The Uptake of Prevention of Mother-to-Child HIV Transmission Programs in China: A Systematic Review and Meta-Analysis

Background No systematic review of prevention of mother to child transmission (PMTCT) in China has been performed. We aimed to estimate the uptake of PMTCT programs services in China. Methods We searched MEDLINE, EMBASE, Cochrane Central Register of Controlled Trials, China National Knowledge Infrastructure, Chinese Biomedical Literature Database, and Wanfang (Chinese) to identify research studies. Only descriptive epidemiological studies were eligible for this study. Results A total of 57 eligible cross-section studies were finally included. We estimated that the mean HIV-positive rate of exposed infants was 4.4% (95% CI = 3.2–5.5), and more than 33% of exposed infants had not undergone HIV diagnostic testing. The percentage of initiating antiretroviral therapy (ART) in HIV-positive women was 71.0% (95% CI = 66.3–75.8), and that for initiating antiretroviral prophylaxis (ARP) in exposed infants was 78.3% (95% CI = 74.9–81.8); also, 31.3% (95% CI = 15.5–47.0) of women with HIV and < 1% of exposed infants received the combination of three antiretroviral drugs. There were bigger gap of uptake of PMTCT programs between income levels, and cities with a low income level had a higher percentage of initiating ART in HIV-positive women (80%) and ARP in exposed infants (85%) compared to cities with high-middle income (57% and 65%, respectively) (P<0.05). Conclusions This paper highlights the need to further scale up PMTCT services in China, especially in regions with the lowest coverage, so that more women can access and utilize them. However, some estimated outcome should be interpreted with caution due to the high level of heterogeneity and the small number of studies.

Introduction same search strategies were used with each database.We placed no language restrictions on the searches or search results. Additional strategies included hand searches of journals that were not indexed in the electronic sources, web-based searches, and screening of reference lists of retrieved studies for additional potentially relevant articles.
Only descriptive epidemiological studies were eligible for this study. When a study reported the results from different subpopulations, we treated them independently. We excluded meetings, literature reviews, discussions, editorials, research overviews, book reviews, letters, and news articles. We excluded qualitative studies, modeling studies, studies where uptake of PMTCT was assessed by interview, and cost-effectiveness studies. We excluded studies that did not provide useable data. Studies with fewer than 30 participants were excluded to improve the efficiency of the analysis.
Two of the authors (MJ and HZ)independently screened the titles and abstracts of all identified studies. Studies that appeared to be relevant were selected, and the same two reviewers (MJ and HZ)independently assessed the full-text versions. Disagreements were resolved by consensus or the involvement of a third reviewer(ZH). Fig 1 shows the flowchart for selecting articles .

Data Collection
We developed and modified a data abstraction form after a training exercise for investigators. We extracted the following data from the eligible studies: characteristics of the studies (years, design, ART regimen, and regions), participants (age, sex, and the status of HIV infection), and primary and second outcomes. We also gathered data on potential explanatory variables (i.e., variables that might explain the variance in uptake of PMTCT; Table 1). We also obtained data from the website of the Chinese administrative region; these data were used to categorize regions into regions (EasternChina, south-central China, northChina, northwestChina, southwestern China, northeastChina, and special district of Taiwan, Hongkong, and Macao) and gross domestic product in $US per head in 2013 (low income (<5520), low-middle income (5635-6750), high-middle income (6892-9961), and high income (>10915)) [66,67].
For each study, one reviewer(ZD) extracted the data, a second reviewer (SZ)checked the accuracy, and a third reviewer (JH)evaluated the data for disagreements.

Primary Outcomes
The primary outcomes for pregnant women were the following: (1) HIV-positive ratein the present pregnancy (including antenatal, intrapartum, and postpartum); (2) underwent voluntary HIV-counselingin the present pregnancy (including antenatal, intrapartum, and postpartum); (3) underwent voluntary HIV-testingin the present pregnancy (including antenatal, intrapartum, and postpartum); (4) initiating antiretroviral therapy (ART) from 28 weeks of pregnancy (5) selecting termination of pregnancy; (6) elective cesarean section; and (7) artificialfeeding. The primary outcomes for children were the following: (8) HIV-positive rate (9) HIV diagnosis of the exposed infants between 12 and 18 months by PCR or antibody test; and (10) initiating ARP in exposed infantsin the first 72 hours after delivery.

Quality assessment
The quality of eligible literature was assessed according to the criteria of observational studies in recommended by Agency of Healthcare Research and Quality AHRQ included 11-items with a yes/no/unclear response option: the "Yes" would be scored "1", "No" or "unclear" was scored "0". Articles were scored as follows: low quality (0-3), moderate quality (4-7), high quality (8)(9)(10)(11) [68].

Statistical Analysis
We calculated the pooled rate or percentage with 95% CIs with a random effect model (DerSimonian-Laird's method) [69]. We used the Q-statistic and I 2 statistic to estimate the heterogeneity between studies, and we used "small," "moderate" and "large" to describe values of 25%, 50% and 75% for the I 2 [70][71].
We investigated potential sources of heterogeneity with subgroup analyses. In subgroup analyses, we estimated the uptake of PMTCT according to regions (EasternChina, southChina, South-central China, northChina, northwest China, southwestern China, northeastChina, and special district of Taiwan Hongkong and Macao), per capita GDP (low income, low-middle  income, high-middle income, and high income), pregnancy stage (premarital checkups, antenatal care, and at delivery), and antiretroviral therapy regimens (single regimen vs. combination regimens).
To establish the robustness of the outcome by sensitivity analyses, we applied a fixed effects model;used the trim-and-fill method; and excluded studies with a low number of participants [72].A funnel plot was used to explore the publication bias. Funnel-plot asymmetry was further assessed by the method of Begg' test and the modified Egger's linear regression test [73].We performed all analyses using the software STATA (version 11.0).

Characteristics of Eligible Studies
We identified 4459papers from a database search, 28 papers through internet and hand searches, and 15 papers through checking reference lists. No unpublished data that met our inclusion criteria were identified. During the step of screening the abstracts, 2884 papers were excluded, leaving 639full text papers that were assessed for eligibility. We excluded 40 papers with fewer than 30 participants; in the end, a total of 57 papers fulfilled our inclusion criteria and were included in the meta-analysis (Fig 1). These 57 studies originated from 31 of 34 provinces in China. Table 1 presents the characteristics of every analysis outcome, 13 studies were of high quality, and other studies were of moderate quality.

Estimated HIV-positive rate
Thirty-four studies, including 15,994,415 pregnant women, reported the HIV-positive rate of pregnant women (Fig 2). We estimated that the mean HIV-positive rate of pregnant women was 0.11% (95% CI = 0.1-0.12) with a high level of heterogeneity between the rate estimates HIV-testing; (4) initiating antiretroviral therapy (ART); (5) selecting termination of pregnancy; (6) elective cesarean section; and (7) artificialfeeding. The primary outcomes for children were the following: (8) HIV-positive rate (9) HIV diagnosis of the exposed infants between 12 and 18 months by PCR or antibody test; and (10) initiating ARP in exposed infants. (Q = 5049.43, P<0.001; I 2 = 99.30%). Sixteen studies, including 1,480 infants, reported the HIV-positive rate of exposed infants. We estimated that the mean HIV-positive rate of exposed infants was 4.4% (95% CI = 3.2-5.5) with a low level of heterogeneity between the rate estimates (Q = 13.96, P = 0.602; I 2 = 7.10%) (Fig 2 and Table 2).

Estimated Uptake of PMTCT Programs Services
The estimated percentages of voluntary HIV-counseling and HIV-testing of pregnant women were 79.3% (95% CI = 75.4-83.3) and 81.2% (95% CI = 77.8-84.6), respectively. The studies were very heterogeneous in both of the analyses (Fig 2). We identified 35 studies providing data for the percentage of initiating ART in HIV-positive women and 32 studies providing data for the percentage of initiating ARP in exposed infants. We estimated that the percentage of initiating ART in HIV-positive women was 71.0% (95% CI = 66.3-75.8) and initiating ARP in exposed infants was 78.3% (95% CI = 74.9-81.8); these estimates were also associated with a high level of heterogeneity (Fig 2).Other outcomes were also estimated, and the results are displayed in Fig 2. Subgroup Analyses of the Uptake of PMTCT Programs Table 3 and Table 4 Table 5). The region with the highest HIV-positive rate of pregnant women was the Xinjiang Uygur [Uighur]       Table 5).
The estimated HIV-positive rate of pregnant women was the highest in studies from low income regions(0.20%, 95%CI = 0.17-0.22), which was followed by low-middle income regions(0.15%, 95%CI = 0.07-0.29). Studies from high-middle income and high income regions reported the lowest prevalence( Table 5)

Sensitivity Analyses
We used the fixed effect model and trim and fill analysis, and we excluded studies with fewer participants to perform sensitivity analyses of the uptake of PMTCT programs, which gave similar results to the primary analysis.

Meta-regression analysis, assessment of publication bias
We noted significant heterogeneity within studies(P<0.001, I 2 = 95.1%-100%) except for the outcome of estimated HIV-positive rate of exposed infants (P = 0.602, I 2 = 7.1). In univariate  and multivariable meta-regression analyses (Table 6), we usedvariables including year of publication, sample size, regions (eastern China, south-central China, northwest China, southwestern China, and trans-regional), income level, and quality score. We notedthat regions, income level, and sample size were significantly associated with the estimated HIV-positive rate of pregnant women; year of publication was significantly associated with the percentage of voluntary HIV-testing (R 2 = 25.35%, P = 0.004); year of publication (R 2 = 30.63%, P = 0.056) and sample size (R 2 = 38.68%, P = 0.032) were significantly associated with the percentage of voluntary HIV-counseling; income level were significantly associated with the percentage of initiating ARP in exposed infants (R 2 = 12.36%, P = 0.043); year of publication was significantly associated with the percentage of selecting termination of pregnancy (R 2 = 17.59%, P = 0.007). Furthermore, in multivariable analysis these variables still were significantly associated with the heterogeneity of these main outcomes ( Table 6).
Egger'slinear regression test (P = 0.093) and Begg's test (P = 0.204) shown significantpublication bias among the contributing studies in terms of the outcome of the estimated HIV-positive rate of pregnant women, the percentage of initiating ART in HIV-positive women, the percentage of initiating ARP in exposed infants, and the percentage of artificial feeding (Table 7).

Discussion
This is the first comprehensive overview of PMTCT programs in China at the national level. We included 57 studies covering all provinces in China, and there was no report on PMTCT programs before 2004, which may be because the Chinese ministry of health first issued guidelines on PMCTC in 2004 [6].Our review showed that the overall uptake of PMTCT programs was low and did not reach the 80% target that was setby the United Nations General Assembly Special Session(UNGASS) [74]. The estimated percentage of antiretroviral therapy in HIV-positive pregnant women was still unsatisfactory. We estimated that the MTCT rate in China was 4.4% (95% CI, 3.2-5.5), which was substantially higher than in the U.S. and Europe (less than 1%), while it was lower than some low-and middle-income countries in Sub-Saharan Africa (11%) [1,75]. Great successes in reducing the MTCT rate has been achieved in China (reduced from 11.8% in 2005 to 4.2% in 2009) due to the scale up of PMTCT programs since the Ministry of Health launched PMTCT programs in 2003. In 2001 UNGASS set a goal that would reduce theproportion of HIV infected infants by 50% in ten years, and in order to achieve this target they estimated that 80% of pregnant women andtheir children need to receive PMTCT programs service [74]. Despite efforts to increase the uptake of PMTCT interventions services, coverage is still lower than desired in China. Many international non-governmental organizations, such as the UN Secretary General, G8 countries, the Global Fund toFight AIDS, WHO and so on have committed to further develop and improve the quality and effectiveness ofPMTCT service coverage in low-and middle-income countries [75]. Integration of PMTCT with other healthcare services, such as maternal, newborn, and child health may be a crucialcomponent of the strategy to scale up PMTCT programs. In China, PMTCT services were integrated with antenatal care and perinatal care, pregnant women were provided with HIV testing and counseling and HIV positivepregnantwomen were provided with antiretroviral prophylaxis in antenatal care and attending labor ward.Discouragingly, a recent review noted very limited, non-generalizableevidence of improved PMTCT intervention uptake in integratedPMTCT programscompared to non-or partially integrated services [76].A reviewassessing the role of family planning in eliminating new pediatric HIV infectionsreported that integrating family planning and HIV services is an effective strategy for increasing access to contraception among women with HIV who do not wish to become pregnant, which could   accelerate the ending of new pediatric HIV infections [77]. In recent years, PMTCT servicewas also integrated with family planning service in China; HIV testing was provided forwomen of childbearing age when attending national free pre-pregnancy eugenic health check project, and the HIV-positive women were suggested that they should receive antiretroviral therapy before pregnancy or prevent unintended pregnancies by the use of contraception. But the coverage of integrated family planning/HIV service was very low in China, which prompt the Chinese government to call for increaseddomestic and international financing to expand the uptake of PMTCT. We noted significant heterogeneity within studies in this meta-analysis, and meta-regression analyses revealed that year of publication was the main factors that explained much of the heterogeneity between studies.PMTCT programs have changed significantly over the years, and there are guidelines for starting ART for both women and infants. While this is acknowledged, treating all studies from 2000 to 2014 as equivalent leads to very heterogeneous data. For example, if the risk of MTCT was high in 2000, when presumably no ART was administered, it is unfair to combine these data with recent guidelines to provide ART if CD4<500 or lifelong ART (Option B+). Therefore, we performed subgroup analysis by years and present the historical and current policies for PMTCT services.Study region and income level may also account for much of the heterogeneity between studies. For example, northwest China had the highest MTCT rate (5.9%)and lowest uptake of PMTCT programs services; the high income region usually reported the highest percentages of voluntary HIV-counseling (93.2%) and voluntary HIV-testing (94.9%).The variation between regions and income levels were consistent with the global uptake of HIV testing; the low income countries are reported to have low uptake of HIV testing, and the coverage for early infant diagnosis of HIV was below 6% in some low income countries (Angola, Nigeria, Malawi, Democratic Republic of Congo, and Chad) in 2012 [78,79].Such variation could be explained by lower government financial investment, limited health resources and inefficient programs implementation strategy in regions with low uptake of PMTCT services.Voluntary counseling and testing (VCT) was effective in reducing MTCT as well as cost effective as a PMTCT intervention. The WHO had recommend that all children younger than 2 years old who are living with HIV should be treated, and the important first step was to identify the HIV-infected children through developing effective strategies for HIV testing. Although HIV testing facilities have increased over the past ten years, the uptake of VCT has still been low; UNAIDS reported in 2012 that 50% of individuals living with HIV were unaware of their HIV status [80]. Encouraging data are presented in this review, including that the estimated percentage of voluntary HIV-testing in pregnant women continually increased from 57.9% in 2004 to 98.5% in 2012 and that the estimated infant free HIV antibody testing rate over 18 months of age increased from 81% in 2005 to 97.7% in 2012. The number of new HIV-infected children in the 21 priority African countries in the UN Programme on HIV/AIDS (UNAIDS) global plan decreased by 38% between 2009 and 2012 because of increased access to antiretroviral treatment to prevent MTCT [81,82]. We estimated that more than 70% of HIV-infected pregnant women never received antiretroviral treatmentin antenatal care in China, which is significant lower than that in low-and middle-income countries with only 55% (range 22-99%) of HIV positive women starting highly active antiretroviral therapy in antenatal care [75,83]. An encouraging result in this meta-analysis was that initiating ARP in exposed infants continually increased from 77.0% in 2005 to 98.1% in 2012, which achieved the goal of the 12thFive-Year action plan on containment, prevention and control of HIV/AIDS that the percentage of HIV exposed infants who received ARV would be more than 90% by the year 2015 [84]. According to recently WHO report, the use of combination ART during pregnancy is preferable to single therapy [85]. The combination ART is more effective at PMTCT, and it has the advantages of reducing sexual HIV transmission and HIV-associated morbidity and mortality [86]. In China, we estimated that approximately 31.3% (95% CI, 15.5-47.0) of women with HIV received the combination of three antiretroviral drugs (AZT+3TC/ AZT+NVP), and < 1% of exposed infants received the combination of three antiretroviral drugs.
Despite the great successes in reducing MTCT in China, we are facing many challenges, such as low coverage of PMTCT programs, stigma and discrimination, drug resistance, and delayed infant HIV diagnosis [5]. This review had several limitations; firstly, significant heterogeneity between studies was observed. Although we performed subgroup analyses by publication year, geographical area, and income level, and these factors may be the sources of between-study heterogeneity. However other unmeasured characteristics in study population and limitations of the included studies likely influence the detected heterogeneity; unfortunately, we did not obtain enough information about these aspects for further analysis. Secondly, we have only conducted the search in electronic databases. Studies published in local journals which are not indexed in electronic databases might have been missed out in this review. The third, results from Begg's funnel plot and Egger's test are different but the funnel plot and Trim and Fill methods suggested the presence of a potential publication bias, a language bias, and inflated estimates by a flawed methodological design in smaller studies.The last, some pooled estimates in subgroups should be interpreted with caution due the small number studies.
In conclusion, PMTCT programs have scaled up quickly in recent years in China. However, antiretroviral therapy in HIV-positive pregnant, antiretroviral prevention and HIV diagnosis in exposed infants were still unsatisfactory. Moreover, there was a big gap of uptake of PMTCT programs between regions and income levels. These results highlight the need to further scale up PMTCT services, especially in regions with the lowest coverage, so that more women can access and utilize them.