Evaluating the progress to eliminate mother-to-child transmission (MTCT) of syphilis in Hunan Province, China: A study based on a health service delivery model

To prevent mother-to-child transmission (MTCT) of syphilis, Hunan Province launched a free syphilis screening and treatment programme in 2011. Thus far, the programme has been implemented for 6 years. This study aimed to assess progress toward the elimination of MTCT of syphilis in Hunan Province from 2011–2016. Estimates of syphilis-related adverse pregnancy outcomes (APOs) were based on the health service delivery model developed by the WHO, which were then translated into disability-adjusted life years (DALYs). Default values in the model were replaced by a Chinese version. The progress of this programme was assessed through the reduction of estimated DALYs with and without screening and treatment services. The results showed that the estimated number of syphilis-related APOs in Hunan Province from 2011 to 2016 was 3,840, more than 70% of which occurred among women who had at least one antenatal care visit but were not screened or treated for syphilis during pregnancy. The public health burden resulting from maternal syphilis-related APOs was 192,528 DALYs over six years, and with the current screening and treatment coverage, approximately 163,794 expected DALYs (46%) were averted. Our estimates indicate that in Hunan Province, syphilis in pregnancy continues to be an important cause of APOs, which can lead to substantial perinatal morbidity and mortality. Approximately half of the expected public health burden resulting from syphilis-related APOs was averted by the current screening and treatment services, which suggests progress toward the elimination of MTCT of syphilis in Hunan Province.


Introduction
In this study, we estimated the disease burden resulting from syphilis-related APOs in Hunan Province from 2011 to 2016 and assessed the progress made in the prevention programme.

Materials and methods
The primary objective of this analysis was to assess the progress made in the programme of preventing MTCT of syphilis in Hunan Province, which was achieved through the estimation of three values: the number of syphilis-related APOs, the estimated number of APOs in the form of DALYs, and the reduction in DALYs achieved by screening and treatment services. The secondary objective was to calculate the number of DALYs averted in the optimal situation in which the coverage of syphilis screening and treatment services has met the WHO-proposed targets for eliminating MTCT of syphilis (i.e., coverage of syphilis testing of pregnant women ! 95% and treatment of syphilis-seropositive pregnant women ! 95%).

Part 1 Estimate the number of syphilis-related APOs in Hunan Province
Estimation model. In this analysis, we estimated the number of syphilis-related APOs via the health service delivery model, which was developed by the WHO and reviewed and approved by the Child Health Epidemiology Reference Group. This estimation included calculations of the number of pregnant women with syphilis infection and the number of syphilisrelated APOs. Details of the model have been described elsewhere [12]. In this model, syphilis infection was defined as seropositivity on both the Treponema pallidum (TP) and non-TP tests. The number of pregnant women with syphilis infection was calculated as the product of the prevalence of syphilis in pregnant women and number of pregnant women. The number of syphilis-related APOs was calculated as the product of the number of pregnant women with syphilis infection and the probability of those women having syphilis-related APOs (the difference between the probability of these women having APOs and the probability of the general population of women having the same APOs).
Estimation process and data sources. Based on the health service delivery model, this estimation involved two steps (Fig 1): calculation of the number of pregnant women with syphilis infection in Hunan Province and calculation of the number of syphilis-related APOs in Hunan Province. For Step A, we estimated the number of pregnant women in Hunan Province with syphilis infection with no antenatal care (ANC) and with at least one ANC visit.
Data on the syphilis seroprevalence among ANC attendees (P ss , Step A 1 in Fig 1) in Hunan Province were obtained through China's Information System of Prevention of Mother-to-Child Transmission of Syphilis Management, a nationwide health facility-based case report system used to monitor and evaluate the prevalence of syphilis infection in mother and their offspring in China. Maternal syphilis is monitored through mandatory case reporting by medical and health institutions in 31 provinces, municipalities, and autonomous regions in China, including general hospitals, maternal and children's hospitals and other health agencies across the country. Diagnosis of maternal syphilis was based on the standard of health industry in the People's Republic of China (WS 273-2007) [13]. All women registered in the system were seropositive in both TP and non-TP tests, regardless of specific test methods. We assumed that the percentage of women with at least one ANC visit (P preg_ANC , Step A 3a in Fig 1) was equal to the antenatal examination rate, which was available from the Hunan Province's Annual Report of Maternal and Child Health. Thus, the number of pregnant women (N preg , Step A 2 in Fig 1) were estimated as the number of women attending their first ANC services divided by the antenatal examination rate. For Step B, we estimated the number of syphilis-related APOs among syphilis-infected women who had no ANC visits.
We also estimated the number of syphilis-related APOs among syphilis-infected women who had at least one ANC visit and were not tested or not treated (if seropositive), and were both tested and treated.
APOs included spontaneous abortion, stillbirth, preterm birth/low birth weight, neonatal death and congenital syphilis. The proportion of pregnant women who were both tested and Evaluating the progress to eliminate MTCT of syphilis treated for syphilis (P tt ) was calculated as the product of the syphilis screening rate and the syphilis treatment rate which were available from the Hunan Province's Annual Report of Maternal and Child Health. Since there was no nationwide or representative data available in China, in this study, we conducted a meta-analysis to estimate the summary probabilities of untreated syphilis-related APOs (P untreated ), as well as the summary probabilities of treated syphilis-related APOs (P treated ) occurring in Chinese pregnant women. Details of the methods and results of the meta-analysis were shown in S1 File. According to our meta-analysis (S1 File), P untreated for each adverse outcome category was 12.6% for spontaneous abortion, 20.0% for stillbirth, 12.9% for preterm birth/low birth weight, 8.0% for neonatal death, and 34.0% for congenital syphilis; P treated for each adverse outcome category was 3.6% for spontaneous abortion, 1.7% for stillbirth, 2.8% for preterm birth/low birth weight, 0.2% for neonatal death, and 8.8% for congenital syphilis.
Year-specific data inputs are displayed in Table 1.

Part 2 Translated the estimated number of APOs into DALYs
The DALYs are a comprehensive index that measures health loss due to both fatal and nonfatal disease burden. One DALY can be considered as one year of 'healthy' life lost [14]. In this analysis, we translated the estimated number of APOs into DALYs according to the methodology outline in the report of the WHO. For a specific cause (i.e., a disease or health condition), DALYs are calculated as the sum of the years of life lost (YLLs) due to premature mortality caused by that cause and the years of life lived with disability (YLDs) for people living in states of less than optimal health resulting from the specific cause [14]. The YLLs due to spontaneous abortion, stillbirth and neonatal death were calculated as the product of the neonatal standard loss function of 86.01 years and the number of incremental cases, calculated by the health service delivery model. To estimate the YLDs, the number of incremental cases were multiplied by the average duration of the disease (i.e., local life expectancy at birth) and its disability weights, which reflect the severity of the disease on a scale from 0 (perfect health) to 1 (dead). Based on the Global Burden of Disease 2004 Update, the disability weights were 0.106 for low birth weight (all sequelae) and 0.315 for congenital syphilis [15]. Since there was no evidence of disability weight for preterm birth, we set the disability weight of preterm birth/low birth weight equal to 0.106 as an alternative.

Part 3 Estimate the reduction in DALYs achieved by screening and treatment service
In this analysis, three screening and treatment coverage scenarios, including scenarios without any screening and treatment services for syphilis, scenarios of current screening and treatment coverage, and scenarios with coverage to meet the WHO-proposed targets for eliminating MTCT of syphilis, were used as the baseline, current, and best-case scenarios, respectively, to calculate the estimated DALYs resulting from adverse outcomes associated with maternal syphilis. The estimated reduction in DALYs achieved by screening and treatment services was calculated by subtracting the estimated DALYs in the current or best-case scenarios from the estimated DALYs in the baseline scenarios.

Results
Year-specific model results are displayed in Table 2. From 2011 to 2016, the estimated number of APOs associated with maternal syphilis in Hunan Province were 876, 676, 548, 600, 506 and 634, respectively, with a total of 3,840 cases over the six years. Approximately 74% (2,817) of these cases occurred among women who had at least one ANC visit but were not screened or treated for syphilis during pregnancy. The estimated numbers for each adverse outcome category over the six years were as follows: 602 cases of spontaneous abortion, 784 cases of stillbirth, 577 cases of preterm birth/low birth weight, 292 cases of neonatal death, and 1,585 cases of congenital syphilis. Table 3 showed the public health burden resulting from APOs associated with maternal syphilis in Hunan Province from 2011 to 2016. The total estimated number of syphilis-related APOs were translated into 192,528 DALYs, with an annual average of 32,088 DALYs. The specific breakdown of DALYs over the six years by outcome (% of total) was 51,778 DALYs (26.9%) for spontaneous abortion, 67,432 DALYs (35.0%) for stillbirth, 5,260 DALYs (2.7%) for preterm birth/low birth weight, 25,115 DALYs (13.0%) for neonatal death, and 42,943 DALYs (22.3%) for congenital syphilis. Approximately 76% (146,002 DALYs) of the total DALYs over the six years occurred among women who had at least one ANC visit but were not screened or treated for syphilis during pregnancy. Fig 2 shows the estimated DALYs resulting from APOs associated with maternal syphilis between 2011 and 2016. In the baseline scenario without any screening or treatment services, there would be 356,320 DALYs resulting from APOs associated with maternal syphilis over the six years. From 2011 to 2016, using the current scenario for syphilis screening and treatment, approximately 163,794 DALYs (46%) were averted, including 33,458 DALYs for spontaneous abortion, 67,861 DALYs for stillbirth, 3,984 DALYs for preterm birth/low birth weight, 28,986 DALYs for neonatal death, and 29,505 DALYs for congenital syphilis (Table 4) (Table 4).

Discussion
As we know, due to the paucity of accurate and reliable demographic and epidemiological data, it is hard to accurately calculate the public health burden due to APOs associated with maternal syphilis in China, even at the subnational level. For the first time, this analysis produced subnational estimations of the public health burden associated with syphilis-related APOs in China based on a mathematical model and determined the extent of progress made in the programme of preventing MTCT of syphilis in Hunan Province through comparison analysis.
In this analysis, we adopted the health service delivery model to estimate the number of pregnant women with syphilis infection, as well as the number of associated APOs in Hunan Province. This model took the existing screening and treatment service into account and removed background morbidity and mortality. The number of APOs calculated by the model were attributable to maternal syphilis infection, which truly reflected the effect of syphilis infection on the foetus and neonates. Furthermore, the model allowed us to replace the default values with more appropriate data, if it were available. Since there is no evidence of whether or not maternal syphilis has the same effect on pregnancy outcomes among Chinese and non-Chinese women, we integrated and analysed the related data to establish the extent of syphilis- related APOs among women with and without penicillin treatment in China through metaanalysis (S1 File). The results suggested that the summary estimates of the proportion of APOs due to untreated maternal syphilis among Chinese women (20.0% for stillbirth, 8.0% for neonatal death, 12.9% for preterm birth or low birth weight, and 34.0% for congenital syphilis) were inconsistent with those among non-Chinese women (21.0% for stillbirth or early foetal death, 9.3% for neonatal death, 5.8% for preterm birth or low birth weight, and 15.5% for congenital syphilis [16]). In addition, the default values of the effectiveness (E) of syphilis screening and penicillin treatment in averting APOs in the model were derived from the metaanalysis conducted in pregnant women with active syphilis (i.e., rapid plasma reagin titre ! 1:8) [5]. However, the model had not taken the baseline non-TP titre of pregnant women into account when calculating, which might lead to underestimation of the number of syphilis-related APOs. Moreover, in that meta-analysis, all pregnant women in the intervention arm received at least 2.4 million units of penicillin at least 28 days prior to delivery [5]. Nevertheless, the model did not involve the treatment time when estimating the number of pregnant women who were both tested and treated for syphilis during pregnancy, given that  current studies demonstrated a time length between the completion of the first penicillin treatment and delivery of less than 28 days as a risk factor for syphilis-related APOs [17,18], which might lead to an underestimate in the number of syphilis-related APOs as well. Therefore, in this analysis, we used the data derived from our meta-analysis (S1 File) to calculate the number of syphilis-related APOs in Hunan Province instead of the default values given in the model. In addition, as recommended, the number of pregnant women was calculated as the sum of the number of live births, stillbirths (after 28 week gestation) and early foetal deaths (22-28 weeks gestation), which did not include first trimester losses (miscarriages) [12]. However, due to the absence of the estimates of early foetal deaths in China and the fact that among women who have had at least one visit to ANC in China, 70% had their first ANC during early pregnancy ( 12 gestational weeks) [19], we could not estimate the number of pregnant women in Hunan Province by the recommended method. As an alternative, we assumed that the percentage of women with at least one ANC visit was equal to the antenatal examination rate used to calculate the number of pregnant women in Hunan Province which theoretically covered individuals who had miscarriages. Therefore, when calculating the number of syphilis-related APOs, we included the category of spontaneous abortions.
Based on the health service delivery model, our estimates suggested that over 3,800 syphilisrelated adverse pregnant outcomes occurred between 2011 and 2016, of which approximately 600 were spontaneous abortion, 780 were stillbirth, 580 were preterm birth/low birth weight, 290 were neonatal death and 1,580 were congenital syphilis. These results indicate that syphilis continues to be an important cause of substantial perinatal morbidity and mortality. Additionally, it is worth noting that this analysis does not include the additional infant deaths that may occur after the first month of life due to low birth weight, preterm birth, and syphilis infection in new-borns. As estimated by McDermott et al. [20], these events may result in approximately 10% additional infant deaths by 1 year. These findings indicate that preventing MTCT of syphilis is not only an effective way to reduce perinatal morbidity and mortality but also an important measure to decrease infant mortality during the first year of life. In addition, in this analysis, approximately 74% of the syphilis-related APOs occurred among women who had at least one ANC visit but were not tested or treated for syphilis if seropositive. This findings highlights the need to improve the quality of ANC while ensuring universal access to early ANC.
In this analysis, we found that approximately 192,528 DALYs occurred in Hunan Province between 2011 and 2016, with an average of 32,088 DALYs per year, which equated to 72% of DALYs associated with tuberculosis in the province in 2005 [21], 36% of DALYs associated with cardiovascular disease in the province in 2011 [22], and 30% of DALYs associated with HIV infection and acquired immunodeficiency syndrome (AIDS) throughout the country in 2007 [23]. These figures suggested that syphilis-related APOs had brought a heavy public health burden upon the society that even exceeded the burden of HIV/AIDS; this finding reminded policymakers charged with resource allocation to regard the elimination of MTCT of syphilis as a public health priority.
Existing screening and treatment services were able to avert nearly half of the expected public health burden resulting from syphilis-related adverse pregnancy outcomes in Hunan Province between 2011 and 2016; if WHO-proposed targets for eliminating MTCT of syphilis for screening and treatment had been met over the six years, more than 70% of the expected burden could have been averted. However, even though the target had been reached, a certain number of syphilis-related APOs still occurred each year, and the incidence of congenital syphilis was significantly higher than 50/100,000 live births. Therefore, further studies are needed to determine how many cases of congenital syphilis and what level of health service delivery are needed to achieve the ultimate goal of eliminating congenital syphilis as a public health problem. In addition, given that maternal syphilis infection is the primary cause of syphilis-related APOs, and that screening and treatment services for preventing MTCT of syphilis is not 100% effective, the prevention of maternal syphilis infection should also be taken as an important strategy to achieve the goal of full elimination of congenital syphilis. Due to the dynamics of syphilis transmission, the prevalence of syphilis among both the general and high-risk populations affects the prevalence of syphilis among pregnant women [24]. Therefore, to prevent pregnant women from syphilis infection, as well as to avert syphilis-related APOs fundamentally, more health resources should be invested into the prevention and control of syphilis in Hunan Province. Meanwhile, target interventions for the general and high-risk populations are necessary to improve control of the elimination of MTCT of syphilis. In addition, although the National Health and Family Planning Commission of the People's Republic of China demands medical and health institutions to provide services such as notification and testing for partners of pregnant women with confirmed syphilis infection, more than 50% of these partners were not tested for syphilis in Hunan Province. This poor enforcement may be an obstacle to the elimination of MTCT of syphilis. ANC workers should improve the testing for syphilis among partners of infected women and ensure adequate and prompt treatment for both infected women and their partners.
Although the proportions of syphilis screening and treatment during pregnancy have increased each year in Hunan Province since the implementation of the programme to prevent MTCT of syphilis, the DALYs caused by maternal syphilis infection increased in 2016 due to the increase in the number of pregnant women with syphilis infection. This trend further emphasizes the importance of preventing pregnant women from acquiring syphilis infection.
Our estimates are subject to some limitations. On the one hand, since there was neither a national or regional record of miscarriages nor a percentage of pregnant women with at least one ANC visit, we presumed that the percentage of pregnant women with at least one ANC visit was equal to the antenatal examination rate as an alternative to estimate the number of pregnant women, which might lead to a miscalculation. However, due to the limitation of information, we could not estimate the error caused by the assumption. On the other hand, in this analysis, when calculating the number of syphilis-related APOs based on the health service delivery model, we used the data derived from our meta-analysis (S1 File), which was considered to be more appropriate than the default values given in the model. Theoretically, we should compare the estimated number with case reports to see if there are significant differences and to verify whether the substitution is appropriate. However, due to the lack of appropriate valid data, we could not complete this task. Finally, since there was no data on the disability weight for preterm birth at present, we took that of low birth weight as an alternative, the calculation of which might be inconsistent with the actual disability weight. These limitations highlight the need to improve the quality of the data through strengthening the national monitoring and surveillance systems.

Conclusions
Our estimates indicate that syphilis in pregnancy continues to be an important cause of APOs, which can lead to substantial perinatal morbidity and mortality in Hunan Province. Approximately half of the expected public health burden resulting from syphilis-related APOs was averted by the current screening and treatment services, which suggests progress towards the elimination of MTCT of syphilis in Hunan Province. Nonetheless, untreated maternal syphilis remains a substantial cause of preventable perinatal morbidity and mortality, even in women receiving ANC. Improving access to quality antenatal care, including syphilis testing and treatment, and ensuring adequate and prompt treatment for infected women and their partners are all important for achieving the elimination of MTCT of syphilis.
Supporting information S1 File. Meta-analyses of the proportions of adverse pregnancy outcomes among women with treated and untreated syphilis. Contains the PRISMA Checklist. (DOC)