Introduction

Regional initiatives to control Chagas disease by eliminating blood-borne transmission and domestic vector infestation have had a remarkable impact on the burden of this neglected disease [1]. Estimates for Trypanosoma cruzi infection prevalence have fallen from 16–18 million in 1990 to 8 million in 2005 [2,3]. Nevertheless, a third or more of those infected are reproductive-aged women and girls who can transmit the parasite to offspring during successive pregnancies. In 1990, the vast majority of new T. cruzi infections were vector-borne, but in 2005, 26% of new infections were attributed to mother-to-child transmission; as vector and blood bank control continue to improve, the proportion due to congenital infection is expected to grow [3,4].

A major unanswered question is why reported transmission rates from T. cruzi-infected mothers vary so widely among study populations, ranging from 1% to >15% in published studies [5–14]. Studies have consistently shown that women with high parasite loads during pregnancy are more likely to transmit to their infants, but the factors that govern maternal parasite load are largely unknown [15–17]. Better understanding of this phenomenon could enable prioritization of screening programs to target women most at risk of transmission to their infants.

Congenitally infected infants are usually asymptomatic or have mild symptoms, but they have a 20–30% risk of chronic cardiac and/or gastrointestinal disease decades later [4]. Treatment is well tolerated and highly effective early in life, but many congenital infections go undiagnosed [15,18]. We studied congenital T. cruzi transmission in a cohort of women giving birth in an urban hospital in Bolivia, the country with the highest Chagas disease prevalence [4]. We evaluated factors associated with maternal parasite load during pregnancy and implications for congenital transmission.

Materials and Methods

Results

A total of 596 pregnant women were enrolled from June 1, 2010 to October 31, 2011. Of these, 128 (21.5%) had confirmed T. cruzi infection. Compared to uninfected women, T. cruzi-infected women were older (median (interquartile range [IQR]): 28 (22–34) vs 22 (19–29), p<0.001), had higher parity (median (IQR): 3 (2–5) vs 2 (1–4), p<0.001) and higher percentage of cesarean births (54.7% vs 47.7%; p = 0.16). Infected women were less likely to have any secondary education (58% vs 74%; p = 0.006), more likely to have lived in an infested house (56% vs 25%; p <0.0001) and slightly more likely to have lived in a rural area (64% vs 57%; p = 0.17) than uninfected women.

Congenital transmission occurred from 15 (11.7%) of 128 infected women, including 3 (42%) of 7 women with twin births (p = 0.03 for twin vs singleton births). Three of 3 bichorionic twin births were associated with transmission compared to 0 of 4 monochorionic twin births (p = 0.29). In each twin birth, both infants were infected, bringing the total number of infected babies to 18. Four infants had ambiguous infection status: 2 infants with a single positive result by PCR who were subsequently lost to follow-up and 2 infants treated by the control program at 6 months based on IHA at a titer of 1:128 with ELISA absorbance of 0.2 in one case and negative ELISA result in the other. These mother-baby pairs were excluded from analyses of transmission risk. Among the 18 confirmed infected infants, diagnosis was based on micromethod in 4 (all positive by qPCR and TESA-blot as well) and qPCR in 12 (10 of whom also had positive IgM TESA-blot); two infants had high titer positive serology by two different assays at 6 months of age.

The transmission rate was higher from infected women with no reported vector exposure compared to those who ever lived in an infested house, but this difference did not reach statistical significance (Table 1). There was no significant difference in age between women who transmitted and those who did not. qPCR data were available for 120 infected women, of whom 55 (45.8%) had positive results with median load of 36.5 parasites/mL (IQR 22.2–80.8). Parasite loads were significantly higher among women who transmitted compared to those who did not and showed a strong dose-response relationship. Congenital transmission occurred from 31.3% (9/29), 15.4% (4/26) and 0% (0/62) of women with high, moderate and low parasite load, respectively (χ² for trend 18.2; p<0.0001).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Maternal factors associated with congenital Trypanosoma cruzi transmission risk.

https://doi.org/10.1371/journal.pone.0119527.t001

Infected women without reported vector exposure had significantly higher parasite loads than those who had lived in an infested house (median 26.4 vs 0 parasites/mL; p<0.001) with an inverse relationship between years of living in an infested house and parasite load (Table 2). Parasite loads were lower among women who had ever lived in a house with thatch roof, mud walls or earth floors, materials associated with triatomine infestation. Housing characteristics had high correlations with each other (coefficients 0.79 to 0.86; p values <0.0001) and moderate but significant correlations with history of having lived in an infested house (coefficients 0.31 to 0.37; p values <0.0001). Mothers of twins had higher parasite loads than those with singleton births with borderline statistical significance. There were no significant differences in parasite load by maternal age or rural residence history.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Factors associated with estimated parasite load in blood of T. cruzi-infected mothers at the time of delivery.

https://doi.org/10.1371/journal.pone.0119527.t002

Discussion

In our data, as in previous studies [15–17], maternal T. cruzi load strongly predicted congenital transmission, with a remarkable dose-response relationship. However, ours is the first study to link absence of vector exposure to increased maternal parasite load, and by implication, congenital transmission risk. A study in Argentina demonstrated the highest transmission rate in women from never-infested areas, with a lower rate in those from previously infested areas and the lowest rate in those from areas with ongoing infestation, but parasite load was not measured [25]. The authors hypothesized that in areas without vector-borne transmission the circulating parasite population may be enriched with strains better adapted to transplacental transmission [25]. In contrast, a study in Bolivia reported more frequent positive hemocultures in women living in highly infested municipalities, suggesting a positive relationship between parasite loads and current vector exposure; however, the study did not quantify parasite load by qPCR, and vector data were reported at the municipality level only [26]. In our data, a woman’s cumulative years of living in an infested house was a stronger predictor of parasite load than current home infestation. We found a significant inverse dose-response relationship between years of residence in an infested house and parasite load.

We hypothesize that sustained vector-borne parasite exposure and repeated superinfection by T. cruzi may act as an immune booster, allowing women to maintain effective control of the parasite despite the down-regulation of late pregnancy [17,27]. Mothers of congenitally infected infants have been shown to have lower interferon-γ production and decreased T-cell and monocyte activation compared to infected women without transmission, both at the time of delivery and post-partum [17]. The association of twin births with higher parasite load and increased transmission risk may also be related to intensified down-regulation in mothers carrying twins [28].

Parasite genetic differences are hypothesized to contribute to the wide variation in reported congenital transmission rates [29,30], but no strain differences have been detected between women who transmitted to their infants and those that did not [16,31,32]. However, little variation was seen in the genotypes infecting the women in these studies, and nearly all available data come from the Southern Cone, where T. cruzi II, V and VI are predominant [33,34]. Data from areas where T. cruzi I is the predominant genotype are too sparse to enable analysis of risk based on parasite lineage [35,36]. Furthermore, molecular tools with higher discriminatory power may be needed to adequately address this issue.

The low sensitivity of standard early diagnostic methods for congenital Chagas disease is a major challenge to control programs and research in the field. Microhematocrit in a single neonatal specimen likely misses more than 50% of infected infants [37,38]. Serology in late infancy has high sensitivity, but follow-up rates are often below 20% [18,39]. While PCR has much higher sensitivity than micromethod for early diagnosis, the technique requires a well equipped laboratory and experienced personnel, and is beyond the resources of most control programs. In research studies, infants not diagnosed early in life may be misclassified as uninfected if specimens are not examined at 6–9 months by serology. Heterogeneity of follow-up may therefore contribute to differences in apparent transmission rates between study sites [37]. The relatively low number of congenital transmissions limited our statistical power to show associations. Future investigations would be helpful to track vector exposure more precisely, measure parasite load and immune responses at different times during pregnancy and augment congenital transmission data to provide more statistical power.

Our data suggest that cumulative vector exposure may be a determinant of T. cruzi blood levels in pregnant women, with an inverse relationship between years of exposure and parasite load. Although it is impossible to assess entomological parameters for exposures that occurred in the past, the prevalence of domestic infestation and T. cruzi infection in domestic vectors in southern Bolivia has been extremely high, even in recent years [40]. In our study, 44% of infected women and 60% of those who transmitted to their infants reported never having lived in an infested house. Although it cannot be proven, many of these women may have been themselves infected congenitally. Successful vector elimination will not eliminate risk for the remaining infected women and their infants [41]. Vigilance in screening for congenital transmission must be maintained in regions that have eliminated vector-borne T. cruzi transmission and in non-endemic countries, such as the United States and Spain, with infected immigrants.

Acknowledgments

The authors thank Edith Malaga, Sara Quispe, Edith Hinojosa, Margot Ramirez and Eliana Saenza.