To estimate the prevalence of Chagas disease in pregnant women and the risk of congenital transmission of Trypanosoma cruzi infection in Brazil, through a systematic review and meta-analysis.
To estimate the prevalence of Chagas disease in pregnant women and the risk of congenital transmission of Trypanosoma cruzi infection in Brazil, through a systematic review and meta-analysis.
We searched electronic databases, grey literature and reference lists of included publications to identify epidemiological studies on the prevalence of Chagas disease in pregnant women and on the congenital transmission rate of T. cruzi infection in Brazil published between January 1980 and June 2013. Pooled estimates and 95% confidence intervals (95% CIs) were calculated using fixed- and random-effects models.
Sixteen articles were included – 12 studies on the prevalence of Chagas disease in pregnant women (549 359 pregnant women) and nine on congenital transmission rates (1687 children born to infected mothers). Prevalence of Chagas disease in pregnant women ranged from 0.1% to 8.5%, and congenital transmission rates from 0% to 5.2%. The pooled prevalence of Chagas disease among pregnant women across studies was 1.1% (95% CI: 0.6–2.0); the pooled congenital transmission rate was 1.7% (95% CI: 0.9–3.1). In 2010, 34 629 pregnant women were estimated to be infected with T. cruzi, and 312–1073 children born (mean: 589 cases) with congenital infection.
Congenital Chagas disease is a neglected public health problem in Brazil. Systematic congenital Chagas disease control programs through routine prenatal screening for T. cruzi should be widely implemented in Brazil's endemic areas, to identify infected pregnant women and newborns at risk of congenital infection.
Estimer la prévalence de la maladie de Chagas chez les femmes enceintes et le risque de transmission congénitale de Trypanosoma cruzi au Brésil, à l'aide d'une revue systématique et méta-analyse.
Nous avons recherché dans des bases de données électroniques, dans la littérature grise et dans les listes de références des publications incluses pour identifier les études épidémiologiques sur la prévalence de la maladie de Chagas chez les femmes enceintes et sur le taux de transmissions congénitales de l'infection à T. cruzi au Brésil, publiées entre janvier 1980 et juin 2013. Les estimations poolées et les intervalles de confiance à 95% (IC95%) ont été calculés en utilisant des modèles à effets aléatoires et fixes.
16 articles ont été inclus: 12 études sur la prévalence de la maladie de Chagas chez les femmes enceintes (549 359 femmes enceintes) et 9 sur les taux de transmissions congénitales (1687 enfants nés de mères infectées). La prévalence de la maladie de Chagas chez les femmes enceintes variait de 0,1% à 8,5% et les taux de transmissions congénitales de 0% à 5,2%. La prévalence poolée de la maladie de Chagas chez les femmes enceintes dans toutes les études était de 1,1% (IC95%: 0,6–2,0), le taux de transmissions congénitales poolé était de 1,7% (IC95%: 0,9 à 3,1). En 2010, il était estimé que 34 629 femmes enceintes étaient infectées par T. cruzi et 312 à 1 073 (moyenne: 589 cas) enfants sont nés avec une infection congénitale.
La maladie de Chagas congénitale est un problème de santé publique négligé au Brésil. Des programmes de contrôle systématique de la maladie de Chagas congénitale par le dépistage prénatal de routine de T. cruzi devraient être largement implémentés dans les zones endémiques du Brésil pour identifier les femmes enceintes infectées et les nouveau-nés à risque d'infection congénitale.
Calcular la prevalencia de la enfermedad de Chagas en mujeres embarazadas y el riesgo de transmisión congénita de la infección por Trypanosoma cruzi en Brasil, mediante una revisión sistemática y meta-análisis.
Hemos realizado una búsqueda en bases de datos electrónicas, literatura gris y las listas de referencias de las publicaciones incluidas, para identificar estudios epidemiológicos sobre la prevalencia de la enfermedad de Chagas en mujeres embarazadas y sobre las tasas de transmisión congénita de la infección por T. cruzi en Brasil, publicados entre Enero de 1980 y Junio del 2013. Los cálculos de datos agrupados en intervalos de confianza del 95% (IC 95%) se realizaron utilizando modelos de efectos fijos y aleatorios.
Se incluyeron 16 artículos – 12 estudios sobre la prevalencia de la enfermedad de Chagas en mujeres embarazadas (549,359 mujeres embarazadas) y 9 con tasas de transmisión congénita (1,687 niños nacidos de madres infectadas). La prevalencia de la enfermedad de Chagas entre mujeres embarazadas estaba entre un 0.1% y un 8.5%, y las tasas de transmisión congénita entre el 0% y el 5.2%. La prevalencia de la enfermedad de Chagas en mujeres embarazadas, calculada a partir de los datos agrupados de todos los estudios era del 1.1% (IC 95%: 0.6–2.0); la tasa de transmisión congénita a partir de datos agrupados era del 1.7% (IC 95%: 0.9–3.1). En el 2010, se calculó que 34,629 mujeres embarazadas estaban infectadas con T. cruzi, y entre 312 a 1,073 bebés nacidos (media: 589 casos) con infección congénita.
La enfermedad de Chagas congénita es un problema de salud pública olvidado en Brasil. Los programas de control sistemático de la enfermedad de Chagas congénita mediante el cribado prenatal rutinario de T. cruzi deberían implementarse ampliamente en áreas endémicas del Brasil, con el fin de identificar a las mujeres embarazadas y a los recién nacidos en riesgo de infección congénita.
Chagas disease (American trypanosomiasis), caused by the protozoan parasite Trypanosoma cruzi, is endemic in 21 Latin American countries, where an estimated 8–10 million people are infected (PAHO 2006; WHO 2012). In Brazil, there are about 2.9–7.2 million people living with Chagas disease (Martins-Melo et al. 2014), causing about 6000 deaths annually (Martins-Melo et al. 2012a,b,c). Chagas disease has become a global emerging problem due to large-scale international migration of Latin Americans to non-endemic countries, particularly to the USA, Canada, Europe, Australia and Japan (Schmunis & Yadon 2010).
Trypanosoma cruzi is transmitted by infected faeces of blood-sucking triatomine bugs, blood transfusion, organ transplantation, consumption of contaminated food or drink and from mother to child (Rassi et al. 2010; WHO 2012). With the control of vector and blood-borne transmission in most endemic areas, congenital transmission has increased in importance (Gurtler et al. 2003; Howard et al. 2014).
Prevalence of T. cruzi infection in pregnant women ranges from 1% to 40%, (Torrico et al. 2004, 2005; Salas et al. 2007; Carlier & Truyens 2010), and about 1.8 million women of childbearing age are infected in Latin America (PAHO 2006). Congenital transmission rates of T. cruzi infection range from 0% to 28.6% (Howard et al. 2014). Recent estimates indicate that about 14 400 newborns are congenitally infected annually in Latin America (PAHO 2006), and 2000 in North America (Buekens et al. 2008). The transmission can be repeated at each pregnancy and during the entire fertile period of a woman's life (Carlier & Torrico 2003). While most cases are asymptomatic, congenital T. cruzi infection may result in preterm birth, low birth weight, stillbirths and clinical manifestations of the disease at birth (Bittencourt 1976, 1992; Carlier & Torrico 2003; Torrico et al. 2006). As congenital transmission cannot be prevented, early diagnosis and treatment of congenital cases are high priorities in congenital Chagas disease control programs (Bern et al. 2009; Carlier et al. 2011).
There are no systematic estimates of burden of Chagas disease in pregnant women and transmission of congenital infection for most endemic areas (Gurtler et al. 2003). In Brazil, prevalence data on Chagas disease in pregnant women are limited, and risk of congenital transmission in most endemic and non-endemic areas is unknown. This study aimed to estimate the prevalence of Chagas disease in pregnant women and the risk of congenital transmission of T. cruzi infection in Brazil, through a systematic review and meta-analysis.
We conducted a systematic literature review to identify relevant publications about prevalence of Chagas disease in pregnant women and congenital transmission of T. cruzi infection in Brazil. We searched the electronic databases PubMed, Web of Science, Scopus, Science Direct, LILACS and SciELO for reports published between 1 January 1980 and 30 June 2013. The search terms used were as follows: ‘Chagas disease’, Trypanosoma cruzi, ‘pregnant women’, ‘congenital transmission’, ‘vertical transmission’, ‘prevalence’ and ‘Brazil’. There were no restrictions on language of publication. Reference lists of review articles and other documents were hand-searched for additional relevant studies. Furthermore, we screened journals not indexed in electronic databases, abstracts from congresses of tropical and infectious disease societies, reports of control programs and grey literature. Our analyses were performed in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines (Liberati et al. 2009).
We included all types of observational studies from Brazil that reported prevalence of Chagas disease in pregnant women and/or congenital transmission rate of T. cruzi infection. We further included only studies in which the diagnosis of infection by T. cruzi in pregnant women was carried out by combination of two or more conventional serological tests. For children born from infected mothers, we included studies if diagnosis of congenital infection was made using one or combination of the following diagnostic methods: direct parasitological examination; serological tests performed on children after 6–9 months of age; polymerase chain reaction (PCR), xenodiagnosis and hemoculture (Howard et al. 2014). Studies with no detailed description of sampling, case reports, anecdotal evidence and reviews were excluded. Duplicated studies using the same data were identified, and the most recent and/or more complete data were included.
From each article included, the following data were extracted: author names, publication year, study period, study sites, sampling method and sample size, study setting (population/community based or hospital based), age group, number of T. cruzi cases in pregnant women or children, diagnostic methods used, follow-up/time of diagnosis of congenital infection, origin of the blood sample for diagnostic screening (venous blood, umbilical cord and peripheral/capillary blood dried on filter paper), and prevalence of disease in pregnant women and/or congenital transmission rate (or data needed to calculate them) with their respective 95% confidence intervals (95% CI). The 95% CI were directly extracted from articles, when available, or calculated using an exact binomial method. Searches and data extraction were performed by a single investigator (FRM). Disagreements were resolved by consensus between the authors or through consultation with the corresponding author of relevant studies.
Pooled estimates with their 95% CI were calculated using the fixed- and random-effects models (Hedges & Vevea 1998). Prevalence of Chagas disease in pregnant women was defined as the number of cases of pregnant women with T. cruzi infection divided by the total of pregnant women evaluated. Congenital transmission rate was defined as the number of congenitally infected children divided by number of children born to infected mothers. In order to minimise bias caused by small sample size, in the quantitative analysis we excluded studies with a sample size <10 children born to infected mothers. Heterogeneity between studies was assessed using the Cochran's Q test (reported as χ2 and P values) and I2 statistic. Heterogeneity was considered statistically significant at a P value <0.10. I2 values of >25%, 50% and 75% show low, moderate and high heterogeneity, respectively (Higgins & Thompson 2002; Higgins et al. 2003). When significant heterogeneity was absent, the pooled estimates were calculated using a fixed-effects model; otherwise, a random-effects model was used. Publication bias was assessed using Begg's funnel plots and Egger's regression test (Egger et al. 1997). If significant publication bias was found, the Duvall and Tweedie's trim and fill method was performed by adding studies that appeared to be hypothetically missing (Duval & Tweedie 2000). Adjusted pooled estimates were computed after the addition of potential missing studies.
Subgroup analyses were performed to investigate potential sources of heterogeneity among studies and included the following variables: Brazil's five geographical regions, maternal age group (<20, 20–29 and >30 years), study period (1980–1989, 1990–1999 and after 2000), sample size, study type, urban/rural area, diagnostic method and origin of the blood sample for diagnostic screening. If a study did not report the year of data collection, the publication year was used.
We also estimated the expected number of women of childbearing age, pregnant women and children born with T. cruzi infection using data on the general population and women of childbearing age (10–49 years) from the 2010 National Demographic Census of the Brazilian Institute of Geography and Statistics (Instituto Brasileiro de Geografia e Estatística – IBGE; http://tabnet.datasus.gov.br/cgi/deftohtm.exe?ibge/cnv/popuf.def); live births in Brazil in 2010 obtained from the Information System on Live Births (Sistema de Informação sobre Nascidos Vivos – SINASC; http://tabnet.datasus.gov.br/cgi/deftohtm.exe?sinasc/cnv/nvuf.def); and the estimated birth rate in 2010. The number of pregnant women was estimated using the total population multiplied by the birth rate, adding a correction of 10% due to losses resulting from abortion, stillbirths and underreporting (http://nutricao.saude.gov.br/calculo_bvg.php). To estimate the potential number of pregnant women and women of childbearing age with T. cruzi infection in Brazil and regions in 2010, we applied the result of the estimated pooled prevalence of Chagas disease in pregnant women at national level, extrapolating to the women of childbearing age population and estimated pregnant women. To estimate the expected number of children born with congenital infection in 2010, we used the estimated pooled congenital transmission rate at national level. Potential geographic and temporal variations and effects of the parasite were not considered for calculation. In addition, we assumed that congenital transmission rate was not modified by mother's age, as most of women of childbearing age are in the indeterminate or chronic phases, and levels of parasitemia are low.
Data were analysed using Stata software, version 11.2 (StataCorp LP, College Station, TX, USA) and Comprehensive Meta-Analysis software, version 2.0 (Biostat, Englewood, NJ, USA). Thematic maps presenting the study sites and prevalence/transmission rate estimates for Brazilian states were created using ArcGIS software, version 9.3 (Environmental Systems Research Institute, Redlands, CA, USA).
We identified a total of 621 records. After removal of duplicates and initial screening, 33 articles were reviewed in full (Figure 1). After exclusion of not eligible articles and inclusion of additional articles identified by manual search of reference lists, 16 publications were eligible for analysis (Bittencourt 1984; Arteaga-Fernandez et al. 1987; Vaz et al. 1990; Medina-Lopes 1992; Mota 1993; Santos et al. 1995; Gontijo et al. 1998a,b, 2009; Nisida et al. 1999; Reiche et al. 2000; Rassi et al. 2004; Figueiró-Filho et al. 2007; Botelho et al. 2008; Araújo et al. 2009; Gomes Filho et al. 2009). These included 21 studies: 12 on prevalence of Chagas disease in pregnant women (Bittencourt 1984; Vaz et al. 1990; Santos et al. 1995; Gontijo et al. 1998a,b, 2009; Reiche et al. 2000; Figueiró-Filho et al. 2007; Botelho et al. 2008; Araújo et al. 2009; Gomes Filho et al. 2009) and nine on congenital transmission of T. cruzi infection (Bittencourt 1984; Arteaga-Fernandez et al. 1987; Medina-Lopes 1992; Mota 1993; Gontijo et al. 1998a, 2009; Nisida et al. 1999; Rassi et al. 2004; Araújo et al. 2009) (Figure 1).
Characteristics of studies for both target groups are summarised in Tables 1 and 2. The 12 studies on prevalence of Chagas disease in pregnant women included a total of 549 359 pregnant women, with 3033 cases of T. cruzi infection. Sample size ranged from 351 to 272 335 (mean: 45 780; median: 2141) and observed prevalence of Chagas disease in pregnant women from 0.1% to 8.5% among studies. The studies were conducted between 1981 and 2007 in seven Brazilian states. Most data were collected in the south-east region (41.7%), in the 1980s and after 2000 (41.7%) (Table 1).
|Authors/publication year||Study period||Study site (locality, city or municipality)||State||Region||Study setting/sample type||Area||Serological test||Blood sample origin (screening)||Age group (years)||Sample size||Cases||Prevalence (%)a||95% CI|
|Bittencourt (1984)||January 1981 to August 1982||Salvador (02 maternity hospitals)||Bahia||Northeast||Hospital based||Urban/rural||ELISAb; IFAb||Venous blood||NA||2651||226||8.5||7.5–9.6|
|Vaz et al. (1990)||April to October 1988||São Paulo (08 health centres)||São Paulo||Southeast||Healthcare service||Urban||IFAb; IHAb||Venous blood||14–46||481||14||2.9||1.6–4.8|
|Santos et al. (1995)||November 1990 to June 1991||Salvador||Bahia||Northeast||Population/Community based||Urban||IFA; IHAb||Venous blood||<20–>30||1024||24||2.3||1.5–3.5|
|Gontijo et al. (1998a)c||March to May 1997||Statewide (407 municipalities)||Minas Gerais||Southeast||Population/Community based||Urban/rural||ELISAb; IFA; IHA||Capillary blood dried on filter paper||NA||18 414||175||0.9 (0.1–33.0 among municipalities)||0.8–1.1|
|Gontijo et al. (1998b) (1)||July 1995 to June 1997||Belo Horizonte (maternity hospital)||Minas Gerais||Southeast||Hospital based||Urban/rural||IFAb; IHAb||Umbilical cord blood||NA||1631||23||1.4||0.9–2.1|
|Gontijo et al. (1998b) (2)||July 1995 to June 1997||Uberaba (maternity hospital)||Minas Gerais||Southeast||Hospital based||Urban/rural||IFAb; IHAb||Umbilical cord blood||NA||1266||54||4.3||3.2–5.5|
|Reiche et al. (2000)||June 1996 to June 1998||Londrina (ambulatory obstetric)||Paraná||South||Hospital based/Healthcare service||Urban/rural||ELISAb; IFAb||Venous blood||12–58||1164||11||0.9||0.5–1.7|
|Figueiró-Filho et al. (2007)||November 2002 to October 2003||Statewide (78 municipalities)||Mato Grosso do Sul||Central-West||Population/Community based||Urban/rural||ELISAb; IFA; IHA||Peripheral blood on filter paper||11–49||32 512||43||0.1||0.0–0.2|
|Botelho et al. (2008)||January 2004 to December 2007||Statewide (78 municipalities)||Mato Grosso do Sul||Central-West||Population/Community based||Urban/rural||ELISAb; IFA; IHA||Peripheral blood on filter paper||9–54||153 857||512||0.3||0.30–0.36|
|Araújo et al. (2009)||2004||Pelotas (03 hospitals)||Rio Grande do Sul||South||Hospital based||Urban/rural||ELISAb; IFA; IHA; TESA-Blot||Umbilical cord blood||NA||351||1||0.3||0.0–1.6|
|Gomes Filho et al. (2009)||September 2003 to May 2008||Statewide (245 municipalities)||Goiás||Central-West||Population/Community based||Urban/rural||ELISAb; IFA; IHA||Peripheral blood on filter paper||NA||272 335||1418||0.5||0.49–0.55|
|Gontijo et al. (2009)c||August 2005 to October 2006||Statewide (853 municipalities)||Minas Gerais||Southeast||Population/Community based||Urban/rural||ELISAb; IFA; IHA||Capillary blood dried on filter paper||NA||63 673||532||0.8 (0.1–23.0 among municipalities)||0.77–0.91|
|Authors/publication year||Study period||Study site (locality, city or municipality)||State||Region||Study setting/sample type||Area||Diagnostic method||Blood sample origin (screening)||Follow-up time/diagnosis||Sample size||Cases||Transmission rate (%)a||95% CI|
|Bittencourt 1984||January 1981 to August 1982||Salvador (02 maternity hospitals)||Bahia||Northeast||Hospital-based||Urban/rural||Parasitological (microhematocrit); xenodiagnosis||Peripheral blood (heel puncture)||Birth||186||3||1.6||0.3–4.6|
|Arteaga-Fernandes et al. 1987||NA||São Paulo (01 hospital)||São Paulo||Southeast||Hospital-based||Urban||Serological (CFR, IHA, IFA); xenodiagnosis||NA||Birth, >6 months||129||4||3.1||0.8–7.7|
|Medina-Lopes 1992||NA||Brasília||Distrito Federal||Central-West||Hospital-based||NA||Parasitological (Strout method); xenodiagnosis||Umbilical cord blood/peripheral blood||0–5 days, until 4 years||178||2||1.1||0.1–4.0|
|Mota 1993||1983–1990||Belo Horizonte (01 hospital)||Minas Gerais||Southeast||Hospital-based||NA||Parasitological; serological (CFR, IFA), xenodiagnosis||NA||Birth to 5 years||64||2||3.1||0.4–10.8|
|Gontijo et al. 1998a||March to May 1997||Statewide (407 municipalities)||Minas Gerais||Southeast||Population/Community-based||Urban/rural||Serological (ELISA/IFA/IHA)||Capillary blood dried on filter paper||5–7 days, 6–12 months||175||3||1.7||0.3–4.9|
|Nisida et al. 1999||NA||São Paulo (03 obstetric clinics)||São Paulo||Southeast||Hospital-based||Urban/rural||Parasitological (microhematocrit, QBC); serological (ELISA, TESA-blot); xenodiagnosis; immunohistochemistry||Venous blood/umbilical cord blood||Birth, 1 month, >6 months||58||3||5.2||1.1–14.4|
|Rassi et al. 2004||1975–2004||Goiânia||Goiás||Central-West||Hospital-based||NA||Serological (CFR, IHA, IFA, ELISA); xenodiagnosis||Venous blood||Retrospective (3 months –40 years old)a||278||2||0.7||0.1–2.6|
|Araújo et al. 2009||2004||Pelotas (03 hospitals)||Rio Grande do Sul||South||Hospital-based||Urban/Rural||Serological (ELISA, IFA, IHÁ, TESA-Blot)||Umbilical cord blood||Birth, >6 months||1||0||0.0||0.0–97.5|
|Gontijo et al. 2009||August 2005 to October 2006||Statewide (853 municipalities)||Minas Gerais||Southeast||Population/Community-based||Urban/Rural||Serological (ELISA, IHA, IFA, ELISA), PCR, haemoculture||Capillary blood dried on filter paper||5–7 days, 3 months, 6–9 months||541||1||0.2||0.0–1.0|
Figure 2(a) shows the geographical distribution of study sites for studies on prevalence in pregnant women. These were concentrated in the Southeast and Central-West regions. Some states (such as Minas Gerais, Goiás and Mato Grosso do Sul) contain a large number of study sites, while there are no data available from many other states (Figure 2a). Not a single study was identified from the North region. Available data for municipalities in the states of Goiás (all 245 municipalities) (Gomes Filho et al. 2009) and Mato Grosso do Sul (all 78 municipalities) (Figueiró-Filho et al. 2007; Botelho et al. 2008) are derived from prenatal screening programs, while most of the data of Minas Gerais were derived from surveys based on neonatal screening conducted on state level in 1997 (407 municipalities) (Gontijo et al. 1998a) and 2005 (all 853 municipalities) (Gontijo et al. 2009) (Figure 2a, Table 1).
The nine selected studies on congenital transmission of T. cruzi infection included a total of 1610 children born to infected mothers and 20 cases of congenital transmission. Congenital transmission rates ranged from 0% to 5.2% among studies (Table 2). Sample size ranged from 1 to 541 children born to infected mothers (mean: 179; median: 175), while the number of diagnosed congenital cases ranged from 0 to 4 children. Studies were conducted between 1975 and 2006 in six Brazilian states. Most data were collected in the south east region (55.6%), in the period 1980–1989 (44.4%) (Table 2). Most samples were hospital based (77.8%), using peripheral blood for diagnostic screening (33.3%) (Table 2). Most study sites were in the south east region, especially in the state of Minas Gerais (Figure 2b, Table 2).
Pooled prevalence of Chagas disease in pregnant women from the 12 studies was 1.1% (95% CI: 0.6–2.0) (Table 3), with high level of heterogeneity (I2: 95%, P < 0.001), which remained even after analysis of estimates by subgroups (Table 3). There was no evidence of significant publication bias (Egger's test, P = 0.15). Prevalence estimates for the study period ranged from 5.2% (95% CI: 1.8–14.2) in 1980–1989, 1.7% (95% CI: 0.8–3.3) in 1990–1999, to 0.4% (95% CI: 0.2–0.6) after 2000. The highest estimated regional prevalence was observed in the north east region (4.6%, 95% CI: 1.3–15.3) (Table 3). Among the states with available data, the highest estimated prevalence was observed in Bahia (4.6%, 95% CI: 1.3–15.3) and São Paulo states (2.9%, 95% CI: 1.7–4.9) (Figure 3a).
|Characteristics||Number of studies||Rangea||Pooled Chagas disease prevalence||Heterogeneity|
|Case||Pregnant women||Prevalence (%)||95% IC||I2 (%)||P-value (Cochran's Q)||Model|
|Overall prevalence||12||0.1–8.5||3033||549 359||1.1||0.6–2.0||99.5||<0.001||Random|
|Age group (years)b|
|Region of Brazil|
|>10 000||05||0.1–0.9||2680||540 791||0.5||0.3–0.7||98.9||<0.001||Random|
|Study setting/sample type|
|Population/Community based||06||0.1–2.3||2704||541 815||0.6||0.4–0.9||98.8||<0.001||Random|
|Blood sample (screening)|
|Peripheral blood/capillary blood dried on filter paper||05||0.1–0.9||2680||540 791||0.5||0.3–0.7||98.9||<0.001||Random|
|Umbilical cord blood||03||0.3–4.3||78||3248||1.7||0.6–4.9||92.3||<0.001||Random|
Information on the age distribution was available in four studies. The highest prevalence was found in pregnant women >30 years (1.4%, 95% CI: 0.4–4.7) (Table 3).
In quantitative analysis, eight studies were included in the meta-analysis. One study was excluded due to sample size <10 children (Araújo et al. 2009). Pooled congenital T. cruzi transmission rate among studies was 1.7% (95% CI: 0.9–3.1), with low degree and significant heterogeneity (I2: 44.5%, P = 0.08) (Table 4). Evidence of significant publication bias was found (Egger's test, P = 0.02). Duval and Tweedie's trim and fill method (random effect) imputed one hypothetically missing study, with an adjusted congenital transmission rate of 2.0% (95% CI: 1.0–3.9). Congenital transmission estimates for the study period varied between 2.0% (95% CI: 1.0–3.6) in 1980–1989, 2.3% (95% CI: 1.2–4.6) in 1990–1999 and 0.2% (95% CI: 0.0–1.3) after 2000 (Table 4). The highest estimation of regional congenital transmission rate was observed in the Southeast region (2.1%, 95% CI: 0.9–5.0) (Table 4). The states of São Paulo (3.9%; 95% CI: 1.8–7.9) and Minas Gerais (1.2%; 95% CI: 0.3–4.7) had the highest estimates (Figure 3b). Estimates were higher in studies performed in urban areas (3.1%, 95% CI: 1.2–8.0), hospital-based samples (2.2%, 95% CI: 1.4–3.6) and sample sizes of <100 children born to infected mothers (3.5%, 95% CI: 1.8–9.8) (Table 4).
|Characteristics||Number of studies||Rangea||Pooled congenital T. cruzi transmission rate||Heterogeneity|
|Case||Sample size||Infection rate (%)||95% IC||I2 (%)||P-value (Cochran's Q)||Model|
|Overall transmission rate||08||0.2–5.2||20||1609||1.7||0.9–3.1||44.5||0.082||Random|
|Region of Brazil|
|Study setting/sample type|
|Blood sample (screening)|
|Peripheral blood/capillary blood dried on filter paper||03||0.2–1.7||07||902||1.2||0.6–2.5||52.2||0.124||Fixed|
|Umbilical cord blood/peripheral blood||01||–||02||178||1.1||0.3–4.1||0.0||1.0||–|
|Umbilical cord blood/venous blood||01||–||03||58||5.2||1.1–14.4||0.0||1.0||–|
In 2010, the estimated number of women of childbearing age infected with T. cruzi in Brazil was 683 217 (95% CI: 372 664–1 242 213), with an expected number of 34 629 infected pregnant women (Table 5). Considering the estimated pooled congenital transmission rate, there was an expected number of 589 (95% CI: 312–1073) newborns with congenitally acquired T. cruzi infection in Brazil in 2010 (Table 5). The incidence was estimated at 18.7 congenital cases per 100 000 live births (95% CI: 9.9–34.1).
|Region||Number of women of childbearing age (10–49 year-olds)a||Estimated number of women of childbearing age infected with T. cruzi (95% IC)b||Number of live births recordedc||Estimated numbers of pregnant women infected/newborns from infected mothersb||Expected number of newborns infected with T. cruzid|
|Lower limit||Middle limit||Upper limit|
|North||5 206 854||57 275 (31 241–104 137)||306 422||3708||33||63||115|
|North-east||17 476 138||192 238 (104 857–349 523)||841 160||10 178||92||173||316|
|South-east||2 597 7400||285 751 (155 864–519 548)||1 123 593||13 595||122||231||421|
|South||8 735 418||96 090 (52 413–174 708)||369 905||4476||40||76||139|
|Central-west||4 714 827||51 863 (28 289–94 297)||220 788||2672||24||45||83|
|Brazil||62 110 637||683 217 (372 664–1 242 213)||2 861 868||34 629||312||589||1073|
This systematic review provides a comprehensive overview of the prevalence of Chagas disease in pregnant women and of the congenital transmission rate of T. cruzi infection in Brazil, over a period of more than three decades. The data indicate an increasing urbanisation of the disease, higher prevalence in women with more advanced age, and considerable regional and temporal variations of congenital transmission risk.
With the development of national programs in recent decades, focusing on systematic entomological surveillance and screening of blood donors, control of vector and blood-borne transmission was achieved in most endemic areas (Gurtler et al. 2003; Ramos Jr et al. 2010; Martins-Melo et al. 2012a). Consequently, congenital transmission has an increasing epidemiological importance (Reiche et al. 1996; Brazilian Ministry of Health 2005; Ostermayer et al. 2005; Andrade & Gontijo 2008; Bern et al. 2009). In Argentina, it has been estimated that congenital cases are at least 10 times more frequent than acute cases by vector transmission (Gurtler et al. 2003).
In this context, the higher prevalence in pregnant women at more advanced age (>30 years) reflects the trend of ageing of patients with chronic Chagas disease, after successful control of vector and blood-borne transmission in the last decades (Martins-Melo et al. 2012a,c, 2014). A national seroprevalence survey conducted in Brazil's rural areas (2001–2008), including about 105 000 children aged 0–5 years (Ostermayer et al. 2011), has found a prevalence of 0.03% (32 confirmed cases) – 0.02% (20) with probable congenital transmission, mostly from Rio Grande do Sul state, and only 0.01% (11) with probable vector transmission (Ostermayer et al. 2011). Currently, about 60–90% of Brazil's population infected with T. cruzi is living in urban areas (Dias 2007; Martins-Melo et al. 2014). As a result, congenital transmission, as confirmed by our study, has been an increasing problem of urban centres.
Latin American urbanisation and migratory movements from endemic to non-endemic countries, including women of reproductive age, also have changed transmission dynamics (Schmunis & Yadon 2010; Dias 2013). Migration of women of childbearing age to non-endemic countries is partially responsible for spread of Chagas disease worldwide through congenital transmission (Schmunis & Yadon 2010). For example, congenital cases in non-endemic countries have been reported from Spain, USA Sweden, Switzerland and Japan (Pehrson et al. 1981; Jackson et al. 2009; CDC 2012; Merino et al. 2013; Imai et al. 2014).
It can be assumed that congenital transmission will continue being a public health problem in Latin American countries for years (and in countries receiving migrants from endemic areas), despite the decline of prevalence of Chagas disease over the past decades. As long as a significant proportion of women of childbearing age are still seropositive for T. cruzi, congenital transmission will occur (Gontijo et al. 2009; Raimundo et al. 2010).
The prevalence of Chagas disease in pregnant women among studies ranged from 0.1% (Mato Grosso do Sul state, 2002–2003) (Figueiró-Filho et al. 2007) to 8.5% (Salvador city, 1981–1982) (Bittencourt 1984), and congenital transmission rate from 0% (Pelotas city, 2004) (Araújo et al. 2009) to 5.2% (São Paulo city, 1999) (Nisida et al. 1999). The observed and pooled congenital transmission rates are lower than in other Latin American endemic countries: 0.75–17% in Argentina; 3.4–11% in Bolivia; 0.49–19% in Chile; 1.44–10% in Paraguay; and 0.13–1.57% in Uruguay (Yadon & Schmunis 2009; Howard et al. 2014), and non-endemic countries (children born to immigrant Latin American pregnant women): 0–28.6% in Spain, and 25% in Switzerland (Howard et al. 2014). These variations may be caused by the degree of parasitemia of the infected mother, differing parasite strains, exposure to vector-borne re-infections during pregnancy, occurrence of acute form of the disease during pregnancy, different origins of the population studied (Latin American immigrants in non-endemic countries), maternal age, as well as different diagnostic techniques and study designs (Bittencourt 1992; Andrade et al. 1994; Gurtler et al. 2003; Hermann et al. 2004; Moretti et al. 2005; Torrico et al. 2006; Salas et al. 2007; Yadon & Schmunis 2009; Carlier & Truyens 2010; Howard et al. 2014).
The Brazilian Consensus on Chagas disease and the World Health Organization (WHO) reinforce the need for implementation of routine screening programs for Chagas disease in pregnant women and newborns at risk (Brazilian Ministry of Health 2005; Gontijo et al. 2009; Carlier et al. 2011). In fact, screening pregnant women at risk is a cost-effective intervention, not only for early diagnosis of congenital Chagas disease, but also for improvement in quality of life and prognosis for patients (Billot et al. 2005; Sicuri et al. 2011; Alonso-Vega et al. 2013). High cure rates of specific treatment of congenital cases justify the efforts needed to detect infection by T. cruzi in mothers and their newborns (Gontijo et al. 2009). However, these recommendations are neglected in most endemic and non-endemic countries receiving Latin American immigrants (Buekens et al. 2008; Pinto et al. 2011; Imai et al. 2014). In Brazil, only the states of Goiás and Mato Grosso do Sul have established systematic prenatal screening programs in all municipalities (Figueiró-Filho et al. 2007; Botelho et al. 2008; Gomes Filho et al. 2009). Specific antiparasitic treatment is mandatory in all congenital cases, but is not recommended during pregnancy, due to toxicity and teratogenic risks of available drugs (benznidazole and nifurtimox) (Brazilian Ministry of Health 2005; Gontijo et al. 2009; Carlier et al. 2011).
There are several factors limiting the estimation of prevalence of Chagas disease in pregnant women and congenital transmission rate in Brazil (Martins-Melo et al. 2012a, 2014). First, only acute forms of the disease are subject to compulsory notification, and contemporary data on the prevalence of Chagas disease on national or regional level are scarce (Camargo et al. 1984; Martins-Melo et al. 2014). Second, systematic control programs on congenital Chagas transmission through prenatal and neonatal screening are not routinely established at national level, nor in most endemic areas (Gurtler et al. 2003). Lastly, as most children with congenital infection are asymptomatic (Oliveira et al. 2010), the number of congenital cases is still underestimated.
There were only a small number of relevant studies included, with a clear geographic focus. Consequently, the included studies may reflect the situation in some areas or regions in Brazil, but due to the lack of data from other regions not the reality of the entire country. Furthermore, the studies were conducted during a period of about 30 years (1980–2013). This long time period was necessary because of limited data availability in some areas of Brazil. Thus, prevalence figures should be interpreted and compared with care, especially in those areas where due to the lack of more recent studies, current data were not available. Other limitations of our analysis refer to differences in diagnostic techniques used, different settings and populations, case definitions and methods used to detect infection in children born to infected mothers.
We assumed that the estimated prevalence can be generalised to the general population of women of childbearing age and that the prevalence and transmission rates did not vary between regions, period and mother's age. While there is evidence that the prevalence of Chagas disease was different over decades and among regions, this assumption was applied for the sake of simplicity (Yadon & Schmunis 2009; Martins-Melo et al. 2014). Thus, the numbers presented provide a general overview of the burden of congenital Chagas disease, indicating the importance of this mode of transmission (Yadon & Schmunis 2009).
There is a need to implement public health programs directed to congenital transmission at national and regional levels. Chagas disease in Brazil and some other endemic countries is suffering from the so-called curse of success, where reduction in public health importance also reduces public and political interest and budgets (Massad 2008). Clearly, the potential re-emergence of vector transmission and the importance of congenital transmission in the maintenance of Chagas disease in Brazil are underestimated by policy makers.
Adequate access to health services and social assistance should be guaranteed for the large number of individuals affected by chronic Chagas disease (Ramos et al. 2010; Martins-Melo et al. 2012a). Most infected pregnant women present in the chronic phase of the disease, especially in the indeterminate form. This reinforces the need for the introduction of compulsory notification also of chronic forms of Chagas disease (Martins-Melo et al. 2012a).
In conclusion, our findings show that congenital Chagas disease is an important but neglected health problem in Brazil, with significant regional differences. Systematic implementation of surveillance and control of congenital Chagas disease is needed, including routine screening of infection by T. cruzi in pregnant women and newborns, early case detection, prompt treatment and follow-up of children with congenital infection. Epidemiological data on the extension of maternal and congenital T. cruzi infection are lacking from many endemic areas.
We thank the Fundação Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES/Brazil) for granting a PhD Scholarship to FRM and Master Scholarship to MSL. JH is research fellow at the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq/Brazil).