluated in resource-limited settings had specificities over 95% and sensitivities from 64% to 100%.¹² HIV testing is already conducted using a rapid POC diagnostic in these settings.

The Elizabeth Glaser Pediatric AIDS Foundation, partnering with the Centre for Infectious Disease Research in Zambia, collaborated with the Ministries of Health (MOH) of Uganda and Zambia to study feasibility and acceptability of introducing RSTs into PMTCT services, with a focus on the impact on existing HIV services. For the purposes of this study, feasibility was defined as use of RST to improve identification and treatment of syphilis in pregnant women without compromising HIV services and ability to conduct multiple rapid tests concurrently, and acceptability was defined as health workers' satisfaction with and correct and consistent use of RST.

METHODS

Study Population

The study population included all pregnant women attending their first ANC visit and their male partners. The study was conducted in 2 countries where syphilis screening was specified in the national ANC guidelines but inconsistently available among pregnant women. Sites with established PMTCT services were collaboratively selected with the MOH from each country as follows: 9 rural sites (Mongu District) and 6 urban sites (Lusaka District) in Zambia; and a high-volume urban site Mulago National Referral Hospital (MNRH) and 8 rural facilities (Jinja District) in Uganda.

Study Design

We used a pre–post intervention design with retrospective review of routine data in Zambia (April to September 2009) and baseline data collection in both countries (November 2009 to January 2010) before RST introduction. Aggregate syphilis and HIV data were extracted from modified ANC registers during the intervention (February to June 2010 in Uganda and March to July 2010 in Zambia).

Intervention Procedures

All pregnant women attending ANC at study clinics received HIV and syphilis services per national guidelines using a same-day testing and treatment (STAT) strategy. After baseline data collection and health worker training, RSTs were introduced into ANC/PMTCT services at study sites. Rapid syphilis and HIV testing following standard operating procedures were performed concurrently using venipuncture or finger-prick whole blood.^17,18

After group pre-test counseling, HIV and syphilis rapid tests were conducted. Women received individual post-test counseling, and treatment if needed.^19,20 ARV prophylaxis provided to women for PMTCT of HIV followed the current country guidelines of antiretorviral therapy for one's own health or combination prophylaxis (nevirapine and zidovudine). In Uganda, additional consent was obtained for treatment based on RST results. Partner notifications were given to syphilis-positive women in Zambia to encourage partner testing. In Uganda, general invitation letters were sent home with ANC attendees inviting partners to discuss syphilis, HIV, and other health issues. Following counseling, partners were offered rapid HIV and syphilis testing. Syphilis-positive partners received same-day treatment and HIV-positive partners were referred to care. Presumptive syphilis treatment was available for exposed male partners testing negative.

Data Collection

In Zambia, baseline data were used to compare HIV service statistics; however, rapid plasma reagin (RPR) test stock-outs necessitated retrospective baseline data collection to compare syphilis care and treatment services. Data collected from ANC registers during the intervention included number of ANC attendees, number of women and partners tested for HIV and syphilis, results, number receiving results, syphilis treatment rates, and referrals for HIV care. Facility assessments documented service and patient volumes, staffing levels, location, duration of PMTCT services, and stock-outs of benzathine penicillin and rapid test kits.

Syphilis Testing and Quality Management

RST was performed using SD BIOLINE Syphilis 3.0 rapid syphilis antibody test (Standard Diagnostics, Inc., Yongin-Si, South Korea).²¹ The manufacturer reported 99.3% sensitivity and 99.5% specificity using serum. In a clinic-based evaluation using whole blood, SD BIOLINE sensitivity was 85.7% and specificity was 98.1%.¹² Test kits were transported, stored, and performed on whole blood following manufacturer instructions and study-specific standard operating procedures. Study-specific RST external quality assurance and quality control management systems were established.²² Internal quality control included weekly testing using archived syphilis seropositive and seronegative sera aliquots. External quality assurance included repeat confirmatory testing of 2% of samples using SD BIOLINE at a central laboratory. Health workers completed proficiency testing using dried tube specimens. Investigators reviewed discrepant results with clinics; health workers who scored below 100% underwent remedial training on RST use and interpretation.

Data Analysis

Quantitative data were entered into a Microsoft Access database. All statistical analyses were generated using SAS/STAT software, version 9.1 of the SAS System for Windows (SAS Institute Inc, Cary, NC). Contingency tables and χ² tests were used to summarize data and test for associations between event frequencies. Pearson χ² test statistics were calculated and reported for 2-way tables. Two-way test statistics were used to test the null hypothesis of no association between row and column variables. If more than 20% of expected cell counts were less than 5, Fisher exact test was applied.

Ethical Review

The Uganda study protocol was approved by Uganda's National Council for Science and Technology. The Zambia protocol was approved by the University of Zambia Biomedical Research Ethics Committee and University of Alabama at Birmingham's Institutional Review Board. Both protocols were approved by the World Health Organization Research Ethics Review Committee and by each country's MOH.

RESULTS

Feasibility of Integrating RST into PMTCT Services

During baseline, only 1 of 9 study sites in Uganda had performed any syphilis testing in the past year, testing symptomatic women only. Baseline data were collected from this site but are not used in analysis (Table 1). During intervention, 13,131 of 14,540 (90.3%) women attending ANC for the first time were tested and 690 of 13,131 (5.3%) were syphilis-positive; 715 of 690 (103.6%) women were treated and 708 of 715 (99.0%) of those treated received STAT. Women who tested syphilis-negative but whose partner tested positive were treated, resulting in over 100% of positive cases in women treated (690 positive cases vs. 715 treated).

TABLE 1: Pre-intervention and postintervention Results for Syphilis Testing and Treatment in Uganda

TABLE 2: Pre-intervention and Postintervention Results for Syphilis Testing and Treatment in Zambia

Due to RPR stock-outs, which occurred immediately before the intervention in Zambia, retrospective data were used for baseline as follows: 12,761 of 15,967 (79.9%) first-time ANC attendees received RST, 523 of 12,761 (4.1%) tested positive, and 267 of 523 (51.1%) of the positive cases were treated (Table 2). During intervention, 11,460 of 11,985 (95.6%; P < 0.0001) first-time ANC attendees received RST, 1050 of 11,460 (9.2%) tested positive, and 1000 of 1050 (95.2%) of positive cases were treated. Of those, 958 of 1000 (95.8%) received STAT.

Impact on Existing HIV Services

One Jinja District site in Uganda experienced significant ARV stock-outs and was excluded from HIV service pre- and post analyses. Significant increases in the percentages of pregnant women tested for HIV (11,192; 96.4%; P = 0.009) and receiving ARV prophylaxis (964; 83.6%; P = 0.007) were noted after RST introduction; no significant changes in HIV service referrals were noted (118; 16.1%; P = 0.74) (Table 3).

TABLE 3: Pre-intervention and postintervention Results for HIV Services in Uganda

TABLE 4: Pre-intervention and Postintervention Results for HIV Services in Zambia

During the first 5 months of implementation, RST was provided in ANC in Zambia, there was a significant increase in percentage of women tested for HIV at first ANC visit (11,151; 97.7%; P < 0.0001) (Table 4). Of those tested, 2034 (18.2%) tested positive, and there were significant increases in the percentage of women receiving maternal ARV prophylaxis (2036; 100.1%; P < 0.0001) and referrals to care and treatment (1721, 84.6%; P < 0.0001). A total of 576 known HIV-positive women attended ANC. The percentage of women receiving maternal prophylaxis and referrals were inflated, as illustrated by percentages over 100%, because of the inclusion of some women with known HIV status in the numerator but not in the denominator. Due to the aggregate nature of the data collected at the facilities, it is not possible to identify those individuals and exclude them from the analysis.

Male Involvement

In Uganda, a general invitation letter was distributed for partners of ANC attendees. In ANC, 1436 of 14,450 (9.9%) partners received syphilis testing; 118 of 1436 (8.2%) tested positive and 153 were treated (117 testing positive for syphilis plus 36 treated presumptively due to partner's positive RST). There was a small but significant increase in the percentage of partners receiving HIV testing (9.8% to 12.5%; P < 0.0001).

In Zambia, a targeted invitation letter for partners of syphilis-positive women was distributed. Of 1050 positive women, 363 male (34.6%) partners were tested for syphilis; 179 of 363 (49.3%) tested positive, and 160 of 179 (89.4%) were treated. There was no significant change from baseline in percentage of partners tested for HIV (3.9% to 4.0%; P = 0.73).

Coinfection With HIV and Syphilis

During the Uganda intervention, 99 of 14,540 (0.7%) first-time ANC attendees tested positive for syphilis and HIV. Among syphilis-positive pregnant women, 14.3% were HIV-positive; conversely, 7.5% of HIV-positive pregnant women were syphilis-positive. In Zambia, 254 of 11,985 (2.1%) first-time ANC attendees tested positive for syphilis and HIV. Among syphilis-positive pregnant women, 24.2% were HIV-positive; conversely, 9.7% of HIV-positive pregnant women were syphilis-positive. For both countries, coinfection with HIV and syphilis varied by urban and rural districts (Fig. 1).

FIGURE 1: Magnitude of HIV and syphilis comorbidity among ANC clients in Jinja and Kampala Districts (Uganda) and Mongu and Lusaka Districts (Zambia). Each colored bar represents a distinct cohort of women. The blue bar represents the percentage of syphilis-positive pregnant women who also had HIV. The red bar represents the percentage of HIV-positive pregnant women who also had syphilis; and the green bar represents the percentage of pregnant women who were HIV and syphilis coinfected at first visit.

DISCUSSION

This study demonstrated the feasibility and acceptability of RST within busy urban and rural sites in 2 countries. Significant increases in syphilis testing and treatment using a POC test and improvements in measures of HIV testing and receipt of ARV prophylaxis were demonstrated. In Uganda, where access to RPR is extremely limited, all levels of health facilities were able to add RST with high STAT coverage. In Zambia, RST was introduced in high-volume PMTCT sites in both the capital and remote areas. Health workers incorporated RST into routine PMTCT services in settings with high HIV disease burdens with no significant negative impacts on HIV service delivery. Documentation of both positive change and lack of negative impact is an important measure of acceptability and feasibility.

To sustain improved HIV services, it is critical to understand any potential impact when programmatic changes are introduced into settings with high HIV disease burdens. Increases in HIV service delivery during intervention are likely attributable to additional health worker training and improved logistics systems. To the authors' knowledge, there were no notable changes in service delivery, health policy or practice, or other external conditions in the countries that would have influenced HIV service delivery. This also did not seem to be part of an emerging trend though a potential Hawthorne effect may have been introduced. It should be noted that our objective was to monitor any significant decreases in HIV services to ensure they did not deteriorate with the addition of RST; such deterioration was not found.

Male involvement is seen as an important component of comprehensive PMTCT services. Through use of letters, male partners came to ANC for HIV and syphilis testing, and syphilis treatment. Limiting partner letter distribution to only syphilis-positive women in Zambia resulted in fewer men coming for testing than in Uganda where the general letter was used. A small but significant increase in male partner HIV testing was observed in Uganda.

The SD Bioline rapid test is Treponema pallidum-specific, producing a positive result regardless of successful past treatment; therefore, use of this test could result in overtreatment. In Zambia, the baseline syphilis rate was 4.1% compared with 9.2% with RST, with the assumption that RST is providing positive results for past infections. Given the serious consequences of untreated syphilis in pregnancy, the benefit of STAT for syphilis outweighs the risk of overtreatment. The common diagnostic for syphilis, RPR, detects antibodies that are not specific to Treponema pallidum and can therefore result in false positives, also resulting in overtreatment.²³A high-quality treponemal/nontreponemal rapid test would address these drawbacks.

This study showed high rates of syphilis and HIV coinfection in pregnant women, with 24.2% and 14.3% of syphilis-positive women also HIV-positive in Zambia and Uganda, respectively. The observed high coinfection rates in this study and those reported elsewhere³ and the increased risk of MTCT of HIV among women with syphilis coinfection^4–6 justify the inclusion of syphilis STAT in PMTCT programs.

This operations research study was strengthened by a mixed methods approach²⁴ and implementation within “real world” resource-limited settings but had some limitations, such as a reliance on routinely collected data. Cross-sectional data and collection in busy clinics may have affected accuracy and completeness of information at various service points, although data verification by a study monitor reconciled inconsistencies. Data collection did not link syphilis and HIV-related information between women and partners. Use of facility records also tends to result in data inflation; notably, in percentages greater than 100% when known HIV-positive women received prophylaxis without testing or where presumptive syphilis treatment was given based on a partner's positive test. However, we assume low variation in inflated data between baseline and intervention periods, allowing for valid comparisons. Another limitation was the possibility that increases in HIV service delivery may have been attributable at least in part to an unknown factor or an emerging trend—though this does not seem likely based on the authors' experience and knowledge of the countries. This is supported by the fact that the study was conducted in different countries and in different settings within each country (eg, urban versus rural), yet yielding similar results. Potentially wider health systems strengthening that were included with the introduction and integration of RST, such as health worker training and quality management procedures may have led to the improved HIV indicators. It is difficult to know about the sustained effects of these health systems strengthening measures on HIV services over time, but the authors would advocate that these activities be included to support integration.

This study provides compelling evidence that RST can be acceptably and feasibly incorporated into urban and rural PMTCT programs. The results reinforce findings of other studies that successfully introduced RST in ANC^25,26 and also demonstrate feasibility and acceptability in HIV services. Additional research is recommended to measure incidence of congenital syphilis, HIV, and other related outcomes (eg, stillbirths) in infants. Introducing RST alongside HIV rapid testing facilitates women's receipt of STAT and reduces risk of transmitting both HIV and syphilis to their unborn infants. Combining POC HIV and syphilis testing in ANC is a tangible aspect of integration of HIV and MCH.

ACKNOWLEDGMENTS

The authors would like to acknowledge the support provided by the Uganda MOH (Dr Jenifer Wanyana, Dr Zainab Akol, Dr Godfrey Esiru, Dr Collins Tusingwire, Dr Benon Biryahwaho, and Mr John Kalema); Mulago Hospital staff and the Jinja District Health Team, Zambia MOH (Dr Reuben Mbewe, Dr Mary Nambao, and Dr Maximillian Bweupe); Mongu and Lusaka District Health Teams; and Mr Ignatius Kunda and the Kalingalinga Laboratory for their role in the study of integrating RST into PMTCT services. The authors also want to thank the following individuals for their invaluable technical support for the study: Catherine Wilfert, Rosanna Peeling, David Mabey, Laura Guay, Chuck Hoblitzelle, Blair Kauffmann, Helen Kelly, Kimberly Sollis, Manya Magnus, John Changalucha and the trainers at the National Institute of Medical Research Tanzania and Sandra Garcia and the Population Council-Mexico Team. The authors also thank Luis Cuevas and Izabela Suder from the UNICEF/UNDP/World Bank/World Health Organization Special Programme for Research and Training in Tropical Diseases for their ongoing support throughout this study.