1. Background

Trypanosoma cruzi, a flagellated protozoan parasite (class Kinetoplastida), is the etiological agent of Chagas disease (ChD), also known as American trypanosomiasis (trypanosomosis), a serious, potentially fatal, infection. An estimated 6 to 7 million people are infected with T. cruzi worldwide, with about 10,000 deaths annually (https://www.who.int/news/item/14-04-2022-world-chagas-disease-day-bringing-a-forgotten-disease-to-the-fore-of-global-attention; website updated in April 2022). Although this infection occurs predominantly in Latin American countries, the incidence in other global regions, particularly North America and Europe, is rising; cases outside Latin America are largely associated with migrants from endemic countries bringing the infection with them [1].

Transmission of this disease is complex and involves multiple pathways elucidated in greater detail below. Whereas the conventionally accepted (“traditional”) route of transmission is vector-borne, oral transmission, potentially by contaminated food or drink, seems to be of increasing importance and becoming more recognized. This is partly because vector-borne transmission is being reduced by initiatives such as improved housing; simultaneously, there is greater human encroachment on wilderness areas, resulting in closer contact with sylvatic transmission routes [2].

In 2006, the World Health Organization (WHO) developed a strategy to estimate the global burden of foodborne diseases according to etiological agent and constituted an expert group (The Foodborne Disease Burden Epidemiology Reference Group (FERG)) to advise on this task. The first estimates were published in 2015 [3]. In 2021, WHO decided that there was a requirement to update the estimates based on more recent data and with an extended list of etiological agents. In order to undertake this task, FERG membership was reconstituted (FERG for 2021 to 2025), with some of the previous team and some new members, to oversee collection of relevant data for this global effort.

In the 2015 FERG estimates, T. cruzi was excluded from the list of parasites due to lack of resources [4]. A subsequent publication noted that, as the focus of the foodborne burden estimates was on global impact, some important pathogens with restricted distribution were not included [5].

Indeed, a crude-level calculation of the disability-adjusted life-years (DALYs) associated with foodborne ChD—based on ChD burden reported in Global Burden of Diseases (GBD) [6] and data on the proportion of cases that are likely orally transmitted—gave an estimate of 273,000 DALYs per annum attributable to foodborne transmission globally [5]. This exceeds the DALYs associated with foodborne transmission of several of the other pathogens that were included in the first FERG estimates (see Section 4; [3,4,7]). Given that the vast majority of the ChD burden is restricted to Latin America, the burden per 100,000 of exposed population living in the endemic region is even higher. Further crude estimates based on more recent GBD data [1] are discussed in Sections 4 and 6.

The question now arises on whether and how T. cruzi should be included in this new round of estimating the global burden of foodborne disease. It is clear that although foodborne ChD is gaining recognition, the importance of the foodborne infection route is not widely understood. Indeed, many articles continue to be published in which oral/foodborne transmission is not included in the introductory text that describes the parasite, the disease, and infection routes (e.g., [8–10]), or describes oral transmission as being an infrequent infection route (e.g., [11]). Even the current English Wikipedia page states that vector-borne transmission is “the parasite’s only transmission route”; also today’s Spanish and Portuguese Wikipedia pages only describe vector-borne transmission.

More importantly, the WHO Road Map for neglected tropical diseases 2021 to 2030 [12] identifies only 4 transmission routes (vectoral, transfusion, transplantation, and congenital) in the first table regarding disease-specific targets. WHO lists 5 main objectives for eliminating ChD by 2030. These are: verification of interruption of vectorial domiciliary transmission; verification of interruption of transmission by blood transfusion; verification of interruption of transmission by organ transplantation; elimination of the congenital form of the disease; and 75% coverage of antiparasitic treatment of the eligible population. None of these addresses the importance of foodborne transmission.

2. Methods

Here, we describe the transmission routes of T. cruzi to people and, by a narrative review of the literature, discuss why foodborne ChD should be considered separately in terms of disease burden. We also explain why estimates of the ChD burden will be higher than previously calculated when clinical differences associated with oral transmission are taken into account. One limitation of our purposive approach is that, by necessity, not all relevant studies and results are mentioned.

3. Routes of transmission of ChD

The lifecycle of T. cruzi is relatively complicated and includes various routes of infection for ChD (see Fig 1). These are:

1. Vector-borne: This transmission route is often considered the “traditional” infection path in endemic areas and is related to the ability of different triatomine (reduviid bug) species to adapt to, and colonize, different ecotopes, particularly home environments [13]. During transmission, an infected vector defecates into a wound (usually, but not always, a bite wound from the vector itself), introducing infective trypomastigotes. These invade host cells, where they replicate to non-flagellated amastigote forms that may remain quiescent and intracellular, or may develop to trypomastigotes that return to the bloodstream. From here they can both infect other cells and be acquired by the vector. In addition, vectors may contaminate food; this is considered under foodborne transmission in this list (point 6b).
2. Transplacental/congenital: This usually occurs after the 12th week of gestation, when the maternal blood supply becomes diffuse and continuous in the entire placenta; high maternal parasitemia is a proposed risk factor, and transmission is thus more likely to occur when infection is acquired during pregnancy (acute phase of disease). Coinfection with HIV and/or Plasmodium vivax is associated with increased risk of congenital transmission [14].
3. Blood transfusion: Apart from plasma derivatives, all blood components from an infected donor may transmit T. cruzi and the parasite remains viable for considerable periods at room temperature, refrigerated, and frozen. Although only a few hundred cases have been published, in some places (particularly where the parasite is not endemic), it can be an important transmission route. Transmission likelihood from an infected donor depends on various factors, including: extent of parasitemia at donation, transfusion volume, strain, recipient immune status. This transmission route can be mitigated against by use of low-risk donors (questionnaires, donor screening) and use of pathogen-reduction technologies on donated blood components, e.g., UV irradiation [15–16].
4. Organ/bone marrow transplant: This transmission route is less common than blood transfusion transmission, as organ transplants occur less frequently and screening mitigates against this infection route. However, a lack of suitable donated organs has encouraged use of organs from infected donors. Reactivation of T. cruzi in donated organs is likely exacerbated by immunosuppression to limit rejection of the transplant and seems to vary according to organ and degree of immunosuppression [11].
5. Laboratory accident: The predominant stage of T. cruzi in laboratory culture is the non-infectious epimastigote stage, although infective trypomastigotes occur, as well as intermediate forms (transitional epimastigotes) that are also infectious [17]. Despite occasional reports of laboratory-acquired T. cruzi infections, predominantly from a needle or other sharp object that pierced the skin [18] (but also from handling of infected mammals; [19]), this is the least common transmission route. However, it can have severe consequences, and at least one of 65 laboratory-accident cases has proven fatal [18].
6. Foodborne/oral: Although it has been argued that foodborne/oral transmission is another form of vector-borne transmission [20], this transmission route does not necessarily involve triatomines and is therefore more usefully considered separately. In addition, even when triatomines are involved in foodborne transmission, the entry site differs (skin abrasion versus digestive tract), and this can be of pathophysiological and clinical relevance.

The 3 main routes of foodborne transmission are [21]:

1. Meatborne transmission: Ingestion of inadequately cooked meat or organs from animals infected with T. cruzi has been indicated as a transmission route, but in many cases, it is very difficult to exclude other infection routes. A systematic review of ChD transmission from consumption of game meat [22] identified only 5 reports (in 6 papers) where transmission due to ingestion of meat seemed likely. Results from 2 different animal experiments [23,24], although superficially contradictory, indicate that inadequately cooked bloody meat containing trypomastigotes could be a source of infection, but non-bloody muscle, tissue, and organs seem less likely to be so [21]. ChD infection transmitted by drinking the blood of wild animals (e.g., armadillos) as a part of religious rituals or traditional medicine has been reported from rural areas of Latin America [25].
2. Contamination of food by infected vectors: This seems to be the most common oral infection route and has thus resulted in oral transmission being considered an extended part of vector-borne transmission. This is unfortunate as the “traditional” vector-borne route results in a different clinical picture, and therefore, the 2 routes cannot be considered as one and the same. For oral transmission, fruit juices tend to be the most common, but not the only, vehicle (other products reported as vehicles include sugarcane juice, ice cream, soup, and shrimps). Food contamination is either with the feces of the vectors or, more commonly, particularly for juices, the entire vector is incorporated into the food product. As triatomine nymphs are only a couple of millimetres in size, they may be overlooked in preparation of juices [21,25].
3. Contamination of food by infected marsupials: Different species of opossum, particularly Didelphis marsupialis (common opossum), are common reservoirs of T. cruzi. When these marsupials are infected with T. cruzi, in addition to trypomastigotes circulating in blood and amastigotes colonizing tissues, the parasite will reproduce and transition to metacyclic trypomastigote within the opossum’s anal gland secretions. The lifecycle is thereby completed in the absence of the triatomines. The parasites are infective when expelled in aerosolized secretions, and contamination of food may occur when Didelphis spp. have access to open kitchens with exposed foods. This transmission pathway has been suggested (but not proven) to have occurred during outbreaks in which the oral route is suspected, but triatomines were apparently absent and infected opossums were present [26–28].

Download:

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

Fig 1. Lifecycle of T. cruzi demonstrating the predominant infection routes of people via vector-borne and foodborne transmission.

In this illustration, food is contaminated either via an infected triatomine (reduviid bug) or via secretions from an infected reservoir host. Note: Fig 1 was initially created in BioRender (www.biorender.com) and then modified. Less frequent transmission routes (not included in this figure) are: transplacental/congenital infection, blood transfusion and organ/bone marrow transplant infection, and laboratory accidents. In addition, consumption of inadequately cooked meat or blood from infected animals may act as a transmission route. An infected triatomine vector takes a blood meal from a person and infects them by defecating in or near to the wound with trypomastigotes in its feces (A). The trypomastigotes (B) invade cells and differentiate into intracellular amastigotes (C) that multiply by binary fission. These differentiate into trypomastigotes that are released into the circulation (D) and infect further cells (E) from various tissues, where they again transform into intracellular amastigotes and replicate (C). The triatomine vector becomes infected by feeding on an infected person (F). Inside the midgut of the vector (G), the ingested trypomastigotes (H) transform into epimastigotes (I). Here they multiply, before differentiating into infective metacyclic trypomastigotes in the hindgut (K) that may then be defecated into a feeding wound (A). Food, particularly fruit, may also be contaminated by being colonized by infected triatomines (L). In addition, reservoir hosts (such as opossums) can also be infected from trypomastigotes in the feces of feeding vectors (M). The circulating trypomastigote-intracellular amastigote cycle (N, O) then occurs. Food may be contaminated by trypomastigotes in the secretions of the opossums (P). Food, particularly fruit juices, contaminated by trypomastigotes (Q), either via infected triatomines (L) or secretions from infected opossums (P), may act as a vehicle for oral infection.

https://doi.org/10.1371/journal.pntd.0011898.g001

4. More severe clinical outcomes suggest that foodborne ChD should be considered separately from ChD transmitted by other routes of infection in estimating the burden of disease

Burden of disease is usually presented as DALYs, which includes both reduction in life expectancy and diminished years of healthy life [29]. Thus, the DALY burden for a particular condition is the sum of YLL (years of life lost due to premature mortality) and YLD (years lived with disability, weighted for the severity of the illness) as spread across a population (often global or national).

Many pathogens that were included in the 2015 FERG estimates could be transmitted both by food and other transmission routes. In order to estimate the burden associated with the component associated with foodborne transmission, first the entire DALY burden due to all cases of the specific hazard was estimated. The proportion of cases associated with foodborne transmission was then estimated [30], and incidence, mortality, and DALYs associated with foodborne transmission could then be calculated, assuming that clinical outcomes were independent of transmission route. Estimating DALYs is difficult and with considerable uncertainty for all pathogens, and estimating the proportion that is foodborne (for those pathogens that are not solely foodborne) is even more challenging. For various pathogens that were covered in the FERG estimates for 2015, including parasites such as Echinococcus and intestinal protozoa, a structured expert elicitation approach was used to estimate the proportion of cases likely to be foodborne [30]. The data indicate substantial uncertainty (uncertainty from individual experts and disagreement between experts) for most parasites and very high uncertainty for others (e.g., Echinococcus multilocularis).

Thus, for example, for ascariosis (disease associated with Ascaris infection), the total global burden of DALYs was estimated to be 1,317,535 (95% uncertainty interval (UI): 1,182,187 to 2,700,572), whereas the total global burden of DALYs due to foodborne transmission was estimated to be 605,278 (95% UI: 410,668 to 1,301,114) based on the proportion of foodborne cases of ascariosis, and thus DALYs being 45% (95% UI: 31 to 59) of all cases. For E. multilocularis, the equivalent calculation of DALYs associated with foodborne transmission was 312,461 (95% UI: 9,083 to 640,716), with 47% (95% UI: 4 to 75) of all cases estimated as being foodborne [4,30]. For diseases where all cases are foodborne (e.g., trichinellosis), the total number of DALYs and the number of DALYs due to foodborne transmission are equal [4].

In the crude-level estimate of a burden of 273,000 DALYs per annum attributable to foodborne transmission of ChD mentioned earlier [5], the calculation was based on the ChD burden reported in GBD, 2010 of 546,000 DALYs (95% UI: 271,000 to 1,054,000) [6], and data from a narrative review [31], in which of 959 cases, most (638) were due to oral transmission. This approximation is higher than the DALY burden estimated for many other foodborne pathogens at around the same time and included in the outputs of FERG in 2015, such as: Bacillus cereus, Brucella spp., Clostridium botulinum, Clostridium perfringens, Echinococcus granulosus, Entamoeba histolytica, Fasciola spp., Giardia, intestinal flukes, Listeria monocytogenes, Opisthorchis spp., Shiga toxin-producing E. coli, Staphylococcus aureus, and Trichinella spp. [3,4,7]. The majority of these have a global distribution and thus an even lower DALY per exposed person than foodborne ChD. More recent GBD data now available [1], give a slightly different picture—as described in Section 6.

However, extrapolation from an estimated burden of total (transmission-route independent) DALYs to the foodborne burden, based only on estimating the proportion that are foodborne is insufficient for T. cruzi. This is because for this pathogen the course and outcomes of clinical disease vary considerably according to transmission route. In particular, for foodborne transmission the rate of symptomatic acute ChD (defined as occurring relatively rapidly after infection, when trypomastigotes can be found in the circulating blood) is higher, the symptoms are more severe, there is greater cardiac involvement, and mortality is higher [21,32,33].

In general, mortality from vector-borne transmission ChD is estimated to be between 5% and 10%, whereas ChD from the foodborne infection route is estimated to be associated with 8% to 40% mortality [34], with children, pregnant or postpartum women, and the elderly accounting for most fatalities [35]. Whereas acute ChD in childhood due to vector-borne transmission often presents as a nonspecific, flu-like disease, infection via oral transmission causes more severe symptoms; fatality during acute ChD is not uncommon, particularly in children [35–38].

Comparisons of clinical signs and symptoms between ChD patients from oral infections and from vector-borne infections have been published [34,39,40] and together provide relatively clear information on important differences regarding severity of symptoms, and thus disease burden, depending on transmission route (summarized in Table 1).

Download:

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

Table 1. Summary of Chagas disease manifestation differences following vector-borne or foodborne transmission (references and greater detail included in the text below).

https://doi.org/10.1371/journal.pntd.0011898.t001

For vector-borne triatomine transmission, the acute phase of disease is broadly considered asymptomatic, with some nonspecific somatic signs (CDC https://www.cdc.gov/dpdx/trypanosomiasisamerican/index.html). However, between 3% and 60% of cases have been reported to exhibit some relatively mild symptoms (fever, headache, mononuclear phagocytic system involvement, and bilateral palpebral and/or leg oedema).

In contrast, in foodborne disease, acute symptomatic ChD occurs in nearly all patients, with close to 100% experiencing fever; other common symptoms include: myalgia, headache, leg and/or facial oedema, pericardial effusion, and abdominal pain [33,34,39,40]. In some cases of oral transmission, diarrhea, skin rash, dyspnoea, palpitations, hepatomegaly, splenomegaly, and hemorrhagic jaundice are also reported.

Cardiac manifestations (acute myocarditis) may occur in up to 10% of acute ChD cases with infection by the vector-borne route, particularly in children. However, in cases infected by the oral route, myocardial involvement occurs early, is often severe, and includes cardiac arrhythmias, congestive heart failure that may progress to cardiogenic shock, pericardial effusion, and pleural effusions [34]. Among patients with heart disease due to ChD, sudden cardiac death is the most common cause of death (55% to 60%), followed by heart failure (25% to 30%) [41].

Chronic ChD is generally considered to be the symptomatic phase following vector-borne transmission and may occur years or decades after initial infection. For most patients (60% to 70%) infected by the vectorial route, the chronic phase is asymptomatic or indeterminate, with amastigote invasion of smooth muscle occurring in the minority and leading to a cardiac form (20% to 30%) or a digestive form (megaoesophagus/megacolon) or mixed (both cardiac and digestive) in 5% to 15% of patients [39].

Long-term sequelae in the chronic phase of ChD following oral transmission are less well defined, but rapid progression to long-term cardiac or gastrointestinal dysfunction has been indicated [34].

5. Why does oral infection result in more severe symptoms?

Reasons for the differences in clinical outcome according to transmission route have been considered. One commonly cited reason is a greater parasitic load associated with oral infection compared with vectorial infection. In foodborne infection, a single triatomine crushed into a food or drink may have an infectious load of over 600,000 trypomastigotes; this is a considerably higher amount than occurs in fecal material from an infected triatomine [34,42]. In addition, a proportion of parasites entering through the skin does not survive, whereas those that are ingested may enter through cells in the mucous membrane of the oral cavity or via the stomach wall, reproducing in Peyer’s patches before migrating into the bloodstream. Both these latter routes are associated with a shorter incubation period and a higher parasite load [21]. Experimental studies in mice have shown that the 50% infectious dose (ID50) is 100-fold lower for oral challenge than for cutaneous challenge, indicating the greater efficiency of oral infection [43].

The few documented cases of meatborne infection (from ingestion of meat/blood of infected animals) also indicate greater severity in the acute phase than from vector-borne infection [20]. Interestingly, experimental infections in mice have suggested that entry of T. cruzi through the oral cavity results in more severe symptoms than through the stomach (delivered by gavage), indicating that in natural foodborne infection there are 2 entry routes that may result in different clinical responses. Differences in the mucosal pathways associated with infection site and, thus, components of the immune response, may be another reason why foodborne infection results in more severe disease than vector-borne transmission [44,45]. These authors also note that the frequent observation of facial oedema in foodborne ChD may indicate invasion already occurring in the buccal cavity.

Other considerations for why oral infection should result in greater symptom severity than infection via vector-borne infection include the association of different vectors with different transmission routes and genetic differences in T. cruzi itself.

6. Current estimates of burden of disease due to ChD and how to estimate the burden from foodborne ChD

A recent paper used data from the GBD Study (2019) to estimate both the global prevalence of ChD and the burden of ChD in terms of DALYs [1]. The authors report a global reduction in prevalence from 7,292,889 cases in 1990 to 6,469,283 cases in 2019, and a reduction in DALYs, with the estimates falling from 360,872 (95% UI: 153,746 to 450,827) in 1990 to 275,377 (95% UI: 184,453 to 459,354) in 2019 [1]. Results were provided by the Institute for Health Metrics and Evaluation (IHME), with ChD case definitions based on relevant International Classification of Disease (ICD)-9 and ICD-10 codes, and seroprevalence study results for nonfatal estimates. Burden was based on DALYs extracted directly from GBD, with YLL and YLD estimated by age, sex, and country, and all measures reported as counts and rates per 100,000 population (along with UI). The article notes that the GBD study estimates differ vastly from those calculated by other studies published in the literature and suggests various reasons for this. In particular, the authors highlight differences in immigration data for cases outside Latin America that may result in an underestimate in the prevalence of disease in non-endemic countries.

If these more recent GBD estimates [1] on ChD burden (275,377 DALYs) are used to deduce DALYs associated with foodborne transmission (as described in Section 3 [5]), with the same proportion considered foodborne (a conservative approximation of 50%; [5]), then we can derive a global foodborne ChD burden of around 138,000 DALYs (see Table 2). Although this is considerably lower than the previously estimated figure (273,000 DALYs [5]), based on older GBD data [6], it still exceeds those DALY burdens described in the FERG outputs of 2015 for most (11 of 15) of the pathogens listed as being exceeded in burden in Section 4 (namely: Bacillus cereus, Brucella spp., Clostridium botulinum, Clostridium perfringens, Echinococcus granulosus, Fasciola spp., Giardia, Listeria monocytogenes, Shiga toxin-producing E. coli, Staphylococcus aureus, and Trichinella spp. [3,4,7]).

Download:

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

Table 2. Estimates of DALYs associated with foodborne Chagas disease using older and newer global data, but not taking into account the greater disease severity associated with foodborne transmission.

https://doi.org/10.1371/journal.pntd.0011898.t002

Of particular relevance in the recent ChD estimates of burden [1], is that the authors do not mention transmission route. It is not possible to determine from the IHME database whether this was considered, but this is important given that oral ChD has increasingly replaced vector-borne transmission. In addition, there should be a heavier weighting for ChD transmitted orally than via vector bite, resulting in a greater burden of disease, in terms both of YLL and YLD. This is due to the greater fatality (particularly in younger age groups) and also more serious symptoms associated with foodborne transmission.

The data available in the IHME database do not seem to reflect this, as, for example, YLL values, have been dropping over time, in particular for younger age groups (Fig 2, [62]), especially in Tropical Latin America (Brazil and Paraguay). Although it is difficult to determine the extent of foodborne versus vector-borne burden of ChD, a synthesis of reported cases from 2002 until 2012 [31] provides 73 reports, which was double that reported in the previous decade. Among these, oral infection was the most common transmission route reported from Brazil; of 514 cases listed, 482 (94%) were described as oral transmission [31]. Similarly, the majority of reported cases from Colombia (25/49; 51%) were reported as oral transmission [31]. It has been argued that there may be more cases of vector-borne disease than foodborne disease, but foodborne cases are more visible because they result in more severe symptoms and when large outbreaks occur there is investigation and reporting bias [20]. This may be true, but as burden is linked to severity of symptoms and not only numbers of cases, and as the foodborne path may result in many cases rather than individual sporadic infections, this also contributes to the overall burden.

Download:

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

Fig 2. Change in rate of YLL due to ChD for children 0–14 years in 3 different regions of Latin America (shaded areas indicate uncertainty): data from IHME and graph created in Global Health Data Exchange GBD 2019 website (http://ghdx.healthdata.org/gbd-results-tool) [62].

In the absence of systematically sourced data on the proportion of foodborne ChD, whether these declines are associated with all potential exposures to ChD or primarily due to declining vector-borne disease cannot be estimated, particularly given the limited interventions to address the burden of foodborne ChD.

https://doi.org/10.1371/journal.pntd.0011898.g002

7. Conclusions

ChD remains a considerable public health issue in endemic countries of Latin America. In efforts aimed at quantifying the global health burden from foodborne diseases, it is important that it is not overlooked, simply because of its relatively restricted geographical area of endemicity. Even preliminary conservative estimates suggest that the burden from foodborne ChD is greater than for some other foodborne diseases with a global distribution. Exclusion of foodborne ChD from etiology-based burden of foodborne disease estimates may result in errors when risk ranking these diseases for the purpose of prioritizing interventions in endemic countries. Should FERG for 2021 to 2025 be able to include ChD in their estimates, then it will be essential not only to estimate the proportion of cases of ChD that are foodborne, but also to ensure that the more severe illness associated with this route of infection is taken into account.

Acknowledgments

The intention and scope of this manuscript was conceived within the Parasitic Diseases Task Force of The WHO Foodborne Disease Burden Epidemiology Reference Group (FERG), 2021 to 2025. We acknowledge the support of the whole FERG team, and especially the assistance of the WHO secretariat, in particular Yuki Minato and Charlee Roberts.

The views expressed in this paper do not necessarily represent the opinions of WHO nor those of the entire FERG.