Strategies of testing for syphilis during pregnancy

Sadequa Shahrook; Rintaro Mori; Tumendemberel Ochirbat; Harumi Gomi

doi:10.1002/14651858.CD010385.pub2

Estrategias de pruebas para la sífilis durante el embarazo

Declaraciones de intereses de los autores

Versión publicada: 29 octubre 2014 Historial de versiones

https://doi.org/10.1002/14651858.CD010385.pub2

Contraer todo Desplegar todo

Resumen

disponible en

Antecedentes

Cada año, cerca de dos millones de mujeres embarazadas quedan infectadas por sífilis que se podría prevenir, principalmente en los países en desarrollo. A pesar del aumento de los programas de detección de la sífilis prenatal durante las últimas décadas, la sífilis es aún una preocupación importante de salud pública en los países en desarrollo. Las pruebas para la sífilis en un punto de atención sanitaria pueden ser una estrategia útil para prevenir de manera significativa la mortalidad perinatal asociada con la sífilis, así como otras consecuencias negativas en ámbitos de escasos recursos. Sin embargo, las pruebas de su efectividad se han generado principalmente de diseños de estudios observacionales o se han informado como un efecto de intervenciones mixtas.

Objetivos

Evaluar la efectividad del cribado de la sífilis prenatal para mejorar la captación para las pruebas de detección y el tratamiento, y para reducir la mortalidad perinatal.

Métodos de búsqueda

Se hicieron búsquedas en el registro de ensayos del Grupo Cochrane de Embarazo y Parto (Cochrane Pregnancy and Childbirth Group) (30 de septiembre de 2014) y en las listas de referencias de los estudios recuperados.

Criterios de selección

Ensayos controlados aleatorios (individuales y grupales) que compararan diferentes pruebas de detección realizadas durante el chequeo prenatal habitual versus ninguna prueba de detección. Los diseños de ensayos cruzados (crossover) y estudios experimentales cuasialeatorios no fueron elegibles para inclusión.

Obtención y análisis de los datos

Dos revisores de forma independiente evaluaron los ensayos para la inclusión y el riesgo de sesgo, extrajeron los datos y verificaron su exactitud.

Resultados principales

Se incluyeron dos ensayos controlados aleatorios grupales (tres informes). Ambos ensayos evaluaron las pruebas para la sífilis en puntos de atención sanitaria versus métodos de pruebas convencionales, y juntos incluyeron 8493 embarazadas. Los datos de estos ensayos no fueron apropiados para el metanálisis porque la medida de efectividad se evaluó de una manera no comparable.

Un ensayo asignó al azar 14 consultorios prenatales (incluidas 7700 embarazadas) y se realizó en Ulan‐Bator, Mongolia. El ensayo evaluó las pruebas inmediatas para la sífilis mediante una prueba rápida de treponema y se consideró que el riesgo fue incierto para los métodos de generación de la secuencia aleatoria, la ocultación de la asignación, el informe selectivo y otros sesgos, y que tuvo bajo riesgo de sesgo para los datos de resultado incompletos. El cegamiento no se informó y se evaluó como alto riesgo. Las pruebas realizadas en un punto de atención sanitaria proporcionaron un cribaje, los resultados de la prueba y el tratamiento en el mismo día. El ensayo parece haber ajustado sus resultados para considerar el agrupamiento. Se introdujeron los datos en RevMan con el uso del método de la varianza inversa genérica. La incidencia de sífilis congénita fue inferior en los grupos que recibieron el cribado en el lugar (odds ratio ajustado [ORA] 0,09; intervalo de confianza [IC] del 95%: 0,01 a 0,71) y la proporción de pacientes con pruebas para la sífilis fue mayor en los grupos que recibieron cribado en el lugar en la primera visita prenatal y en la visita al tercer trimestre (OR 989,80; IC del 95%: 16,27 a 60233,05; OR 617,88; IC del 95%: 13,44 a 28399,01). El tratamiento adecuado y el tratamiento a la pareja fueron mayores con el cribado en el lugar (ORA 10,44; IC del 95%: 1,00 a 108,99; ORA 18,17; IC del 95%: 3,23 a 101,20) y se detectaron más casos de sífilis en la vista inicial y al tercer trimestre con el cribado en el lugar (ORA 2,45; IC del 95%: 1,44 a 4,18; ORA 6,27; IC del 95%: 1,47 a 26,69). La mortalidad perinatal, la incidencia de infección por VIH/SIDA, los obstáculos en la captación para el cribado, otros efectos adversos, o el uso de los recursos de atención sanitaria no se informaron en este ensayo.

El segundo ensayo dividió los consultorios en siete pares pareados (que incluyeron 7618 embarazadas, aunque solamente se presentaron los resultados de los casos positivos [793 pacientes]), y dentro de cada par un consultorio se asignó al azar a recibir el cribado in situ y el otro a continuar con las pruebas de laboratorio habituales. El ensayo se realizó en consultorios de atención primaria en KwaZulu‐Natal, Sudáfrica. Se consideró que la generación de la secuencia aleatoria tuvo riesgo bajo de sesgo, pero la ocultación de la asignación y los datos de resultado incompletos se consideraron de alto riesgo. Otros sesgos y el sesgo de informe selectivo son inciertos. El cegamiento no se informó y se evaluó como alto riesgo de sesgo. Este ensayo evaluó el resultado primario de esta revisión (mortalidad perinatal) y los resultados secundarios (resultados adversos; tratamiento adecuado; prevalencia de sífilis) en el subgrupo de embarazadas (793 pacientes) con prueba positiva para la sífilis. Solamente un resultado, el tratamiento adecuado, se ajustó para tomar en cuenta el diseño grupal. Sin embargo, no se proporcionó información suficiente para incluirlo en un análisis que utilizó el método de la varianza inversa genérica. Cuando fue posible, los resultados se presentaron por lo tanto en diagramas de bosque (mortalidad perinatal; tratamiento adecuado), como si los datos provinieran de un ensayo controlado aleatorio paralelo. Estos resultados deben, por tanto, interpretarse con precaución.

El ensayo informó la mortalidad perinatal en las pacientes con resultados positivos de y mostró que el cribado in situ que utilizó una prueba de reagina plasmática rápida no proporcionó pruebas claras de un efecto sobre la reducción de la mortalidad perinatal (odds ratio [OR] 0,63; IC del 95%: 0,27 a 1,48; 18/549 [3,3%] versus 8/157 [5,1%]). Después de las pérdidas durante el seguimiento, 396/618 (64,1%) pacientes con resultados positivos recibieron tratamiento suficiente (dos o más dosis de 2,4 millones de unidades de penicilina benzatínica) en el grupo de intervención versus 120/175 (68,6%) en el control (OR 0,82; IC del 95%: 0,57 a 1,17). No fue posible incluir otros datos sobre los resultados informados en los diagramas de bosque (resultados adversos; prevalencia de sífilis). La incidencia de sífilis congénita, la proporción de pacientes con pruebas para la sífilis, la incidencia de infección por VIH/SIDA, los obstáculos para la captación para el cribado, el tratamiento de la pareja o el uso de los recursos de atención sanitaria no se informaron en este ensayo.

Conclusiones de los autores

Esta revisión incluyó pruebas de dos ensayos con asignación al azar grupal con riesgo alto o incierto de sesgo para la mayoría de los dominios del "riesgo de sesgo". Los datos no se combinaron en el metanálisis porque los ensayos utilizaron medidas de efectividad no comparables.

Las pruebas para la sífilis en un punto de atención sanitaria mostraron resultados alentadores para la detección de la sífilis y las tasas de tratamiento, y para el uso en diferentes contextos. Se encontró que las pruebas realizadas en un punto de atención sanitaria en Mongolia fueron eficaces para aumentar la proporción de embarazadas a las que se les realizaron pruebas para la sífilis y se les proporcionó tratamiento, reducir la sífilis congénita y mejorar el acceso al tratamiento de las pacientes y sus parejas. Por otra parte, en el área rural de Sudáfrica, entre las pacientes con resultados positivos de la prueba no hubo evidencias claras de que las pruebas para la sífilis en un punto de atención sanitaria tengan efecto sobre el aumento de las tasas de tratamiento adecuado de la sífilis y sobre la reducción de la mortalidad perinatal, pero se encontró que las pruebas realizadas en un punto de atención sanitaria redujeron el retraso para buscar tratamiento.

Por lo tanto, se justifican más ensayos para determinar la efectividad de las estrategias de pruebas disponibles para mejorar los resultados adversos asociados con la sífilis en las embarazadas y los recién nacidos, especialmente en las regiones de alto riesgo.

PICO

Population

Intervention

Comparison

Outcome

El uso y la enseñanza del modelo PICO están muy extendidos en el ámbito de la atención sanitaria basada en la evidencia para formular preguntas y estrategias de búsqueda y para caracterizar estudios o metanálisis clínicos. PICO son las siglas en inglés de cuatro posibles componentes de una pregunta de investigación: paciente, población o problema; intervención; comparación; desenlace (outcome).

Para saber más sobre el uso del modelo PICO, puede consultar el Manual Cochrane.

Resumen en términos sencillos

disponible en

Pruebas de detección prenatales para la prevención y el tratamiento de la sífilis

La sífilis es una infección de transmisión sexual causada por la bacteria Treponema pallidum. La afección es un tema importante de salud pública en los países en desarrollo, se desarrolla en cuatro etapas y es potencialmente mortal si no se trata. Una embarazada con sífilis puede transmitir la infección a su hijo, lo que puede dar lugar a una afección grave en los recién nacidos vivos, mortinatalidad o muerte neonatal. La infección por sífilis se puede transmitir por el contacto directo persona a persona a través de heridas abiertas en los labios, la boca, los genitales y otras áreas y durante el coito vaginal, anal u oral. Las heridas abiertas también aumentan el riesgo de infección por el virus de la inmunodeficiencia humana (VIH). El cribado universal de la sífilis dentro del programa de atención prenatal existente se ha recomendado como una manera eficaz de reducir los resultados adversos asociados con la sífilis. Sin embargo, a pesar de las décadas de uso de los programas de pruebas para la sífilis y de los adelantos significativos en la tecnología del cribado, la prevención y el tratamiento exitoso de la sífilis han sido limitados Lo anterior se debe en gran parte a los retrasos en la identificación y el tratamiento de las pacientes infectadas. Las dificultades técnicas y logísticas con las pruebas, la falta de atención prenatal y los servicios de calidad deficiente son posibles factores contribuyentes. Por lo tanto, es fundamental investigar los ensayos controlados aleatorios disponibles para determinar qué estrategias de pruebas son las más eficaces en los países en desarrollo.

Dos ensayos incluidos evaluaron las pruebas para la sífilis en un punto de atención sanitaria versus los métodos de prueba convencionales. El primer ensayo se realizó en Mongolia y comparó la prueba rápida de treponema con la prueba convencional. Las pruebas realizadas en un punto de atención sanitaria proporcionaron el cribado, los resultados de la prueba y el tratamiento en el mismo día. El ensayo informó una mejoría marcada en la cobertura del cribado, la detección de los casos y el tratamiento en la primera visita y en la visita al tercer trimestre, en comparación con el cribado convencional. El segundo ensayo se realizó en consultorios primarios de atención sanitaria en un área rural de Sudáfrica. El cribado en el lugar mediante la prueba de reagina plasmática rápida (RPR) se comparó con la prueba convencional. Entre las pacientes con prueba positiva para la sífilis, no se observó una reducción clara de las muertes perinatales en las que se les realizó una prueba de RPR en comparación con las que recibieron la prueba convencional, y se informaron dificultades técnicas y logísticas.

Ambos ensayos tuvieron principalmente riesgo alto o incierto de sesgo. En el ensayo realizado en Mongolia el cribado en el lugar fue mejor para detectar casos de sífilis. Se justifica la realización de más ensayos, especialmente en regiones donde la carga de morbilidad está aumentando y es probable la coinfección con VIH debido a la alta prevalencia de VIH/SIDA.

Authors' conclusions

Implications for practice

Reported primary outcomes were different between trials and point‐of‐care syphilis testing showed promising results for the detection and treatment of syphilis infection and congenital syphilis (Mongolia), and treatment delay (South Africa). Both trials were of high to unclear risk of bias. More trials are warranted to be able to further determine the effectiveness of testing strategies, especially for the prevention of syphilis and HIV co‐infection in regions with high HIV/AIDs endemicity.

Implications for research

This review includes two cluster‐randomised controlled trials and offers a basis for the integration of future trials investigating antenatal syphilis testing strategies that are indispensable in improving the adverse outcomes of the disease. This is particularly significant for developing context, especially in those locations with a high prevalence of syphilis and HIV/AIDs, a lack of laboratory testing, or poor‐quality antenatal clinic (ANC) services. Future trials should take greater care to avoid common risks of bias and report all important associated outcomes (e.g. HIV/AIDs), including a sufficient number of observations. Provision of HIV screening in syphilis trials would be of immense importance to generate valuable data to prevent syphilis and HIV/AIDs co‐infection. Moreover, assessment of the obstacles/challenges of screening uptake that pregnant women plausibly face both at the individual and health system levels would be important. For example, lack of knowledge or awareness of syphilis infection and screening tests (Munkhuu 2006), living far from antenatal clinics, and poor quality of ANC services (Walker 2002; Wilkinson 1998) lead to many pregnant women remaining unscreened and hence untreated. Neither of the included trials (Munkhuu 2009; Myer 2003) assessed the determinants of syphilis testing and treatment uptake; however, in the South African trial (Myer 2003), pregnant women with negative on‐site testing results were more likely to remain untreated. Therefore, an examination of the potential factors that may drive or prevent women from using syphilis screening services would be necessary. Also, an assessment of the technical and logistical difficulties in on‐site diagnosis using rapid plasma reagin (RPR) testing, e.g. separating the serum from cells and mixing with RPR antigen, has great potential to enhance syphilis detection and treatment in resource‐poor settings. Furthermore, it would also be useful to investigate whether women are charged a fee for screening services and if so, how much the fee costs, whether it is affordable and finally, what effect the fee has on screening uptake. Although antenatal care is usually provided free, syphilis screening may incur out‐of‐pocket payments in some settings, and elimination of such expenses could result in an increase in screening attendance (Cheng 2007). Future trials may also consider following a checklist for good quality reporting such as a modified Consolidated Standards Of Reporting Trials (CONSORT) for cluster‐randomised trials (Campbell 2004) to ensure better quality and generation of important information.

Background

Syphilis is a potentially fatal, sexually transmitted infection (STI) that can be transmitted from a pregnant woman to the fetus. Though preventable, each year about two million pregnant women become infected with syphilis globally, the majority of whom live in developing countries (Schmid 2004). The yearly toll of adverse birth outcomes associated with untreated maternal syphilis is 730,000 to 1,500,000, of which nearly 650,000 deaths occur in fetuses and newborns (Schmid 2007; WHO 2010). Maternal syphilis is less of a concern in developed countries than in developing countries. For example, the seroprevalence of women with syphilis attending antenatal care is estimated to be highest in Latin America (3.90%) and Africa (1.98%) (Schmid 2007). In sub‐Saharan Africa, mother‐to‐child transmission of syphilis is responsible for significant proportion of perinatal mortality count (21%) (Shafii 2008) and associated morbidities e.g. stillbirth and low birthweight (25%) (Watson‐Jones 2002), serious neonatal infection (20%) (Schulz 1987) and neonatal death (35%) (McDermott 1993). Furthermore, concern is deepening in countries such as China where an increase in disease incidence has already been observed (Cheng 2007; Tucker 2010). For example, in 2008 an average of more than one baby was born per hour with congenital syphilis among 9480 total cases; the observed amplification rate was by a factor of 12 during the five preceding years (Tucker 2010). Moreover, people with human immunodeficiency virus (HIV) are more likely to acquire syphilis infection and vice versa (Walker 2001). As a result, the rise in congenital syphilis in many countries in Sub‐Saharan Africa has been aggravated by HIV as this region is highly burdened by HIV infection (WHO 2010).

In approximately 69% of pregnant women with inadequately treated or untreated syphilis, the infection will be transmitted to the fetus causing severe birth outcomes such as spontaneous abortion, prematurity, stillbirth, low birthweight, neonatal death, or serious sequelae in liveborn infants (Hawkes 2011). However, these adverse outcomes are preventable, and existing approaches such as incorporated sexual and reproductive health programs, antenatal syphilis screening, and timely treatment have been suggested as a way to reduce syphilis‐attributable perinatal deaths and stillbirth incidence by about 50% (Bique 2000; Hawkes 2011; Myer 2003; Wilkinson 1998). Routine antenatal check‐ups have been widely promoted to pregnant women in developing countries (UNICEF 2009; WHO 2007), however despite the existence of antenatal screening policies in most countries, policy implementation is typically lacking (Gloyd 2001; Hossain 2007).

Additionally, the control and elimination of syphilis is hindered by a lack of testing among the majority of infected women. Of the women who are tested, most women do not undergo prompt treatment or are missed entirely (Newman 2013; WHO 2007; WHO 2010). Despite the availability of various improved diagnostic tools and cost‐effective prevention therapy (Peeling 2004; WHO 2010), attempts to prevent and eliminate syphilis are predominantly disrupted by the complex natural history of the disease and the absence of precise clinical presentation in infected patients (Peeling 2004). It has also been suggested that the absence of antenatal care, and poor‐quality services are likely to be important factors in the increase of congenital syphilis (Walker 2002; Wilkinson 1998).

Scientific efforts to prevent and eliminate congenital syphilis have been accelerated by the development of reliable and improved diagnostic tools such as on‐site syphilis testing, which provides rapid results and immediate therapy for seropositive women in primary care settings. In addition to laboratory testing, on‐site testing might be a useful strategy to curb congenital syphilis and its associated adverse outcomes by reducing treatment delays and increasing the numbers of seropositive women receiving treatment (Delport 1998; Fonn 1996; Jenniskens 1995). Although the effects of on‐site testing in observational studies (Bique 2000; Temmerman 2000) were positive, one randomised controlled study found no effective impact on either treatment rates or perinatal mortality reduction (Myer 2003). Despite the presence of laboratory access in some developing areas, the number of infected women who receive a full course of treatment remains low (Wilkinson 1997). Furthermore, in developing countries, useful screening tools such as treponemal tests are often obtained only at reference laboratories or large regional centres (Peeling 2004). Hence, syphilis screening has been constrained by varying dynamics and largely due to delays in diagnosis and treatment of infected women (Rotchford 2000). It is therefore crucial to identify from randomised controlled trials the most effective strategies for antenatal syphilis testing in order to improve the uptake of testing and treatment, and to reduce perinatal mortality.

Description of the condition

Syphilis is caused by the bacterium Treponema pallidum. The disease has protean clinical manifestations that may involve any organs and a range of severe health outcomes (CDC 2010). Syphilis infection is transmitted via direct person‐to‐person contact with an open syphilitic sore, and during vaginal, anal or oral sexual intercourse. The external genitals, vagina, rectum or anus are the main organs where sores usually occur, as well as the lips and inside the mouth. The risk of acquiring HIV infection in an individual with syphilis is two‐ to five‐fold if exposed when an ulcer is present, and consequently, individuals involved in high‐risk sexual behavior are likely to have a higher risk of syphilis and HIV co‐infection. Furthermore, the syphilis bacterium can be vertically transmitted to the fetus of a pregnant woman who has a syphilis infection, with at least two‐thirds of all newborns infected with maternal syphilis (Zenker 1990). The likelihood of fetal involvement occurs among women with active syphilis infection (i.e. rapid plasma reagin (RPR) titre greater than 1:4), specifically, with inadequately or untreated infection acquired in the five years prior to pregnancy (Blencowe 2011). Sixty‐nine per cent of women with active infection may experience a variety of adverse birth outcomes (Ingraham 1950; McDermott 1993), i.e. late miscarriage (after 16 weeks) or stillbirth in 25% cases, neonatal death at term in 11%, preterm or low birthweight in 13%, and classic symptoms and clinical signs of congenital syphilis in 20% (Ingraham 1950; McDermott 1993; Schmid 2004; Watson‐Jones 2002). Symptomatic newborns with congenital syphilis are usually born premature and classic signs of the condition include marasmus (acute malnutrition), pot belly, and wrinkled skin, especially on the face (Walker 2001). The severity of adverse birth outcomes associated with congenital syphilis is usually determined by the length of the maternal infection as well as the stage of pregnancy. The majority of pregnant women with syphilis as well as infected newborns at time of birth, are asymptomatic (Peeling 2004). Therefore, if not treated immediately, within a few weeks the disease progression can be fatal (CDC 2010).

Description of the intervention

Early detection and administration of appropriate therapies are at the centre of syphilis prevention strategies: undergoing syphilis screening tests at the first antenatal check‐up within the first trimester and again in the late stage of pregnancy followed by prompt treatment of sero‐positive women with a single dose of long‐acting penicillin before the second trimester (WHO 2010).

Serologic testing is the core strategy of syphilis screening and diagnosis (Hook 1992; Peeling 2004). The two main types of serologic tests are non‐treponemal tests and treponemal tests. Non‐treponemal tests identify active infection by detecting antibodies to imprecise antigens, e.g. cardiolipin present in the sera of patients with syphilis. Non‐treponemal tests such as the RPR test are easy to perform, sensitive, and relatively cheap (Peeling 2004). Furthermore, the non‐treponemal test is quantitative and the treatment response can be monitored over time (Fiumara 1978). On the other hand, treponemal tests are more specific than the former but in most cases, treponemal tests continue to be positive, regardless of treatment administration. In addition, treponemal tests, e.g. enzyme immunoassay (EIAs) are more costly than non‐treponemal tests and can be difficult to perform (Peeling 2004). Seroprevalence data from antenatal screening programmes are used as one of the proxy indicators to track the prevalence of sexually transmitted infections (Peeling 2004). Non‐treponemal tests such as RPR can be performed at a local laboratory but one of the major limitations is that RPR cannot be carried out on whole blood. Conversely, although useful to obtain prevalence rates and surveillance facts, confirmatory assays such as EIAs are usually only available at reference or large regional laboratories in resource‐poor settings. Currently, numerous improved sero‐diagnostic tools are available for the control and treatment of syphilis; for example, RPR and reagents of the Venereal Disease Research Laboratory (VDRL) test can be stored at room temperature. In addition, existing solar‐energy powered rotators have provided the means to carry out these tests in resource‐poor settings where there is a lack of, or no electricity (Peeling 2004). Rapid and easy treponemal tests using whole blood, serum or plasma can be stored at room temperature for six to 12 months, are cost‐effective (Peeling 2004), and overall have a performance comparable to laboratory tests (Fears 2001; Lien 2000). It is noteworthy that syphilis screening and treatment are estimated to be the most economical public health interventions available (WHO 2007).

How the intervention might work

Prevention success lies in the early detection of syphilis in pregnant women and prompt treatment management before the second trimester (WHO 2010). As recommended by the WHO, all pregnant women should undergo antenatal syphilis screening tests; however, women without test results at delivery should also be tested or re‐tested. Women should also be well informed about the importance of being tested for HIV infection. Additionally, HIV testing should also be offered to their partners and treatment planning should be primed in order to protect their infants at birth. Screening pregnant women in the early stage of pregnancy (preferably prior to 24 weeks of gestational age) can substantially avert the burden of associated adverse birth outcomes in many developing countries. Intial screening at routine antenatal check‐ups in the first trimester and again in the late stage of pregnancy, as well as prompt treatment of seropositive women is the desired approach to syphilis prevention. Once diagnosed, syphilis can be cured with a single dose of long‐acting penicillin, which prevents transmission to the unborn infant. Depending on the stage of disease progression, one (primary or secondary disease) or three (latent disease) doses of penicillin can effectively treat maternal syphilis (WHO 2010).

Why it is important to do this review

Evidence from randomised controlled trials on the effectiveness of screening strategies for the detection and treatment of maternal syphilis is scarce, and most knowledge is derived from observational studies. Moreover, earlier reviews of syphilis screening and treatment did not detect an intervention effect on preterm birth reduction (Barros 2010), nor a high grade of evidence (Menezes 2009). Therefore, in this review we attempted to accumulate quality evidence on the effectiveness of syphilis screening strategies in pregnant women and their neonates.

Objectives

To assess the effectiveness of antenatal syphilis screening in improving the uptake of screening tests and treatment, and reducing perinatal mortality.

Methods

Criteria for considering studies for this review

Types of studies

We planned to include all randomised (individual and clustered) controlled trials that compared different screening test strategies to detect and treat syphilis infection in pregnant women during routine antenatal check‐ups. The unit of randomisation could be either the individual pregnant woman or any formal healthcare facilities, e.g. health posts/clinics. We identified and included only two cluster‐randomised trials. Studies presented only as abstracts were considered, only if accompanied with appropriate publication status. Cross‐over trials and quasi‐randomised experimental study designs were not eligible for inclusion.

Types of participants

Eligible participants were either pregnant women at any stage of their pregnancy or healthcare facilities/clinics depending on the randomisation unit in each of the included trials.

Types of interventions

Antenatal syphilis screening tests versus no screening tests or standard screening. We also considered trials investigating the effect of combined screening strategies, only if the difference between arms was that of strategies for testing syphilis i.e. syphilis and HIV/AIDs testing versus HIV/AIDs testing.

Types of outcome measures

Primary outcomes

Perinatal mortality
Incidence of congenital syphilis**
Proportion of pregnant women tested for syphilis and treatment provided

Secondary outcomes

Incidence of HIV/AIDs in pregnant women and neonates
Obstacles/challenges in the uptake of antenatal syphilis screening tests e.g. availability of the useful syphilis tests and living far from the antenatal clinic (ANC) service spots
Any other adverse outcomes reported in the included studies (summarised)
Adequate treatment*
Partner treatment*
Syphilis prevalence*

Economic data for the use of healthcare resources

Pregnant women

Antenatal hospital admission

Neonates

Special care/intensive care admission

* Outcome not prespecified in our protocol.

** Outcome was prespecified in our protocol as a secondary outcome.

Search methods for identification of studies

The following methods are based on a standard template used by the Cochrane Pregnancy and Childbirth Group.

Electronic searches

We contacted the Trials Search Co‐ordinator to search the Cochrane Pregnancy and Childbirth Group’s Trials Register (30 September 2014).

The Cochrane Pregnancy and Childbirth Group’s Trials Register is maintained by the Trials Search Co‐ordinator and contains trials identified from:

monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL);
weekly searches of MEDLINE;
weekly searches of Embase;
handsearches of 30 journals and the proceedings of major conferences;
weekly current awareness alerts for a further 44 journals plus monthly BioMed Central email alerts.

Details of the search strategies for CENTRAL, MEDLINE and Embase, the list of handsearched journals and conference proceedings, and the list of journals reviewed via the current awareness service can be found in the ‘Specialized Register’ section within the editorial information about the Cochrane Pregnancy and Childbirth Group.

Trials identified through the searching activities described above are each assigned to a review topic (or topics). The Trials Search Co‐ordinator searches the register for each review using the topic list rather than keywords.

Searching other resources

We checked the reference lists of retrieved studies.

We did not apply any language restrictions.

Data collection and analysis

The following methods are based on a standard template used by the Cochrane Pregnancy and Childbirth Group.

Selection of studies

Two review authors, S Shahrook (SS) and R Mori (RM), independently assessed all potential studies identified by the search methods to determine whether they met all inclusion criteria. Any disagreement was resolved through discussion.

Data extraction and management

For eligible studies, SS and RM independently extracted data using a specified form. We resolved discrepancies through discussion. We entered data into Review Manager software (RevMan 2014) and checked for accuracy.

When information regarding any of the above was unclear, we contacted authors of the original reports to provide further details.

Assessment of risk of bias in included studies

SS and RM independently assessed risk of bias for each study using the criteria outlined in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We resolved any disagreement by discussion.

(1) Random sequence generation (checking for possible selection bias)

We described for each included study the method used to generate the allocation sequence in sufficient detail to allow an assessment of whether it should produce comparable groups.

We assessed the method as:

low risk of bias (any truly random process, e.g. random number table; computer random number generator);
high risk of bias (any non‐random process, e.g. odd or even date of birth; hospital or clinic record number);
unclear risk of bias.

(2) Allocation concealment (checking for possible selection bias)

We described for each included study the method used to conceal allocation to interventions prior to assignment and assessed whether intervention allocation could have been foreseen in advance of, or during recruitment, or changed after assignment.

We assessed the methods as:

low risk of bias (e.g. telephone or central randomisation; consecutively numbered sealed opaque envelopes);
high risk of bias (open random allocation; unsealed or non‐opaque envelopes, alternation; date of birth);
unclear risk of bias.

(3.1) Blinding of participants and personnel (checking for possible performance bias)

We described for each included study the methods used, if any, to blind study participants and personnel from knowledge of which intervention a participant received. We considered studies to be at low risk of bias if they were blinded, or if we judged that the lack of blinding would be unlikely to affect results. We assessed blinding separately for different outcomes or classes of outcomes.

We assessed the methods as:

low, high or unclear risk of bias for participants;
low, high or unclear risk of bias for personnel.

(3.2) Blinding of outcome assessment (checking for possible detection bias)

We described for each included study the methods used, if any, to blind outcome assessors from knowledge of which intervention a participant received. We assessed blinding separately for different outcomes or classes of outcomes.

We assessed methods used to blind outcome assessment as:

low, high or unclear risk of bias.

(4) Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data)

We described for each included study, and for each outcome or class of outcomes, the completeness of data including attrition and exclusions from the analysis. We stated whether attrition and exclusions were reported and the numbers included in the analysis at each stage (compared with the total randomised participants), reasons for attrition or exclusion where reported, and whether missing data were balanced across groups or were related to outcomes.

We assessed methods as:

low risk of bias (e.g. no missing outcome data; missing outcome data balanced across groups);
high risk of bias (e.g. numbers or reasons for missing data imbalanced across groups; ‘as treated’ analysis done with substantial departure of intervention received from that assigned at randomisation);
unclear risk of bias.

(5) Selective reporting (checking for reporting bias)

We described for each included study how we investigated the possibility of selective outcome reporting bias and what we found.

We assessed the methods as:

low risk of bias (where it is clear that all of the study’s pre‐specified outcomes and all expected outcomes of interest to the review have been reported);
high risk of bias (where not all the study’s pre‐specified outcomes have been reported; one or more reported primary outcomes were not pre‐specified; outcomes of interest are reported incompletely and so cannot be used; study fails to include results of a key outcome that would have been expected to have been reported);
unclear risk of bias.

(6) Other bias (checking for bias due to problems not covered by (1) to (5) above)

We described for each included study any important concerns we have about other possible sources of bias.

We assessed whether each study was free of other problems that could put it at risk of bias:

low risk of other bias;
high risk of other bias;
unclear whether there is risk of other bias.

(7) Overall risk of bias

We made explicit judgements about whether studies were at high risk of bias, according to the criteria given in the Cochrane Handbook (Higgins 2011). With reference to (1) to (6) above, we assessed the likely magnitude and direction of the bias and whether we considered it is likely to impact on the findings. We explored the impact of the level of bias through undertaking sensitivity analyses ‐ seeSensitivity analysis.

Measures of treatment effect

Dichotomous data

For dichotomous data, we have presented results as odds ratio with 95% confidence intervals.

Continuous data

For continuous data, we would have used the mean difference if outcomes were measured in the same way between trials. We planned to use the standardised mean difference to combine trials that measured the same outcome, but used different methods.

Unit of analysis issues

Cluster‐randomised trials

We included two cluster‐randomised trials in the analyses. We were unable to pool the summary effect as the measure of effectiveness was assessed in different ways. In one trial (Munkhuu 2009) the results appear to have been adjusted to account for clustering. The adjusted odds ratios [AOR] and P values were presented (Table 1) and we used these to calculate the log odds ratio and standard error and entered these into RevMan 2014 using the generic inverse‐variance method. In the other trial (Myer 2003), only one outcome, adequate treatment, was adjusted to account for the cluster design. However, not enough information was provided to include this in an analysis using the generic inverse variance method.

Open in table viewer

Table 1. Munkhuu 2009 ‐ Associations between main outcomes and the intervention analysed by multilevel logistic model

Outcome	Intervention n/N (%)	Control n/N (%)	Adjusted OR*	P
Congenital syphilis	1/3632 (0.03)	15/3564 (0.42)	0.09	0.022
Coverage at 1st antenatal visit	3849/3850 (99.9)	3065/3850 (79.6)	989.84	< 0.001
Coverage at 3rd trimester	3670/3683 (97.7)	2357/3796 (62.1)	617.88	< 0.001
Syphilis case at 1st visit	73/3849 (1.9)	27/3065 (0.9)	2.45	< 0.001
Syphilis case at 3rd trimester	20/3670 (0.5)	2/2357 (0.08)	6.27	0.013
Adequate treatment	92/93 (98.9)	26/29 (89.6)	10.44	0.05
Partner treatment	88/93 (94.6)	16/29 (55.2)	18.17	< 0.001

* Adjusted for significant independent variables such as gestational age at 1st antenatal visit and multiple sexual partners with a random effect for antenatal clinics.

OR: odds ratio.

In future updates of this review, if more cluster‐randomised trials are found, we will use effect estimates and their standard errors from correct analyses of cluster‐randomised trials and use these in meta‐analysis using the generic inverse‐variance method. However, if the trials fail to report an appropriate correct analysis, we will approximate analyses of cluster‐randomised trials to take account of the design effect. We plan to adjust their sample sizes using the methods described in the Cochrane Handbook using an estimate of the intracluster correlation co‐efficient (ICC) derived from the trial (if possible), from a similar trial or from a study of a similar population. If we use ICCs from other sources, we will report this and conduct sensitivity analyses to investigate the effect of variation in the ICC. If we identify both cluster‐randomised trials and individually‐randomised trials, we plan to synthesise the relevant information. We will consider it reasonable to combine the results from both if there is little heterogeneity between the study designs and the interaction between the effect of intervention and the choice of randomisation unit is considered to be unlikely.

We plan to acknowledge heterogeneity in the randomisation unit and perform a sensitivity analysis to investigate the effects of the randomisation unit.

Trials with more than two treatment groups

We planned to include trials with more than two intervention groups (multi‐arm studies), if identified. In future updates of this review, if trials with more than two intervention groups are identified, only directly relevant arms will be included. If studies with various relevant arms are identified, groups will be combined to generate a single pair‐wise comparison (Higgins 2011), and the disaggregated data in the corresponding subgroup category will be included. If the control group is shared by two or more study arms, the control group over the number of relevant subgroup categories will be divided to avoid double‐counting the participants (for dichotomous data, we will divide the events and the total population, and for continuous data, we will assume the same mean and standard deviation but will divide the total population). The details will be described in the Characteristics of included studies tables.

Cross‐over trials

For this review cross‐over trials have been judged invalid as they are generally considered to be inappropriate while measuring a primary outcome which is irreversible, such as mortality as described in the Cochrane Handbook for Systematic Reviews of Interventions section 16.4.

Dealing with missing data

For included studies, we noted levels of attrition. In future updates, we will explore the impact of including studies with high levels of missing data in the overall assessment of treatment effect by using sensitivity analysis.

For all outcomes, we carried out analyses, as far as possible, on an intention‐to‐treat basis, i.e. we attempted to include all participants randomised to each group in the analyses, and all participants were analysed in the group to which they were allocated, regardless of whether or not they received the allocated intervention. The denominator for each outcome in each trial was the number randomised minus any participants whose outcomes are known to be missing.

Assessment of heterogeneity

We were unable to conduct meta‐analysis combining the two included trials. However, in future updates of this review, we plan to assess statistical heterogeneity in each meta‐analysis using the Tau², I² and Chi² statistics. We will regard heterogeneity as substantial if an I² is greater than 30% and either a Tau² is greater than zero, or there is a low P value (less than 0.10) in the Chi² test for heterogeneity.

Assessment of reporting biases

In future updates of this review, if there are 10 or more studies in a meta‐analysis, we plan to investigate reporting biases (such as publication bias) using funnel plots. We will assess funnel plot asymmetry visually. If asymmetry is suggested by a visual assessment, we will perform exploratory analyses to investigate it.

Data synthesis

We carried out statistical analysis using the Review Manager software (RevMan 2014). We were unable to meta‐analyse the results as the measure of treatment effectiveness was assessed in a non‐comparable way. In one trial (Munkhuu 2009), the results appear to have been adjusted to account for clustering. The adjusted odds ratios [AOR] and P values were presented (Table 1) and we used these to calculate the log odds ratio and standard error and entered these into RevMan 2014 using the generic inverse‐variance method. In the other trial (Myer 2003), only one outcome, adequate treatment, was adjusted to account for cluster design. However, not enough information was provided to include this in an analysis using the generic inverse variance method. We presented some of these data in forest plots, as if for a parallel randomised‐controlled trial, and for the remaining data used a narrative synthesis of the data and presented them in a consistent manner as described in the Cochrane Handbook (section 11.7.2).

In future updates of this review, we plan to use fixed‐effect meta‐analysis for combining data where it is reasonable to assume that studies are estimating the same underlying treatment effect: i.e. where trials are examining the same intervention, and the trials’ populations and methods are judged sufficiently similar. If there is clinical heterogeneity sufficient to expect that the underlying treatment effects differ between trials, or if substantial statistical heterogeneity is detected, we will use random‐effects meta‐analysis to produce an overall summary if an average treatment effect across trials is considered clinically meaningful. The random‐effects summary will be treated as the average range of possible treatment effects and we will discuss the clinical implications of treatment effects differing between trials. If the average treatment effect is not clinically meaningful, we will not combine trials.

If we use random‐effects analyses in future updates of this review, we plan to present results as the average treatment effect with its 95% confidence interval, and the estimates of Tau² and I².

Subgroup analysis and investigation of heterogeneity

We were unable to conduct either an investigation of heterogeneity or pre‐specified subgroup analyses as only two trials were included. Investigation of publication bias presentation using a funnel plot was also not possible due to the small number of trials. However, in future updates of this review, if we identify substantial heterogeneity, we plan to investigate it using subgroup analyses and sensitivity analyses. We plan to consider whether an overall summary is meaningful by using a random‐effects analysis. In future updates of this review, if sufficient or appropriate data are available, we plan to carry out the following subgroup analyses.

Low‐income versus middle‐income countries.
Study settings, i.e. antenatal clinics versus other healthcare facilities.
HIV/AIDs status of pregnant women and neonates.
Syphilis screening strategies including HIV/AIDs versus without HIV/AIDs screening.

The following outcomes will be used in subgroup analysis.

Perinatal mortality.
Incidence of congenital syphilis.
Proportion of pregnant women tested for syphilis and treatment provided.

We will assess subgroup differences by interaction tests available within RevMan (RevMan 2014). We will report the results of subgroup analyses quoting the Chi² statistic and P value, and the interaction test I² value.

Sensitivity analysis

We were unable to conduct a sensitivity analysis to investigate the robustness of the results as too few trials were detected. However, in future updates of this review, sensitivity analyses will be performed to assess the risk of bias effects (trials with low or unclear sequence generation and allocation concealment and either high levels of attrition or inadequate blinding) on the analyses, by omitting trials rated as 'high risk of bias' for these components.

Results

Description of studies

Results of the search

The search of the Pregnancy and Childbirth Group's Trials Register retrieved three reports of two studies (Munkhuu 2009 based in Mongolia; Myer 2003 based in South Africa) meeting all the inclusion criteria for this review. Rotchford 2000 reported baseline findings of the trial carried out in South Africa (Myer 2003) where perinatal mortality was estimated through modelling (See:Figure 1).

Figure 1

Study flow diagram.

Included studies

Design

Both of the included trials investigated point‐of‐care syphilis testing effect during routine antenatal check‐ups for the detection and treatment of syphilis infection in pregnant women, using cluster‐randomised trials (C‐RCTs) design. The trials were published between 2000 and 2009.

Sample size

The trial by Munkhuu 2009 included 7700 pregnant women in 14 clinics; whereas, in Myer 2003, 7618 pregnant women were eligible, although results were only presented for the positive cases (793 women).

Setting

The Munkhuu 2009 trial was conducted in antenatal clinics in Ulaanbaatar, Mongolia. The Myer 2003 study was a primary healthcare setting‐based intervention carried out in KwaZulu‐Natal, South Africa.

Participants

In the Mongolian trial (Munkhuu 2009), all pregnant women with a single pregnancy attending their first antenatal care were included. Statistical adjustments were made for the significant differences in participants' characteristics, i.e. gestational age at first antenatal visit and multiple sexual partners. In Myer 2003, pregnant women attending their first antenatal care with a positive syphilis test were offered treatment and were followed up and analysed (793 women).

Interventions

Both of the trials assessed the on‐site syphilis testing strategy performed during antenatal check‐ups and compared it with traditional laboratory testing.

Munkhuu 2009 assessed on‐site syphilis screening using rapid treponemal tests at the first antenatal visit and at the third trimester of gestation, and compared on‐site screening with conventional testing using rapid plasma reagin (RPR), and where necessary, Treponema pallidum haemagglutination assay (TPHA) at any district‐level health facilities. The other elements of their one‐stop service included: immediate on‐site treatment for sero‐positive women and their sexual partners, including the administration of G 2.4 million units of intramuscular benzathine penicillin, and counselling before and after the test. The trial included 14 antenatal clinics randomly assigning seven clinics each to the intervention and control group using clinics as the unit of randomisation. In the intervention group, the on‐site rapid treponemal test was assayed with SD Bioline Syphilis 3.0 (Standard Diagnostics Inc., Kyunggi‐do, Korea) as per the manufacturer's specifications, and the results were available the same day and within 10 to 15 minutes. Blood specimens of the women with sero‐positive test results were further collected and sent to the reference laboratory for confirmation testing, which was carried out using RPR tests, with further testing for the positive samples by TPHA if required. After both tests, seropositive women were subsequently administered two free doses of benzathine penicillin at weekly intervals. Women who tested negative on the rapid assays received post‐test counselling by antenatal service providers and were invited for a second screening during the third trimester of pregnancy. Providers in the intervention clinics received two days of workshop training which covered various aspects of antenatal syphilis screening, such as decentralisation of services, case detection and management. After admission to the control clinics, pregnant women were asked to visit any District General Hospital or the National Center of Infectious Diseases for free initial testing with RPR and/or TPHA confirmation test. Providers in the control clinics also received two days of training but on different aspects of the program, such as the project overview and refresher training.

Myer 2003 went on to assess the effectiveness of on‐site syphilis testing using the RPR test complemented by laboratory confirmation, and compared this strategy with routine testing at the provincial reference laboratory. Using clinics as the unit of randomisation, seven matched clinic pairs were each randomly assigned to the intervention and control group. In the intervention group, test results and treatment (the first of three recommended weekly intramuscular injections of 2.4 mega units of benzathine penicillin) were available on the same day and were administered within an hour by clinic nurses. In this test, cells were separated from plasma by standing the samples for 30 minutes, and serum was mixed with RPR antigen using a battery powered rotator as per the manufacturer’s instructions. Positive or negative test results were reported by the clinic nurses using colour photographs as a reference standard. Nurses at the intervention clinics were provided with a one‐day training workshop on a range of on‐site syphilis testing operations, including logistics information, specimen handling and maintenance, and supervised practice. Continued monitoring of test performance and refresher training for new clinic staff were also provided throughout the trial. In the control group, blood specimens were sent to the provincial reference laboratory for routine testing and results were available upon returning to the laboratory two weeks following the test. Upon returning, sero‐positive women were counselled on the effective doses of the treatment regimen and received their first penicillin injection. All women with RPR positive results (on‐site or laboratory tests) were advised to return to the clinics for two follow‐up injections at weekly intervals.

Outcomes

Many pre‐specified outcomes of this review were not reported in the two included trials. Munkhuu 2009 reported two pre‐specified primary outcomes: the proportion of women tested for syphilis and treatment provided at the first antenatal visit and again at the third trimester of gestation, and congenital syphilis (number of cases). In addition to reporting adequate treatment rates (percentage of sero‐positive cases completing three doses of benzathine penicillin before delivery), the trial also reported the proportion of completely treated sexual partners of seropositive women (reported as a non‐pre‐specified review outcome). One of the reports from South Africa by Rotchford 2000 used modelling to estimate the reduction of perinatal mortality risk based on the number of penicillin doses. However, Myer 2003 reported one of the pre‐specified review primary outcomes of perinatal mortality. Treatment delay and maternal syphilis were not pre‐specified in the protocol but were considered later in this review as important outcome data. None of the trials examined obstacles/challenges in the uptake of antenatal syphilis screening tests (review secondary outcome) and incidence of HIV/AIDs in pregnant women and neonates, including economic data on healthcare resource usage, which was a secondary outcome of this review. Other adverse outcomes (review secondary outcome) was reported in one trial (Myer 2003). Comparison with a pooled effect estimate was constrained due to the incomparable modes of action (see also Subgroup analysis and investigation of heterogeneity); however, we presented results in individual forest plots.

Excluded studies

There are no excluded studies in this review.

Risk of bias in included studies

The two included cluster‐randomised controlled trials (C‐RCTs) were judged to be at unclear to high risk of bias overall. A summary of the risk of bias for the two included trials is provided in Figure 2 and Figure 3.

Figure 2

'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Figure 3

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Generation of sequence

Munkhuu 2009 reported that the groups were randomised but the randomisation method was not reported, and therefore, the risk of bias is unclear. However, sequence generation was adequate in Myer 2003, where clinics were randomised through a coin‐toss method.

Allocation concealment

Allocation concealment bias is unclear in the trial by Munkhuu 2009. The authors did not report whether allocation concealment was used and therefore we cannot discard the possibility of baseline imbalance, in spite of the unlikelihood of C‐RCTs to suffer from allocation bias. Also, randomisation of the clusters was not pair‐matched or stratified but baseline characteristics of the groups were reported to be well‐balanced. Since allocation in Myer 2003 was unblinded, we cannot rule out the possibility of high risk of bias. However, we assume that the baseline imbalance in this trial was reduced due to pair‐matched randomisation of the clusters.

Blinding

We assume both of the trials (Munkhuu 2009; Myer 2003) to be at high risk of performance and detection bias since they did not report on blinding of participants, clinicians and outcome assessors.

Incomplete outcome data

Only losses to follow‐up were reported: in Munkhuu 2009, the rates were 5.7% versus 7.4% (intervention versus control) and this was judged to be at low risk of bias. However, in Myer 2003, of 7134 women seeking antenatal care with available results, 793 (11.1%) tested positive for syphilis and the results are presented only for this subgroup of women. It is not clear why outcome data are presented only for those women where syphilis was detected. The denominator here should be the number of women in each trial arm (5201 onsite versus 2417 control clinic). For this reason we judged incomplete outcome data to be at high risk of bias for Myer 2003.

Selective reporting

The risk of reporting bias for the included trials is unclear (Munkhuu 2009; Myer 2003). With no access to the protocols for both, it was not possible to determine whether outcome data for all prespecified outcomes were reported.

Other potential sources of bias

The risk of recruitment bias, baseline imbalance, loss of clusters, early termination of the trial, and incorrect analysis were assessed in both trials. In Munkhuu 2009, the trial appears to have adjusted for the main outcomes using univariate and multivariate analysis. All antenatal care (ANC) service providers in the catchment were included in the randomisation except the Maternal and Child Health (MCH) Centre and two out of 16 ANC clinics. The MCH centre was excluded because it is responsible for the whole country, and the two ANC clinics had small attendance. The study groups were reported to be well‐balanced for the distribution of baseline characteristics, including statistical adjustment made for participants' characteristics such as mean age and marital status. In Myer 2003, baseline imbalance due to one very busy clinic in the intervention arm was reported in terms of participants' age and gravidity. However, adjusted analyses (not shown in the paper) did not alter the trial's findings substantially, as was reported by the authors. We also assumed that the differences in the groups were somewhat reduced by pair‐matched randomisation of the clusters. Nevertheless, contamination between groups in both of the C‐RCTs cannot be excluded as allocation concealment was either not reported (Munkhuu 2009) or was unblinded (Myer 2003). Therefore, the quality of the trials' designs were judged to be at unclear risk of other bias.

Effects of interventions

Both trials assessed point‐of‐care syphilis testing with conventional testing methods and together involved a total of 8493 pregnant women. Meta‐analysis of the two included trials was not possible as the measure of intervention effectiveness was assessed in a non‐comparable way. It is for this reason that the results have been presented separately as different comparisons.

In one trial (Munkhuu 2009) the results appear to have been adjusted to account for clustering. The adjusted odds ratios [AOR] and P values were presented (see Table 1) and these were used to calculate the log odds ratio and standard error and we entered these into RevMan 2014 using the generic inverse‐variance method.

In the other trial (Myer 2003), only one outcome, adequate treatment, was adjusted to account for cluster design. However, not enough information was provided to include this in an analysis using the generic inverse variance method. Where possible, results have therefore been presented in forest plots, as if the data are from a parallel randomised controlled trial. These results should therefore be interpreted with caution, as no adjustments have been made to account for clustering.

On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with rapid plasma reagin (RPR) or treponema pallidum hemagglutination (TPHA)

Primary outcomes

Perinatal mortality

This outcome was not reported by Munkhuu 2009.

Incidence of congenital syphilis (non‐prespecified outcome)

The number of congenital syphilis cases was reported in the Munkhuu 2009 trial in which one‐stop antenatal syphilis screening found a 93.5% (95% confidence interval (CI), 66.0% to 98.6%) reduction in congenital syphilis, and detected only one congenital syphilis case out of 3632 deliveries in the intervention group compared to 15 out of 3552 in the control (P = 0.002), (adjusted odds ratio (AOR) 0.09, 95% CI 0.01 to 0.71), Analysis 1.1.

Proportion of pregnant women tested for syphilis and treatment provided

Munkhuu 2009 reported syphilis testing and treatment as primary outcomes. Both were assessed twice, at the first antenatal visit and again, at the third trimester of gestation; the following changes were detected in the intervention group.

The one‐stop syphilis screening intervention significantly achieved better testing and treatment coverage than the conventional screening test: of 3850 pregnant women recruited in each group, 99.9% (intervention) and 79.6% (control) were screened at the first visit (AOR 989.80, 95% CI 16.27 to 60233.05), Analysis 1.2. Of the remaining participants, 3683 and 3796 pregnant women (excluding loss to follow‐up) from the intervention and control respectively revisited at the third trimester, 99.7% (intervention) and 62.1% (control) had the second test (AOR 617.88, 95% CI 13.44 to 28399.01; P= < 0.001), Analysis 1.3.

Secondary outcomes

Obstacles/challenges in the uptake of antenatal syphilis screening testing was not investigated in Munkhuu 2009. Similarly, Munkhuu 2009 did not report data on the incidence of HIV/AIDs in pregnant women and neonates, any adverse outcomes or healthcare resource usage.

Adequate treatment (non‐prespecified outcome)

One‐stop screening also showed significant impact on improving treatment rates in pregnant women and their sexual partners (Munkhuu 2009). The proportion of women who received adequate treatment (three doses of G 2.4 million units of intramuscular benzathine penicillin) was 98.9% (92/93) in the intervention versus 89.6% (26/29) in the control group AOR 10.44, 95% CI 1.00 to 108.99; P= 0.05), Analysis 1.4.

Partner treatment (non‐prespecified outcome)

On‐site testing significantly improved the completion of recommended treatment doses in sexual partners of seropositive women: 94.6% (88/93) in the intervention versus 55.2% (16/29) in the control clinics (AOR 18.17, 95% CI 3.23 to 101.20; P = < 0.001), Analysis 1.5,

Syphilis prevalence (non‐prespecified outcome)

Munkhuu 2009 reported maternal syphilis cases at the first ANC visit: 73/3849 in the intervention versus 27/3065 in the control (AOR 2.45, 95% CI 1,44 to 4.18; P = < 0.001) Analysis 1.6; and at the third trimester, 20/3670 versus 2/2357, respectively (AOR 6.27, 95% CI 1.47 to 26.69), Analysis 1.7.

On‐site syphilis testing with rapid plasma reagin (RPR) test versus routine laboratory testing with RPR

Primary outcomes

Perinatal mortality

Perinatal mortality, defined as either stillbirth (death at or after 28 weeks’ gestation) or early neonatal death (death up to seven days postpartum), was reported in Myer 2003 but significant on‐site syphilis testing effect on the reduction of perinatal loss was not detected. Among the 618 pregnant women with syphilis in the intervention clinics, pregnancy outcomes were available for 561, of which 549 full‐term pregnancies were reported. In the control clinics, pregnancy outcomes were available for 162/175, of which 157 pregnancies were at full term. Overall, 3.3% perinatal deaths were observed in the intervention group (18/549) versus 5.1% (8/157) in the control (OR: 0.63; 95% CI: 0.27 to 1.48), Analysis 2.1.

Proportion of pregnant women tested for syphilis and treatment provided

On‐site syphilis testing was insignificant to effectively treat maternal syphilis in South Africa (Myer 2003). Out of the 7134 women with available test reports, 793 (11.1%) were detected to be sero‐positive and enrolled in the trial, of which 618 women were allocated to the intervention group and 175 to the control (Myer 2003). After losses to follow‐up, 396/618 (64.1%) women with positive test results received adequate treatment (two or more doses of 2.4 mega units of benzathine penicillin) in the intervention cluster versus 120/175 (68.6%) in the control cluster (odds ratio (OR): 0.82; 95% CI 0.57 to 1.17), Analysis 2.2. The proportion of women receiving inadequate treatment (defined as one dose or no treatment) was reported as 125/618 (20.2%) in the intervention versus 34/175 (19.4%) in the control (OR: 1.05; 95% CI 0.69 to 1.60), Analysis 2.2. Among the infected women receiving inadequate treatment in the intervention clinics, 63/125 needed to receive the recommended treatment dosages (based upon laboratory test reports or clinical decision); however, none obtained the recommended dosages but were found negative in on‐site testing (50%). Conversely, 17% of women who received a minimum of one penicillin dosage showed negative results in on‐site testing (86/493). These findings therefore led the authors (Myer 2003) to conclude that women’s negative sero‐positivity at on‐site testing had a strong influence on whether they remained untreated (ARD, intervention minus control, 28%; 95% CI: 20 to 36).

The included trial did not report data on the incidence of congenital syphilis.

Secondary outcomes

Adverse outcomes

Obstacles/challenges in the uptake of antenatal syphilis screening tests (a secondary outcome of this review) were not investigated in any of the included trials (Myer 2003); however, important and relevant information discussed in the papers is included in our discussion section. Other adverse outcomes were reported in Myer 2003. Of the 723 women who tested positive and for whom available pregnancy outcome data were available (91%), 12 women in the intervention arm (2.1%) and 5 in the control (3.1%) reported miscarriages, and were excluded from the perinatal mortality analysis, Analysis 2.3.

The included trial did not report data on the incidence of HIV/AIDs in pregnant women and neonates, or healthcare resource usage.

Syphilis prevalence (non‐prespecified outcome)

Myer 2003 reported the prevalence of laboratory‐confirmed syphilis as 7.5% which was similar in both intervention and control clinics, after accounting for the matched‐pair design of the study. In both on‐site and laboratory testing, positive syphilis rates were 5% (246) in the intervention versus 6% (149) laboratory‐confirmed cases in the control clinics, Analysis 2.4. Myer 2003 also reported RPR titre data determined from 346 women (55%) from the intervention group and 134 women (75%) from the control group, and observed that a median titre level of 1:16 or higher was documented among 20% of women. On‐site syphilis testing significantly reduced treatment delays in South Africa (Myer 2003). Women in the intervention clinics completed syphilis treatment an average of 16 days earlier compared to women attending control clinics (95% CI 11 to 21; P= < 0.001), Analysis 2.5.

Discussion

Summary of main results

We included two cluster‐randomised controlled trials (totaling 8493 pregnant women) assessing point‐of‐care syphilis testing to improve maternal and perinatal outcomes. The trials were based in Mongolia (Munkhuu 2009) and South Africa (Myer 2003) and judged overall to be of a high/unclear risk of bias. Meta‐analysis of the two included trials was not possible as the measure of intervention effectiveness was assessed in a non‐comparable way. It is for this reason that the results have been presented separately as different comparisons.

The Mongolian trial (Munkhuu 2009) reported on syphilis testing in the first trimester and retesting in the third trimester—a useful strategy to remedy the risk of syphilis reinfection as suggested by experts in the field (Newman 2013; WHO 2007). The trial's one‐stop intervention package was successful in improving the use of antenatal syphilis screening, syphilis case detection and treatment for infected women and their sexual partners, including congenital syphilis cases. Conventional screening approaches are often inconvenient; in Mongolia, at least one additional visit is required for women to be completely tested, in addition to further visits to receive their test results. This approach possibly deters a large number of pregnant women from pursuing screening and treatment. Conversely, the decentralised testing services provide screening, test results and treatment within the same day. The one‐stop approach offers greater convenience to women, and has been indicated by the authors as one of the influencing factors to increase coverage at both testing points during pregnancy. However, the intervention was implemented within a relatively well‐developed antenatal care system with linkages to district general hospitals and a MCH centre (excluded from the sampling), which was likely a key factor in determining the success of this approach. The characteristics of the screening test used might also be influential, for instance, useful screening tools such as rapid treponemal tests are often obtainable only at reference laboratories or large regional centres; perhaps because they are more costly compared to the non‐treponemal test and can be difficult to perform (Peeling 2004). Furthermore, women screened at the first trimester of gestational age were significantly different at baseline, as were those with multiple sexual partners. It has been suggested that gestational age could be one of the predictors of the risks associated with congenital syphilis, which might be reduced by early screening and prompt treatment of women (Liu 2010), for which supporting evidence was offered by Munkhuu 2009. Additionally, one‐stop services showed promise in the notification and treatment of sexual partners, and doctors' active counselling was identified as a key catalyst to achieve success in this area. Partner notification and treatment is also considered useful for the prevention of syphilis reinfection risks and associated adverse outcomes, and can be achieved through the delivery of existing antenatal screening programmes (WHO 2007; WHO 2010). However, negative consequences of partner notification may occur, such as domestic violence. Although not addressed in either of the included trials (Munkhuu 2009; Myer 2003), a cross‐sectional study (Díaz‐Olavarrieta 2007) documented both women's fear of negative reactions from their partners and underreporting of syphilis infection. Future screening programmes may consider integration of effective counselling services to identify potential risk of domestic violence and develop partner notification strategies that protect the safety of pregnant women.

The trial by Myer 2003 was conducted in primary care clinics in rural South Africa. No statistical differences between the intervention and control groups were found in terms of testing or treatment, for which "the relatively high quality of laboratory services provided in the control arm" was indicated as a possibility. The trial found no effect on perinatal mortality reduction, though a 50% reduction was assumed at baseline (Rotchford 2000). As one of the possibilities of this insignificant effect, failure in treating larger proportion of women in the intervention clinics mainly due to technical and logistical difficulties was pointed out. Such difficulties were related to the relatively complex process of on‐site diagnosis using rapid plasma reagin (RPR) testing, e.g. separating the serum from the cells and mixing the serum with RPR antigen. Although a battery powered rotator was used for this process, the test performance could also have been restricted by the fact that in most cases RPR reagents require refrigeration (Peeling 2004). Furthermore, high workloads, limited staffing, and difficulty in maintaining a daily supply of testing materials and holding admitted women in clinics while they awaited test results have also been identified as critical challenges in achieving optimally effective results. The trial distributed standard partner notification cards to seropositive women, but partners' treatment was not reported. Also, miscarriages were documented in both study groups. We contacted Myer 2003 for additional information but did not receive any reply.

Neither trial reported data on healthcare resource use, but both were able to detect a greater number of syphilis cases. Also, neither trial reported whether the participants were encouraged to be screened for HIV infection, which is another viable approach for the prevention of sexually transmitted infections (WHO 2007; WHO 2010) since individuals with HIV/AIDs are highly susceptible to syphilis co‐infection and vice versa (Walker 2001). The trials' findings indicate that the effectiveness of point‐of‐care syphilis testing might depend on varying dynamics such as the setting in which a particular screening test is implemented, as well as the existing service delivery system. Even though the Mongolian trial was carried out in a well‐functioning antenatal service delivery system, and support for health systems strengthening was provided in both trials through training, laboratory support and other activities, success rates varied between both trials, and it was difficult to determine exactly which component/s were effective, and for which particular outcomes.

Overall completeness and applicability of evidence

Three of the pre‐specified review primary outcomes (proportion of pregnant women tested for syphilis and treatment provided, congenital syphilis and perinatal mortality) were reported in the two included trials. Although we were unable to provide an overall summary effect through data pooling, point‐of‐care syphilis testing showed effectiveness in different aspects of syphilis prevention in both included trials. While the one‐stop service in Mongolia significantly improved the screening coverage, case detection of maternal syphilis, treatment quality in seropositive women including their partners (Munkhuu 2009), and reduced congenital syphilis cases, the addition of on‐site testing in South Africa did not improve treatment rates or perinatal deaths, although treatment delays were markedly decreased (Myer 2003). In future updates of this review, we aim to provide a further complete assessment in this critical area of maternal and neonatal health as more data become available.

Quality of the evidence

Munkhuu 2009 was judged to have unclear methods of random sequence generation, allocation concealment, selective reporting, and other bias. However, outcome data were addressed adequately with intention‐to‐treat analysis and multilevel regression modelling. Losses to follow‐up were reported at random only due to logistical and technical difficulties. In the second trial (Myer 2003), methods of random sequence generation were adequate, but allocation concealment was judged to be inadequate. Incomplete outcome data were an issue, because of 7134 women seeking antenatal care with available results, 793 (11.1%) tested positive for syphilis and results are presented only for this subgroup of women. It is not clear why outcome data are presented only for those women where syphilis was detected and why the denominator was not the number of women in each trial arm (5201 onsite versus 2417 control clinic). For this reason, we judged incomplete outcome data to be at high risk of bias for Myer 2003. Selective reporting bias (no published version of the protocol was found) and other bias (baseline imbalance was reported due to a particularly busy clinic in the intervention group) remain unclear. Blinding was not reported in either trial, however it might not have been possible due to the nature of the study design.

Potential biases in the review process

This review evidence is based on two cluster‐randomised controlled trials identified through an inclusive search strategy. However, we cannot eliminate the possibility that additional published or unpublished trials on this topic exist and were not detected. If any such trials are identified, we will include them in future updates of this review.

Agreements and disagreements with other studies or reviews

Our review confirms that the evidence effectiveness of point‐of‐care syphilis testing varies for different outcomes tested under different settings. While point‐of‐care testing was effective in the Mongolian setting to improve syphilis case detection, treatment rates for both women and their partners, and congenital syphilis, it showed non‐significant intervention effect for increased testing and treatment or perinatal mortality reduction in the rural context of South Africa, although treatment delay was reduced. Generally, evidence from randomised controlled trials on the effectiveness of antenatal syphilis testing for the detection, treatment and management of associated adverse outcomes is sparse, with current knowledge mostly derived from observational studies. No Cochrane review has previously attempted to accumulate evidence on effectiveness. Conclusions from other systematic reviews employing rigorous methods were mixed: for instance, one systematic review concluded that 50% of syphilis‐associated stillbirth and perinatal deaths could be averted if effectiveness in both coverage and screening of available antenatal syphilis testing strategies were improved (Hawkes 2011); while other reviews failed to detect any effect of syphilis screening on preterm birth reduction (Barros 2010), or a high grade of evidence (Menezes 2009).

Figure 1

Study flow diagram.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 2

'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Figure 3

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.1

Comparison 1 On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with RPR or TPHA, Outcome 1 Incidence of congenital syphilis.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.2

Comparison 1 On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with RPR or TPHA, Outcome 2 Proportion of pregnant women tested for syphilis ‐ 1st antenatal visit.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.3

Comparison 1 On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with RPR or TPHA, Outcome 3 Proportion of pregnant women tested for syphilis ‐ 3rd trimester.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.4

Comparison 1 On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with RPR or TPHA, Outcome 4 Adequate treatment.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.5

Comparison 1 On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with RPR or TPHA, Outcome 5 Partner treatment.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.6

Comparison 1 On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with RPR or TPHA, Outcome 6 Syphilis prevalence ‐ 1st antenatal visit.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 1.7

Comparison 1 On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with RPR or TPHA, Outcome 7 Syphilis prevalence ‐ 3rd trimester.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.1

Comparison 2 On‐site syphilis testing with rapid plasma reagin (RPR) test versus routine laboratory testing with RPR, Outcome 1 Perinatal mortality.

Ir a la figura de la revisiónAbrir en una pestaña nueva

Analysis 2.2

Comparison 2 On‐site syphilis testing with rapid plasma reagin (RPR) test versus routine laboratory testing with RPR, Outcome 2 Syphilis treatment.

Ir a la figura de la revisiónAbrir en una pestaña nueva


Study
Myer 2003	Of the women with pregnancy outcome data, 12 women in the intervention arm (2.1%) and 5 women in the control arm (3.1%) reported miscarriages.

Analysis 2.3

Comparison 2 On‐site syphilis testing with rapid plasma reagin (RPR) test versus routine laboratory testing with RPR, Outcome 3 Adverse outcomes.

Ir a la figura de la revisión


Study
Myer 2003	Laboratory confirmed syphilis prevalence: 7.5% for both intervention and control clinics (matched pair design accounted for). Syphilis rates positive on both on‐site and laboratory testing were 5% (246) in the intervention versus 6% (149) laboratory confirmed in the control clinics. RPR titre were also determined based on the data available from 346 women (55%), and 134 women (75%) in the control, where 1:16 or higher titre (median) level was documented among 20% of the women.

Analysis 2.4

Comparison 2 On‐site syphilis testing with rapid plasma reagin (RPR) test versus routine laboratory testing with RPR, Outcome 4 Syphilis prevalence.

Ir a la figura de la revisión

Study

Myer 2003

On‐site syphilis testing significantly reduced treatment delays in South Africa: women in the intervention clinics completed syphilis treatment

on average 16 days sooner compared to the women attending control clinics (95% CI 11 to 21; P= < 0.001).

Analysis 2.5

Comparison 2 On‐site syphilis testing with rapid plasma reagin (RPR) test versus routine laboratory testing with RPR, Outcome 5 Treatment delay.

Ir a la figura de la revisión

Table 1. Munkhuu 2009 ‐ Associations between main outcomes and the intervention analysed by multilevel logistic model

Outcome	Intervention n/N (%)	Control n/N (%)	Adjusted OR*	P
Congenital syphilis	1/3632 (0.03)	15/3564 (0.42)	0.09	0.022
Coverage at 1st antenatal visit	3849/3850 (99.9)	3065/3850 (79.6)	989.84	< 0.001
Coverage at 3rd trimester	3670/3683 (97.7)	2357/3796 (62.1)	617.88	< 0.001
Syphilis case at 1st visit	73/3849 (1.9)	27/3065 (0.9)	2.45	< 0.001
Syphilis case at 3rd trimester	20/3670 (0.5)	2/2357 (0.08)	6.27	0.013
Adequate treatment	92/93 (98.9)	26/29 (89.6)	10.44	0.05
Partner treatment	88/93 (94.6)	16/29 (55.2)	18.17	< 0.001
* Adjusted for significant independent variables such as gestational age at 1st antenatal visit and multiple sexual partners with a random effect for antenatal clinics. OR: odds ratio.

Table 1. Munkhuu 2009 ‐ Associations between main outcomes and the intervention analysed by multilevel logistic model

Ir a la tabla de la revisión

Comparison 1. On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with RPR or TPHA

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Incidence of congenital syphilis Show forest plot	1		Odds Ratio (Fixed, 95% CI)	0.09 [0.01, 0.71]

2 Proportion of pregnant women tested for syphilis ‐ 1st antenatal visit Show forest plot	1		Odds Ratio (Fixed, 95% CI)	989.80 [16.27, 60233.05]

3 Proportion of pregnant women tested for syphilis ‐ 3rd trimester Show forest plot	1		Odds Ratio (Fixed, 95% CI)	617.88 [13.44, 28399.01]

4 Adequate treatment Show forest plot	1		Odds Ratio (Fixed, 95% CI)	10.44 [1.00, 108.99]

5 Partner treatment Show forest plot	1		Odds Ratio (Fixed, 95% CI)	18.17 [3.23, 102.20]

6 Syphilis prevalence ‐ 1st antenatal visit Show forest plot	1		Odds Ratio (Fixed, 95% CI)	2.45 [1.44, 4.18]

7 Syphilis prevalence ‐ 3rd trimester Show forest plot	1		Odds Ratio (Fixed, 95% CI)	6.27 [1.47, 26.69]

Comparison 1. On‐site syphilis screening test using rapid treponemal tests versus conventional laboratory testing with RPR or TPHA

Ir a la tabla de la revisión

Comparison 2. On‐site syphilis testing with rapid plasma reagin (RPR) test versus routine laboratory testing with RPR

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Perinatal mortality Show forest plot	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2 Syphilis treatment Show forest plot	1		Odds Ratio (M‐H, Fixed, 95% CI)	Totals not selected

2.1 Adequate treatment	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
2.2 Inadequate treatment	1		Odds Ratio (M‐H, Fixed, 95% CI)	0.0 [0.0, 0.0]
3 Adverse outcomes Show forest plot			Other data	No numeric data

4 Syphilis prevalence Show forest plot			Other data	No numeric data

5 Treatment delay Show forest plot			Other data	No numeric data

Comparison 2. On‐site syphilis testing with rapid plasma reagin (RPR) test versus routine laboratory testing with RPR

Ir a la tabla de la revisión

	Idioma de la Revisión Cochrane Escoja su idioma de preferencia para las revisiones Cochrane y otros contenidos. Las secciones sin una traducción aparecerán en inglés.

	Idioma de la web Escoja su idioma de preferencia para la web de la Biblioteca Cochrane.

Idioma de la Revisión Cochrane

Idioma de la web

Resumen

Antecedentes

Objetivos

Métodos de búsqueda

Criterios de selección

Obtención y análisis de los datos

Resultados principales

Conclusiones de los autores

PICO

PICO

Population

Intervention

Comparison

Outcome

Resumen en términos sencillos

Pruebas de detección prenatales para la prevención y el tratamiento de la sífilis

Resumen visual

Authors' conclusions

Implications for practice

Implications for research

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Economic data for the use of healthcare resources

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

(1) Random sequence generation (checking for possible selection bias)

(2) Allocation concealment (checking for possible selection bias)

(3.1) Blinding of participants and personnel (checking for possible performance bias)

(3.2) Blinding of outcome assessment (checking for possible detection bias)

(4) Incomplete outcome data (checking for possible attrition bias due to the amount, nature and handling of incomplete outcome data)

(5) Selective reporting (checking for reporting bias)

(6) Other bias (checking for bias due to problems not covered by (1) to (5) above)

(7) Overall risk of bias

Measures of treatment effect

Dichotomous data

Continuous data

Unit of analysis issues

Cluster‐randomised trials

Trials with more than two treatment groups

Cross‐over trials

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

Included studies

Design

Sample size

Setting

Participants

Interventions

Outcomes

Excluded studies

Risk of bias in included studies

Allocation

Generation of sequence

Allocation concealment