The first search identified a total of 858 records (363 in MEDLINE/PubMed, 187 in Web of Science, 175 in EMBASE, 117 in SCOPUS, and 16 in LILACS); 611 studies were removed before screening, and 247 papers remained. In total, 71 articles were not considered relevant after title and abstract screening (52) or were not retrieved (19); 176 studies that met the inclusion criteria were reviewed in depth, and 147 studies were excluded for reasons such as lack of diagnostic accuracy data (69), samples other than blood samples (56), treatment monitoring (13), and immunocompromised patients (9). Three records identified by citation searching were added. Finally, 32 papers were found to be eligible for review and meta-analysis (see PRISMA 2020 flow diagram, Additional file 3: Fig. S1). Bibliographic references were ordered alphabetically; in the case of more than one reference by the same author, they were ordered by year of publication in ascending order [5,6,7,8,9, 11, 18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43].

Of the 32 study papers, 27 papers assessed the diagnostic performance of PCR techniques [5,6,7, 11, 18, 21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42], 3 articles assessed T. cruzi-LAMP technology [8, 9, 19], and two studies evaluated both molecular techniques [20, 43]. The characteristics of the included studies are outlined in Additional file 1: Table S1 (PCR) and Additional file 2: Table S2 (LAMP). They were published between 1995 [21] and 2022 [42]. An active interest in this research topic has continued in recent years. Based on the year of publication, the bar chart in Fig. 1 illustrates this increasing trend in journal articles. Regarding study design of the studies included in this review, they are longitudinal comparative studies (some of them multicentre, e.g., Benatar et al. [18], or international, e.g., Ramírez et al. [11]), some with prospective follow-up, e.g., [6, 18, 20, 24, 34, 35, 38, 41, 43], and others retrospective (case‒control), e.g., [8, 9, 19, 22, 23, 26, 28, 33, 37]. With respect to Hernández et al. [31], the sample collection was both retrospective (for the period 2004–2011) and prospective (for the period 2012–2015). Some studies (the older ones, e.g., [5, 7, 21, 27]) do not clearly define the study design.

Graph of the number of papers included in this systematic review

A summary of the study quality assessment using the QUADAS-2 scale can be found in Fig. 2, which shows the quality evaluation of the individual included studies (Fig. 2a) and the risk of bias and applicability concerns of the included studies (Fig. 2b). The risk of bias assessment revealed that most studies carried a low risk of bias. In addition, applicability concerns were also low. Regarding the patient selection aspects, 22 out of 32 studies had a low risk of bias, 7 studies were judged to be unclear, and 3 had a high risk of bias. This result can be explained by the fact that these studies were the oldest and reported insufficient endpoints. In the index test assessment, all studies had a low risk of bias because they clearly mentioned the extraction of data for the index test. Regarding reference standard tests, 29 out of 32 studies were determined to have a low risk of bias, and 30 out of 32 were of low concern in terms of their applicability. For the flow and timing aspects, all studies demonstrated/yielded a low risk of bias in statements regarding the interval time between the reference test and the index test. Therefore, we considered the overall risk of bias to be low, and all included studies generated only low concern regarding applicability in all aspects. This risk of bias disappeared when we analysed subgroups according to diagnostic technique. Thus, the risk of bias only appeared in the cPCR and chronic phase studies.

Study quality assessment using the QUADAS-2 scale: a Quality evaluation of the individual included studies. b Risk of bias and applicability concerns graphical summary

Detailed information on the 32 included studies is summarized in Additional file 1: Table S1 and Additional file 2: Table S2. They were carried out both in endemic areas (Bolivia, Argentina, Brazil, Chile, Colombia, Venezuela) and in nonendemic areas (Spain). The studies carried out in nonendemic areas always included an immigrant population in the sample. All studies indicate the type of population, sample size and the patients' clinical phase, but in many of them, information on the clinical characteristics and description of the samples is not available. Participants were children/infants/neonates born to seropositive mothers (congenital ACD), e.g., [6, 9, 18, 35, 40, 43], and adult patients in the chronic phase, e.g., [27, 37, 39] (patients with Chagasic cardiomyopathy in Duarte et al. [26]), and/or acute phase, e.g., [28, 31, 36, 42]; controls were included for other diseases (leishmaniasis, malaria, toxoplasmosis, etc.), e.g., [36, 42]), infants born to noninfected mothers and healthy individuals from endemic or nonendemic areas, e.g., [36, 42]. The sample size described in the studies varied from 17 in Diez et al. [6] to 708 participants in Hernández et al. [31].

The studies also included information on the type of sample collected from the patients (in all cases, these were blood samples) and the number of samples taken. The blood volume taken ranged from 0.5 mL [20, 23] to 15 mL [7]; 22 of the 32 (68.7%) studies indicated the preservation conditions of the samples; none reported the time period between sample collection and extraction or performance of the diagnostic tests. A total of 6 studies (18.7%) did not specify the type of amplification control of the tests analysed. Only 13 studies (40.6%) indicated the strains used as positive amplification controls; the rest merely indicated the different DTUs used as controls. The parasite load of patients was estimated in 12 of the 32 studies analysed (37.5%). Most of the included studies used proprietary methods except for the studies validating commercial LAMP methods [8, 9, 19, 20,