Expanded Services for the Prevention of Mother-to-Child HIV Transmission

To the Editor: At the start of 2007, an estimated 2.3 million children in sub-Saharan Africa were infected with HIV, most of whom had acquired the virus around the time of birth.1 As services in the region move toward more complicated and efficacious drug regimens for prevention of mother-to-child transmission of HIV (PMTCT),2 acceptability is a critical component to overall program effectiveness. Several studies have observed poor follow-up among patients concerned about stigma or disclosure of their HIV status.3,4 The risk of viral drug resistance after use of drugs such as nevirapine (NVP) may only add to a patient's reluctance to adhere to specific PMTCT drug regimens. We evaluated a pilot program designed specifically to link PMTCT services, provided routinely as part of antenatal care, to long-term HIV care and treatment in Lusaka, Zambia. Kanyama Health Center is among the largest of the Lusaka District's 24 primary care centers and the site of robust government programs for PMTCT5 and antiretroviral therapy (ART).6 To link PMTCT to HIV care and treatment, we implemented specific referral procedures. This began with a scripted talk on the benefits of long-term HIV care and the importance of CD4+ cell count screening. All women were encouraged to enroll into long-term HIV care and treatment. Women interested in enrollment that day were escorted by community health workers to the on-site HIV care and treatment facility and seen immediately. Given the heavy clinic burden, we used dedicated staff members, 2 nurses and 1 clinical officer, to expedite the enrollment of pregnant women into long-term HIV care. Initial evaluation consisted of a medical history, physical examination (including World Health Organization [WHO] staging), and CD4+ cell count screening. A follow-up appointment was made for 2 to 3 weeks later to review laboratory results and determine the need for ART. ART eligibility was based on Zambian National Guidelines:7 CD4+ count <200 cells/mm3 (regardless of WHO stage), WHO stage 4 (regardless of CD4+ cell count), and WHO stage 3 with CD4+ count <350 cells/mm3. Pregnant women who did not qualify for ART were offered 1 of 3 drug regimens for PMTCT: single-dose NVP, short-course zidovudine (ZDV), or a combination of the two (ZDV + NVP). Patients actively participated in decision making. As part of a standardized counseling script, advantages of each regimen were reviewed: higher efficacy (ZDV + NVP), discreetness (NVP alone), and minimal risk for HIV resistance (ZDV alone). Illustrations and patient materials were used to help convey important points. Data were collected from March 1, 2005, the launch of the pilot program, to October 15, 2005, when a new initiative to link ART and PMTCT services was introduced in the health center. This analysis was deemed exempt from human subjects review by the Institutional Review Boards of the University of Zambia and the University of Alabama at Birmingham. Over the course of the pilot program, 680 pregnant women were newly diagnosed with HIV and 433 (64%) were referred from antenatal care by district staff. Patient flow was remarkable for a high degree of attrition (Fig. 1). Among those referred, only 302 (70%) came for enrollment. Of these, 96 (32%) failed to return after the initial visit; 77 had complete ART eligibility information, whereas 19 had missing CD4+ cell counts. Women lost to follow-up after only 1 visit were, on average, later in pregnancy (28.2 vs. 25.9 weeks; P = 0.008) compared with those who subsequently returned. There were no differences with regard to age, gravidity, parity, CD4+ cell count, WHO stage, or baseline hemoglobin between the 2 groups (P > 0.05; data not shown), however.FIGURE 1: Patient flow from antenatal services to Kanyama Health Center pilot program, March 1 to October 15, 2005. ANC indicates antenatal care clinic.Of the 283 women with complete enrollment data, 72 (25%) met criteria for ART initiation and 211 (75%) did not. When compared with those who did not qualify for ART, women who met criteria for ART were, on average, older (28.0 vs. 25.8 years; P = 0.006) and earlier in pregnancy at time of screening (24.8 vs. 27.0 weeks of gestation; P = 0.03). Women who qualified for ART were also less likely to return for subsequent clinic visits (54% vs. 79%; P < 0.001). Only 39 (54%) of 72 of those qualifying for ART returned for review; 33 started therapy and 6 declined. Median gestational age for initiating ART was 28.3 weeks (interquartile range [IQR]: 24.9-35.0 weeks), with a median time of 6.6 weeks (IQR: 2.6-9.1 weeks) from screening to treatment initiation. Of the 211 pregnant women not yet eligible for ART, 155 (73%) returned for their eligibility visit. These individuals were offered counseling on additional PMTCT options, and 140 agreed to take part. Most (99% [n = 138]) chose ZDV + NVP. Only 2 women opted for NVP alone, and none chose ZDV alone (see Fig. 1). Among those choosing ZDV + NVP, the median gestation age at time of initiation was 31.0 weeks (IQR: 29-34 weeks). Based on previous work,8 we originally hypothesized that ZDV + NVP may not be highly acceptable in our setting, because of its high pill burden and its association with nonnucleoside reverse transcriptase inhibitor resistance.9 We were thus surprised to find that after a scripted counseling session presenting risks and benefits of the 3 regimen options, nearly all women (138 of 140) chose the most efficacious regimen available, even when moderately efficacious alternatives that were more discrete (NVP alone) or that provided much less risk of viral drug resistance (ZDV alone) were also available. These findings suggest that protecting their infant from perinatal HIV was the highest priority among these women. Although the external validity is a concern (women adherent to the intensive schedule of visits may be more likely to choose more efficacious PMTCT regimens), we believe that these findings nonetheless provide important insight into the acceptability of PMTCT regimens outside of single-dose NVP in our setting. Like other studies in the region,10,11 we observed a high level of attrition among patients as they navigated the cascade of clinical visits. Because of the drop-out rate, only 171 (39%) of the 433 patients started a PMTCT regimen outside of single-dose NVP (the default PMTCT regimen provided before referral). Most noteworthy and perhaps most concerning was the poor follow-up between the initial screening visit and ART eligibility visit, where nearly one third of HIV-infected women failed to return. The fact that 43% (33 of 77) of women who did not return for their eligibility visit actually qualified for ART represents a critical missed opportunity for long-term maternal health and PMTCT. Because of funding limitations, we did not perform contact tracing for patients with missed visits; as a result, we were unable to determine the causes behind this high drop-out rate. Because we used programmatic data for this analysis, much of it based on simplified clinical registers and forms, we lacked detailed patient information outside of basic health indicators. This limited our ability to describe our patient population or to generalize our findings for different sites. Nevertheless, we believe that our analysis population was representative of many urban settings in Africa. The catchment area of the Kanyama Health Center is one of the most densely populated areas in Lusaka, with a high level of poverty and mobility among its inhabitants. Nearly all women who attend the clinic live in neighboring communities; few travel in from outside the city. The high patient volumes seen daily at the health center are indicative of the limited options for medical care in the area. Whether our findings can be applied to periurban or rural sites is unknown but is of great interest as we implement similar services outside the capital city. We have used these experiences to inform a wider district initiative to link PMTCT and ART services. To reduce the notable lag time between ART screening and ART initiation, for example, we have devised strategies to integrate these services further into the PMTCT/antenatal care infrastructure. “Reflex” CD4+ cell testing (flow cytometry on all specimens testing positive for HIV antibodies) has been instituted for women as part of our HIV testing algorithm. We have trained nurses to screen HIV-infected patients for medical histories suggestive of advanced HIV disease. In this manner, staff in the antenatal clinic can reliably triage referrals to the HIV treatment clinic and ensure the speedy initiation of ART among those who need it most. We also provide ZDV + NVP as the standard of care for PMTCT in Lusaka (in accordance with new government guidelines12) because of its high efficacy and demonstrated acceptability in our setting. In summary, ZDV + NVP was proven to be highly acceptable in this population, suggesting that prevention of perinatal HIV transmission may be a more important factor in antiretroviral drug regimen choice than discreetness of the intervention or concerns of HIV resistance. Although the linking of PMTCT and ART services is important to the long-term health of HIV-infected mothers and their infants, programs may face a high level of attrition from visit to visit. Strategies that better integrate the 2 services are needed and should be critically examined, particularly in settings of severe resource constraint. ACKNOWLEDGMENTS The authors are supported by the US National Institutes of Health (K01 TW06670-01A1, K01 TW05708-01, and K23 AI52481). They thank Felistas Mbewe, Febby Kawamya, Inonge Hatimbula, Jenny Cohen, and the clinical staff at Kanyama Health Center for their hard work and dedication to the expanded PMTCT pilot program. Benjamin H. Chi, MD*†‡ Namwinga Chintu, MD, MMed*‡ Alison Lee, BS* Elizabeth M. Stringer, MD*†‡ Moses Sinkala, MBChB, MPH*§ Jeffrey S. A.