Introduction

Chagas disease (Trypanosoma cruzi infection) can cause fatal cardiomyopathy in up to 30% of infected people [1]. Transmission to mammals occurs via vector, oral, congenital, and/or transfusion/transplantation routes [2]. The triatomine vector, or “kissing bug,” serves as the predominate mode of transmission, particularly in established sylvatic and/or domestic transmission cycles [3]. Over 100 different wildlife mammalian species are competent reservoirs of disease and have been implicated in propagation of sylvatic transmission cycles in nature [4]. Canines, in particular, are important components of peridomestic transmission, resulting in a bridge between sylvatic and domestic transmission cycles [5–7]. Finally, human infections can occur when vectors establish nests inside or near the home, and vectors feed on both humans and domesticated animals [7, 8].

Disease prevalence is highest in impoverished regions of endemic countries due to a plethora of societal factors, including substandard living conditions that result in increased exposure to vectors [9]. While the southern United States is not traditionally considered an endemic area, recent evidence has implicated the establishment of vector transmission cycles, particularly in Texas [10, 11]. Historical evidence of T. cruzi infected vectors and mammalian reservoirs date back to the early 1900s [12]. While the first documented locally acquired human case was published in Corpus Christi, Texas in 1955, the south Texas region, including the Rio Grande Valley, has been the subject of investigation by public health authorities dating back to the 1940s [12].

South Texas has compounding factors that could contribute to this area being a high-risk region for transmission. Within the state, sylvatic transmission cycles have been reported with seven different vector species and 27 sylvatic mammalian reservoirs [10]. The potential for sylvatic spillover to humans in this region has been implicated from increased outdoor exposure and interactions in rural environments [13]. In addition, colonias (primarily Hispanic communities) in this region of Texas have unprecedented poverty rates and living conditions that allow for easy access for vectors to enter and colonize homes, which might place residents at an increased risk of domestic transmission [5, 14]. Despite this compounding evidence of increased potential for Chagas disease in the region, epidemiologic assessments are lacking. The aim of our current assessment was to ascertain the prevalence of T. cruzi in three different mammalian species (coyotes, stray domestic dogs, and humans) and vectors (Triatoma species) to understand the disease burden attributable to Chagas disease among sylvatic, peridomestic, and domestic cycles.

Methods

Results

Chagas Seroprevalence in Coyotes

Coyote samples collected in the Rio Grande Valley had an overall seroprevalence rate of 8% (16 out of 199; 95% CI = 4.2% to 11.8%) (Table 1). Sampled coyotes were evenly distributed by gender (45% female) and all but one were adults. There was no difference in seropositivity by year of sampling. Interestingly, seroprevalence varied with regards to county of collection, with the highest seroprevalence identified in Zapata County (16%; 10/64), followed by Jim Hogg County (14%; 3/22), Dimmit County (10%; 2/20), and Webb County (1%; 1/83) (Fig 1). No positive coyotes were identified in Cameron, Hidalgo, Starr, or Wallacy counties, although sample sizes from each of these counties were low (range 1 to 4, total tested = 10).

Download:

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

Fig 1. Trypanosoma Cruzi (Chagas Disease) Positive Samples By Species And Geographic Origin.

https://doi.org/10.1371/journal.pntd.0005074.g001

Download:

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

Table 1. Trypanosoma Cruzi (Chagas Disease) Prevalence In Coyotes, Shelter Dogs, Human Residents, And Vectors Of South Texas.

https://doi.org/10.1371/journal.pntd.0005074.t001

Discussion

Chagas disease transmission has been identified along the Texas-Mexico border dating back to the 1970s [20, 21]. Our current study is the first to assess the infection status of vectors and seroprevalence among mammalian and human populations all living in the same geographic region of south Texas. Seroprevalence was highest among the sylvatic adult coyote reservoir (8%), moderate among peridomestic juvenile dogs in community shelters (3.8%), and lowest among local residents (0.36%), with one of the three positive CCC participants having a life-long history of living in Texas. In addition to finding evidence of infection in canines and humans, we found a high percentage (56.5%) of vectors carrying the parasite, further solidifying the risk of Chagas disease transmission in the region. Prior case reports have suggested the potential for domestic transmission along the eastern side of the Texas-Mexico border [5, 20], and now our larger regional assessment confirms this risk. Compounding evidence of poverty, substandard housing, rural residential exposure to sylvatic animals, and high infection prevalence of multiple species all can contribute to an increased risk of Chagas disease transmission to local residents [10, 14, 22].

Coyotes (Canis latrans) are den dwelling animals native to North America. Habitat preferences include caves and natural holes, or abandoned domestic structures such as drainage pipes, vacant homesteads and railroad tracks [23]. Similarly, triatomine vectors prefer natural or domestic habitats, living in large numbers within dens that provide constant access to a host meal source [3]. Our finding of 8% seroprevalence among coyote populations in the Rio Grande Valley is slightly lower than a prior study in 1978 which found a 12.8% (20 out of 156) prevalence of infection [20]. A second study published in 1984 found a 14% seroprevalence rate in coyotes from across Texas; however, none of the eastern Rio Grande Valley counties were included in this sampling [24]. Tennessee, Georgia, and Virginia are other southern states with known T. cruzi positive coyote populations [25–27]. Comparable to our study, these more recent studies found seroprevalence rates between 7–10%, suggesting that infection rates might be decreasing with time or current diagnostic tests have better sensitivity-specificity.

Dog (Canis lupus familiaris) populations in the United States can be feral or domesticated; however, both groups can serve as bridge hosts for transferring Chagas disease between sylvatic environments and humans. Dogs serve as important sentinel for disease surveillance purposes as their infection rates can be early predictors of transmission risk to humans, especially considering dogs develop clinical cardiac disease quicker than humans [5, 21, 28–30]. Using public health veterinary shelters as a sampling venue is a convenient methodology to capture feral, community-owned, and domesticated dog populations. The shelter dogs in our study of the Rio Grande Valley had a seroprevalence of 3.8%, which is considerably lower than other published infection prevalence estimates among shelter dog populations from across the state. Over 48 different dog breeds in Texas have demonstrated natural infection with T. cruzi, with prevalence estimates ranging from 8.8–20.3% [31, 32]. In the greater Brownsville, Texas area, infection prevalence of shelter dogs has ranged from 7.5% in 2003 to 6.7% in 2014 [5, 32]. While our prevalence is slightly lower than other studies, the reason is most likely related to our decision to sample dogs that were under 6 months of age, allowing us to estimate incidence related to recent vector-borne or congenitally-acquired infection. By estimating incidence, we can better understand the annual contribution of disease transmission in this geographic area.

The epidemiology and seroprevalence of human infection in the southern United States is largely unknown. Even in endemic areas, human seroprevalence is typically lower than sylvatic and domestic animals due to multiple factors, including increased mammalian-vector habitat exposure, mammalian predilection for oral ingestion of the triatomine vector, and varying defecation behaviors of different triatomine species [3, 30, 33]. While sylvatic transmission cycles between wildlife and vectors have been established in the southern United States, we are still in our infancy of understanding disease burden and transmission source in infected populations. A prior study conducted in 1977 found a seroprevalence of 2.4% (12 out of 500) among eastern Rio Grande Valley residents [20], which is a sharp contrast to our finding of 0.4% (3 out of 841). Our study sampling included random selection of participants, while their study biased their results by recruiting patients at Texas Chest Hospital in Harlingen. It is likely our sampling methodologies influenced the varying rates, especially as other historical random-selection population studies reported 0.01–0.9% seroprevalence [12]. Despite our selection methodology differences, both Burkholder et al.’s study and ours included long-time residents of the Rio Grande Valley, with one positive participant in our study very likely acquiring the infection in Texas. Based on our findings of a seroprevalence estimate of 0.4%, and considering a population of 1.3 million for the Rio Grande Valley, we can estimate that ~4,600 people in this region are currently infected with Chagas, with ~1,300 at risk for developing Chagas-related cardiomyopathy. If this estimate is accurate, then the burden of Chagas disease in the Rio Grande Valley is 23 times higher than what we had previously estimated based on our findings of 1 out of 6,500 (0.02%) blood donors in Texas testing positive for the disease [34]. Future studies should aim to further clarify the true disease burden and rate of autochthonous transmission in the Rio Grande Valley, an area with documented sylvatic and domestic T. cruzi transmission [5].

Our study had a few important limitations notable for discussion. The current World Health Organization guidelines require a minimum of two positive results on different antibody-based assays for diagnostic confirmation [35]. While we used two different assays, neither are currently FDA approved in the United States; however, Stat-Pak rapid immunochromatographic assay has demonstrated efficacy in all three populations of mammals in multiple studies [6, 16–18, 27]. For the purposes of this retrospective study we felt confident in the test results, especially as they were relatively consistent with other published literature. In addition to our finding of a high rate of infection (56.5%) among local vector species, other studies have also confirmed high rates of infection (51–82%) in Triatomine vectors throughout Texas [7, 8, 10]. Provided the retrospective nature of our study, the obvious lack of travel history in these coyote and dog populations, and the establishment of known T. cruzi positive vector populations in our study, we would argue that these are true infections acquired via local vector-sylvatic mammal transmission cycles. Another possible limitation, due to our retrospective sampling of frozen sera collected 8–10 years prior, is the potential for antibody decay resulting in a lower prevalence rate. Handling of the specimens included freezing aliquots to -80°C immediately following collection, constant monitoring of freezer temperature, and adhering to discipline standards during the serum thawing process in an effort to maintain sample preservation. Finally, we cannot rule-out the potential for cross-reaction with leishmaniasis. Rare reports of cutaneous leishmaniasis have been reported in the state [36]; however, none of our three Chagas-positive study participants presented with skin ulcers, lowering the potential for cross-reaction.

In conclusion, we contribute to the growing body of evidence for autochthonous Chagas disease transmission among mammals in south Texas. Coyotes, shelter dogs, and vectors in this region continue to demonstrate high infection rates of T. cruzi. Random sampling of residents also revealed a higher than expected disease burden that had previously been undiagnosed, with one human patient suspected of having locally acquired the disease. With up to 30% of infected individuals developing a potentially fatal cardiac disease, it is imperative that we identify and treat patients before irreversible clinical manifestations have occurred. Future prospective studies are necessary to elucidate and validate the disease burden in the Rio Grande Valley.

Supporting Information

Acknowledgments

The authors wish to thank Kara Chappell and Dr. Sherri Wooding for their assistance with sample collection at the Brownsville and Edinburg animal shelters. We would also like to thank Guy Moore (Texas Department of State Health Service, retired) for collecting and sharing aliquots of sera from coyotes. For the CCC, we would like to thank our cohort team, particularly Rocio Uribe and her team, who recruited and documented the participants, and Emma Molina who performed the echocardiograms. We also thank Marcela Morris and her laboratory team for their contributions, and Christina Villarreal for administrative support. We thank Valley Baptist Medical Center, Brownsville, Texas for providing us space for our Center for Clinical and Translational Science Clinical Research Unit. We also thank the community of Brownsville and the participants who so willingly participated in this study in their city. With regards to the Triatoma species, we thank Susan Kibbe and all land owners that collected and sent triatomine insects for T. cruzi testing.