The Immunoglobulin M-Shed Acute Phase Antigen (SAPA)-test for the Early Diagnosis of Congenital Chagas Disease in the Time of the Elimination Goal of Mother-to-Child Transmission

Abstract

Background

Diagnosis of congenital Chagas disease (CChD) in most endemic areas is based on low-sensitive microscopy at birth and 9-month immunoglobulin G (IgG), which has poor adherence. We aim to evaluate the accuracy of the Immunoglobulin M (IgM)-Shed Acute Phase Antigen (SAPA) test in the diagnosis of CChD at birth.

Methods

Two cohort studies (training and validation cohorts) were conducted in 3 hospitals in the department of Santa Cruz, Bolivia. Pregnant women were screened for Chagas disease, and all infants born to seropositive mothers were followed for up to 9 months to diagnose CChD. A composite reference standard was used to determine congenital infection and was based on the parallel use of microscopy, quantitative polymerase chain reaction (qPCR), and IgM–trypomastigote excreted-secreted antigen (TESA) blot at birth and/or 1 month, and/or the detection of anti–Trypanosoma cruzi IgG at 6 or 9 months. The diagnostic accuracy of the IgM-SAPA test was calculated at birth against the composite reference standard.

Results

Adherence to the 6- or 9-month follow-up ranged from 25.3% to 59.7%. Most cases of CChD (training and validation cohort: 76.5% and 83.7%, respectively) were detected during the first month of life using the combination of microscopy, qPCR, and/or IgM-TESA blot. Results from the validation cohort showed that when only 1 infant sample obtained at birth was evaluated, the qPCR and the IgM-SAPA test have similar accuracy (sensitivity: range, 79.1%–97.1% and 76.7%–94.3%, respectively, and specificity: 99.5% and 92.6%, respectively).

Conclusions

The IgM-SAPA test has the potential to be implemented as an early diagnostic tool in areas that currently rely only on microscopy.

diagnostics, congenital Chagas disease, Trypanosoma cruzi, IgM antibodies, shed acute-phase antigen

The Pan American Health Organization has launched the framework for the elimination of mother-to-child transmission of Chagas as a public health problem in the Americas [1]. Concerning Chagas disease, this initiative’s goal is to increase testing in pregnant women and neonates of seropositive mothers to >90% by 2020 [1]. Detection of congenital Chagas disease (CChD) represents a key opportunity for antiparasitic treatment, which is highly efficient and does not produce the same adverse events in infants that are seen in adults [2, 3].

Diagnosis of CChD in the national control programs of Bolivia and other endemic countries still relies on the microscopy observation of parasites (micromethod) and/or the detection of immunoglobulin G (IgG) antibodies from 6 to 12 months of age [2, 4, 5]. Due to the low sensitivity of microscopy, negative microscopy can only be confirmed by a negative IgG serology. IgG detection has good accuracy in most endemic areas when 2 tests are used in parallel, but they can only be used after maternal IgG antibodies have cleared [2, 4, 5]. This diagnostic algorithm is inefficient, leading to loss to follow-up in 53%–82% of infants at risk for infection [6, 7]. Besides, microscopy detects only 12.5%–33.3% of cases during the first month of life and is highly subjective [2, 4, 5, 8]. Nine months of follow-up represents an important challenge for health systems in high-endemicity areas such as Bolivia, where 18.7%–47.4% of pregnant women have Chagas disease [2, 6, 9, 10]. Although screening at 9–12 months can be part of integrative approaches to infant healthcare, it is usually not implemented.

Two techniques, polymerase chain reaction (PCR) and the immunoglobulin M (IgM) TESA blot (trypomastigote excretory-secretory antigen Western blot for detection of anti–Trypanosoma cruzi IgM antibodies) during the first month of life have been previously evaluated and have shown superior sensitivity over microscopy. However, both are mainly used in research studies [2, 4, 10].

This study evaluates the IgM-SAPA test based on the use of a recombinant shed acute-phase antigen (SAPA) and camelid antibodies. The test is proposed to increase the early detection of CChD at birth in areas where quantitative PCR (qPCR) will be difficult to implement.

MATERIALS AND METHODS

Study Design and Participants

Two studies (a training and a validation cohort) were conducted in 3 hospitals in the department of Santa Cruz, Bolivia. The training cohort was conducted from 2010 to 2014 in 2 hospitals: Hospital Universitario Japones (HUJ) in the city of Santa Cruz de la Sierra, and the Municipal Hospital of Camiri (HMC) in Camiri [2, 9]. The validation cohort was conducted from 2016 to 2018 in the Municipal Women’s Hospital Dr Percy Boland Rodriguez (HPB), Santa Cruz de la Sierra.

From June 2010 to April 2014 for the study in HMC, and the entire duration of activities in HUJ, the cohorts were performed under complete research funding. From May 2014 to December 2014 in HMC, and for the length of the study in HPB, the studies were conducted only under partial research funding.

Both cohorts, the training and validation cohorts, screened women in labor for Chagas disease, and all infants born to seropositive mothers were followed for up to 9 months to diagnose CChD (Figure 1). The IgM-SAPA test was evaluated retrospectively in blood samples of neonates obtained at birth.

Figure 1.

Flowchart of enrollment and follow-up of participants for diagnosis of congenital Chagas in the department of Santa Cruz, Bolivia. Abbreviations: IgG, immunoglobulin G; IgM-SAPA test, immunoglobulin M shed acute-phase antigen test; IgM-TESA blot, Western blot with trypomastigote excretory-secretory antigen for the detection of immunoglobulin M antibodies; qPCR, quantitative polymerase chain reaction.

As part of the National Control Program of Chagas disease in Bolivia, all pregnant women without exception are tested for Chagas disease, and coverage is usually >90% [6].

To determine the specificity of the IgM-SAPA test, we evaluated samples of noninfected neonates born of seropositive mothers. We assessed specimens of neonates without risk of infection born to seronegative mothers from an endemic country (training cohort), and from a nonendemic area, which were part of a study that examines children in the first 2 years of life in Santa Rosa Hospital, Piura, Peru.

The institutional review boards of HUJ, Universidad Catolica Boliviana, HPB, Hospital Santa Rosa, Universidad Peruana Cayetano Heredia, and Johns Hopkins Bloomberg School of Public Health approved the protocol.

Enrollment and Follow-up

Training Cohort Study

Venous blood samples were obtained from women in labor to screen for Chagas disease using 2 rapid tests: the lateral flow assay Chagas Detect Plus (CDP; InBios International, Seattle, Washington) and the PolyChaco indirect hemagglutination assay (IHA; Lemos Laboratories, Santiago del Estero, Argentina) at a single dilution of 1:16. Cord blood was collected from all births to women with positive results on either screening assay. All maternal sera were subsequently tested by an IgG enzyme-linked immunosorbent assay (ELISA) (Chagatest recombinant version 3 or lysate, both from Wiener Lab Group, Rosario, Argentina). Specimens with discordant results were further tested with IgG-TESA blot. Positive results by at least 2 serological tests were required for a confirmed diagnosis [2, 4].

Blood samples from umbilical cords of infants were obtained at 0 months of age, and venous blood samples were collected at 1, 6, and/or 9 months of age.

Validation Cohort

The study protocol for maternal screening and follow-up of infants was similar as described above [2, 4] with the following exceptions: (1) Maternal testing was done using the rapid test CDP and the IHA at a single dilution of 1:16. Only samples with discordant results were evaluated by the Chagatest recombinant version 3; and (2) samples at 0 months from all neonates born from seropositive mothers were obtained within 48 hours of birth by venipuncture.

Composite Reference Standard for the Diagnosis of Congenital Infection

Congenital infection was determined at 0, 1, 6, or 9 months of age in all infants born from seropositive mothers (Table 1).

Table 1.

Cases of Congenital Chagas Disease Determined Using the Composite Reference Standard in the Training and Validation Cohort

Group and Age	Composite Reference Standard	Training Cohort	Validation Cohort
Early diagnosis
0 mo, No.	Positive microscopy	6a	19a
	Positive qPCR and IgM-TESA blot	10	12
0 and/or 1 mo, No.	Positive microscopy	3b	1c
	Positive qPCR at 0 and 1 mo	3	1
	Positive qPCR and IgM-TESA blot at 1 mo	2	2
	Positive qPCR at 0 mo and IgM-TESA blot at 1 mo	1	0
	Positive IgM-TESA blot × 2	1	0
Total, no./No. (%)	…	26/34 (76.5)	35/43 (81.3)
Late diagnosis	…
6–9 mo, no./No. (%)	Positive by 2 IgG tests	8/34 (23.5)	8/43 (18.6)
Total (early and late diagnosis), No.	…	34	43

Neonates were considered to have congenital Chagas disease (CChD) at birth if the result of microscopy was positive and/or the results of qPCR and IgM-TESA blot were positive. Neonates were considered to have CChD at 1 month if (1) microscopy at 1 month was positive; (2) qPCR and IgM-TESA blot were positive at age 1 month; (3) qPCR (or IgM-TESA blot) was positive at 0 month and IgM-TESA blot (or qPCR) was positive at age 1 month; (4) they had 2 positive results by qPCR at different times (0 and 1 month of age); or (5) had 2 positive results by IgM-TESA blot at different times (0 and 1 month of age). Cases were assigned based on the earliest criterion of diagnosis.

Abbreviations: IgG, immunoglobulin G; IgM-TESA blot, Western blot with trypomastigote excretory-secretory antigen for the detection of immunoglobulin M antibodies; qPCR, quantitative polymerase chain reaction.

^aAll neonates in these cells were also positive by qPCR and IgM-TESA blot at birth.

^bTwo of these infants were positive by qPCR at birth and 1 month, and 1 was positive by qPCR at birth and to IgM-TESA blot at 1 month.

^cThe sample was also positive by qPCR at 0 and 1 month.

Composite reference standard at age 0 or 1 month

Early diagnosis during the first month was performed using the micromethod, qPCR in blood clot samples, and IgM-TESA blot, all of them evaluated in parallel. The procedures of the micromethod, qPCR, and IgM-TESA blot have been extensively described [2, 4, 10]. An improved version of the qPCR was used in the validation cohort [11]. For the IgM-TESA blot, the presence of 6 or 4 SAPA bands was used as a positive criterion in the training and validation cohort, respectively [2, 10].

The criteria for a diagnosis at 0 to 1 months of age required positive results by microscopy, or by 2 different tests at the same time point (eg, qPCR and IgM-TESA blot), or 2 positive results by the same test (qPCR or IgM-TESA blot) in samples obtained at 2 different time points (0 and 1 month).

Antiparasitic treatment was administered to all neonates as soon as the diagnosis of CChD was confirmed. All infants born of seropositive mothers who were negative or inconclusive at age 0 months were followed up until 9 months. Neonates with positive diagnoses at birth were also followed up until 6–9 months of age to determine treatment efficacy. Still, since the diagnosis had already been confirmed, their samples were not included at subsequent time points for evaluation of test accuracy.

Composite reference standard at age 6 or 9 months

Late diagnosis of CChD was conducted using the Chagatest ELISA recombinant version 3 and the CDP for detection of IgG-specific antibodies. Negative results for the 2 anti–T. cruzi IgG tests when no antiparasitic treatment was administered ruled out the infection.

Development of the IgM-SAPA Test

Evolutionary Conservation of SAPA Across Genotypes

A previously published SAPA sequence (GenBank accession number M21582) was used to construct a short recombinant SAPA (sh-SAPA) for the IgM-SAPA test by removal of the C-terminal repeat sequence and was used to examine the evolutionary conservation of the region across T. cruzi genotypes. The Basic Local Alignment Search Tool (BLAST) algorithm was used to assess alignment across genotypes using publicly available T. cruzi genomes from TriTrypDB version 46, 6 November 2019 [12]. Alignments were manually inspected, and statistical similarity bit scores (S) and E-values (measures of significance representing the number of different alignments with scores equivalent to or better than S expected to occur by chance) were evaluated. The SAPA sequence was found to align with all T. cruzi genomes irrespective of genotype (all E < 2 × 10⁻³), at a median percentage identity of 93% (range, 75%–100%), suggesting that the region is well conserved across strains and an ideal candidate for diagnostics.

ELISA Methods

IgM detection by ELISA was conducted as follows: sh-SAPA (6.4 ng/µL) was incubated overnight in carbonate/bicarbonate buffer (pH 9.0) in Immulon 4HBX ELISA plates. After 5 washing steps with phosphate-buffered saline Tween 0.05% (PBST), nonspecific binding was blocked with 5% bovine serum albumin in PBST at 37°C for 1 hour. After 4 washing steps with PBST, serum samples were diluted 1:50 in InBios diluent buffer for serum (InBios International), added to specific wells, and incubated at 37°C for 1 hour. After another 5 washing steps with PBST, wells were incubated with biotin anti-IgM conjugate (Thermo Scientific) at a dilution of 1/10 000 in InBios diluent buffer for conjugate for 1 hour at 37°C. This conjugate consists of a 14 kDa llama antibody fragment (affinity ligand) that specifically binds to the μ chain of human antibodies. After 5 washing steps with PBST, an ultra-streptavidin–horseradish peroxidase conjugate was added at a dilution of 1/5000 in InBios diluent buffer for conjugate and incubated for 30 minutes at 37°C. After 5 washing steps, wells were incubated with 1 mg/mL of o-phenylenediamine dihydrochloride (Sigma) and 0.3% of hydrogen peroxide for 30 minutes and stopped by addition of 50 µL of 2N sulfuric acid. Optical density (OD) values were measured at 492 nm.

Statistical Analysis

Maternal characteristics were compared between the 3 hospitals using the Kruskal-Wallis test for continuous variables and Fisher exact test for categorical variables.

Specimens from the training cohort were used to optimize the IgM-SAPA test and determine the appropriate OD cutoff values using the receiver operating characteristic curve (ROC) compared against the composite reference standard. The performance of the test and the calculated OD cutoff value was then evaluated in the validation cohort.

The accuracy of the IgM-SAPA test was calculated against the composite reference standard using 2 × 2 tables, and 95% confidence intervals (CIs) were obtained using the Clopper-Pearson interval.

Sample size calculation was done using the sensitivity reported of the IgM-TESA blot against microscopy, qPCR, and 9-month IgG serology (73.7%) [2]. We hypothesized that the sensitivity of the IgM-SAPA test would be 90%, requiring 32 congenital cases to give 80% power with α set to 5%. Since the prevalence of congenital infection in this department is 7.1%, and 44% of infants born of seropositive mothers will complete the 9-month follow-up program [2, 4], we aimed to recruit at least 703 at-risk neonates. A 2-sided P value of < .05 was considered significant. All statistical analyses were done using Stata statistical software version 14 (Stata Corp, College Station, Texas).

RESULTS

A total of 5505 women were enrolled; of them, 51 of 1286 (4.0%) and 136 of 4219 (3.2%) were excluded from the training and validation cohorts, respectively. Reasons for exclusion were inconclusive results of Chagas serology or inability to make a diagnosis of Chagas disease because of the lack of blood sample (n = 92 [1.7%]), and missing information for >2 epidemiological/clinical variables (n = 98 [1.8%]). Of the ones with missing data, 79.6% (78/98) and 20.4% (20/98) were seronegative and seropositive women, respectively (P < .001).

The seroprevalence of maternal Chagas disease was different in the 3 hospitals (P < .001), with HMC having the highest prevalence (48.8%), followed by HPB (21.4%) and HUJ (18.7%) (Table 2). History of spontaneous abortion or stillbirth was significantly higher in the HPB (6.4%) and HUJ (6.2%) groups compared with HMC (2.5%) (P = .003), because HUJ and HPB are tertiary hospitals. Measures of recent vector exposure indicated that mothers in HMC have higher risk factors for vector exposure as compared to the other hospitals (P < .001).

Table 2.

Characteristics of Women Who Were Received for Their Delivery in Hospital Municipal Camiri, Hospital Japones, and Hospital Percy Boland in the Department of Santa Cruz, Bolivia

Characteristic	Hospital Camiri	Hospital Japones	Hospital Percy Boland	P Value: HMC vs HUJ	P Value: HMC vs HPB	P Value: HUJ vs HPB
Total presenting, No.	515	1286	4219
Total included, N	512	1235	4083
Trypanosoma cruzi seroprevalence	48.8 (250/512)	18.7 (231/1235)	21.4 (875/4083)	<.001	<.001	.039
Seropositive mothers with follow-up at age 1 mo	62.4 (156/250)	76.7 (177/231)	47.3 (414/875)	<.001	<.001	<.001
Seropositive mothers with complete follow-up at age 6 or 9 mo	58.4 (146/250)	59.7 (138/231)	25.3 (221/875)	.765	<.001	<.001
Prevalence congenital cases at birth	1.2a (3/254)	6.8b (17/250)	7.3c (65/891)	.001	<.001	.7888
Prevalence congenital cases at 1 mo	2.6d (4/154)	3.6d (6/166)	2.4d (8/327)	.601	.921	.461
Prevalence congenital cases at 6 or 9 mo	1.4e (3/142)	3.9e (5/127)	3.6e (8/221)	.379	.414	.881
Estimated congenital transmission	5.5f (14/254)	13.6f (34/250)	12.8f (114/891)	.002	.001	.738
Age, y, mean (SE)	24.7 (0.3)	25.4 (0.2)	24.9 (0.1)	.06	.873	.021
History of spontaneous abortion or stillbirth	2.5 (13/512)	6.2 (77/1235)	6.4 (263/4083)	.001	.001	.795
Had received a Chagas test before	71.8 (370/515)	64.3 (794/1235)	87.5 (3571/4083)	.001	<.001	<.001
Had a positive result of Chagas before	27.2 (140/515)	13.7 (169/1235)	10.9 (443/4083)	<.001	<.001	<.001
Lives in a rural area now	34.4 (132/384)	26.3 (324/1232)	20.7 (843/4083)	.003	<.001	<.001
Lives in an infested house now	22.7 (117/515)	5.7 (70/1235)	6.8 (278/4063)	<.001	<.001	.150

Data are presented as percentage (no./No.) unless otherwise indicated. All congenital cases detected in Hospital Municipal Camiri, Hospital Universitario Japones, and 43 cases in Hospital Percy Boland were determined using the diagnostic criteria described in the Methods and Table 1. Due to the high numbers lost to follow-up at 1 and 9 months, the remaining cases in Hospital Percy Boland were assigned based only on the results of quantitative polymerase chain reaction at birth. Bold values denote statistical significance.

Abbreviations: HMC, Hospital Municipal Camiri; HPB, Hospital Percy Boland; HUJ, Hospital Japones; SE, standard error.

^aThe denominator includes 4 twin births that were observed.

^bThe denominator includes 19 twin births that were observed.

^cThe denominator includes 12 twins and 1 triple birth that were observed.

^dCases that were previously detected at birth were not included.

^eCases that were previously detected at birth and 1 month were not included.

^fTo determinate the estimated rate of congenital transmission, we used the prevalence of congenital infection at 1 month and 6–9 months in infants that completed the follow-up period to input the number of congenital cases in infants who were lost to follow-up during those time points. The imputation was done under the assumption that the prevalence of congenital infection was the same between infants who completed and did not complete the follow-up period.

Adherence to 6- or 9-month follow-up was higher in HUJ (146/243 [60.1%]) and HMC (108/146 [74.0%]), which were studies done under complete research funding. Adherence decreased in HMC (38/104 [36.5%]) when research funding decreased (P value < .001). The adherence in HPB was lower (43.8%) (P < .001), since the study in that hospital was always conducted under partial research funding.

A total of 77 congenital cases were evaluated; 34 and 43 were from the training and validation cohorts, respectively. In both cohorts, most cases (training cohort: 26/34 [76.5%; 95% confidence interval {CI}, 58.8%–89.3%; validation cohort: 35/43 [83.7%; 95% CI, 66.6%–91.6%) were diagnosed during the first month of life Table 1. Among mothers of infants that were diagnosed at 6 or 9 months, 64.3% (95% CI, 35.1%–87.2%) and 92.9% (95% CI, 66.1%–99.8%), respectively, had not been living in a rural area and had not seen the vector during the year of their pregnancy, suggesting a low risk of vector transmission, and 80.0% (95% CI, 51.9%–95.7%) had cesareans to deliver their neonates, suggesting a low risk of transmission at the moment of delivery.

Based on the results of the ROC curve in the training cohort, the OD cutoff of 0.23 provided the highest Youden index (Figure 2). We decided to use this cutoff for subsequent analysis.

Figure 2.

Receiver operating characteristic curve in the training cohort to determine the cutoff that provides the best balance between sensitivity and specificity in the immunoglobulin M shed acute-phase antigen test. Abbreviation: ROC, receiver operating characteristic.

Linear regression analysis showed a positive association between IgM and parasitemia levels measured by qPCR (coefficient: 0.19; P < .001) (Figure 3).

Figure 3.

Relationship between optical density values of levels of immunoglobulin M (IgM) antibodies measured by the IgM shed acute-phase antigen test and levels of parasitemia measured by quantitative polymerase chain reaction at birth. Abbreviations: Cof, coefficient of linear regression analysis; ELISA, enzyme-linked immunosorbent assay; IgM, immunoglobulin M; OD, optical density; SAPA, shed acute-phase antigen.

In the validation cohort, the sensitivity of the qPCR and the IgM-SAPA test in infants with early diagnosis of the infection was 97.1% (34/35) and 94.3% (33/35), respectively. qPCR at birth had a sensitivity slightly higher than IgM-SAPA test, but any of the tests achieved a sensitivity >90% when all cases of CChD were included (Table 3).

Table 3.

Sensitivity of Quantitative Polymerase Chain Reaction, Microscopy, Immunoglobulin M (IgM) Trypomastigote Excreted-secreted Antigen Blot, and IgM Shed Acute-phase Antigen Test at Birth in the Diagnosis of Congenital Chagas Disease in the Training and Validation Cohorts

Cohort	Testing at Birth	no./No.	%	(95% CI)
Training cohort
Infected infants diagnosed at 0 or 1 mo (n = 26)	qPCR	22/26	84.6	(65.1–95.6)
	Microscopy	6/25	24.0	(9.3–45.1)
	IgM-TESA blot	18/25	72.0	(50.6–87.9)
	IgM-SAPA test	21/26	80.8	(60.6–93.4)
Infected infants diagnosed at 6–9 mo (n = 8)	qPCR	0/8	0.0	(.0–36.9)
	Microscopy	0/8	0.0	(.0–36.9)
	IgM-TESA blot	0/8	0.0	(.0–36.9)
	IgM-SAPA test	0/8	0.0	(.0–36.9)
All infected infants (n = 34)	qPCR	22/34	64.7	(46.5–80.3)
	Microscopy	6/33	18.2	(7.0–35.5)
	IgM-TESA blot	18/33	54.5	(36.4–71.9)
	IgM-SAPA test	21/34	61.2	(43.6–77.8)
Validation cohort
Infants diagnosed at 0 or 1 month (n = 35)	qPCR	34/35	97.1	(85.1–99.9)
	Microscopy	19/35	54.3	(36.6–71.2)
	IgM-TESA blot	34/35	97.1	(85.1–99.9)
	IgM-SAPA test	33/35	94.3	(80.8–99.3)
Infants diagnosed from 6 to 9 mo (n = 8)	qPCR	0/8	0.0	(.0–36.9)
	Microscopy	0/3	0.0	(.0–70.0)
	IgM-TESA blot	0/7	0.0	(.0–41.0)
	IgM-SAPA test	0/8	0.0	(.0–36.9)
All infected infants (n = 43)	qPCR	34/43	79.1	(64.0–90.0)
	Microscopy	19/43	44.2	(29.1–60.1)
	IgM-TESA blot	34/43	79.1	(64.0–90.0)
	IgM-SAPA test	33/43	76.7	(61.4–88.2)

Abbreviations: CI, confidence interval; IgM-SAPA test, immunoglobulin M shed acute-phase antigen test; IgM-TESA blot, Western blot with trypomastigote excretory-secretory antigen for the detection of immunoglobulin M antibodies; qPCR, quantitative polymerase chain reaction.

The specificity of the IgM-SAPA test in uninfected infants born to seropositive mothers was 98.9% (344/348) in the validation cohort and 92.6% (201/217) in the training cohort. In infants with no risk of CChD, the specificity was 97.8% (90/92) in infants born to seronegative mothers in Santa Cruz, and 98.9% (184/186) in neonates born to mothers in Piura, Peru (Table 4).

Table 4.

Specificity of Diagnostic Tests in Specimens Collected at Birth in Different Groups of Uninfected Infants

Cohort	Testing at Birth	no./No.	%	(95% CI)
Training cohort: Uninfected infants of infected mothers (n = 348)	qPCR	294/296	99.3	(97.6–99.9)
	Microscopy	338/338	92.6	(98.9–100.0)
	IgM-TESA blot	335/346	96.8	(94.4–98.4)
	IgM-SAPA test	344/348	98.9	(97.1–99.7)
Training cohort: Uninfected infants of seronegative mothers (n = 92)	qPCR	NP	NP	NP
	Microscopy	NP	NP	NP
	IgM-TESA blot	NP	NP	NP
	IgM-SAPA test	90/92	97.8	(92.4–99.7)
Validation cohort: Uninfected infants of infected mothers (n = 217)	qPCR	212/213	99.5	(97.4–99.9)
	Microscopy	189/190	99.5	(97.8–100.0)
	IgM-TESA blot	144/149	96.6	(92.3–98.9)
	IgM-SAPA test	201/217	92.6	(88.3–95.7)
Uninfected infants from nonendemic area (n = 186)	qPCR	NP	NP	NP
	Microscopy	NP	NP	NP
	IgM-TESA blot	186/186	100.0	(98.0–100.0)
	IgM-SAPA test	184/186	98.9	(96.2–99.9)

Abbreviations: CI, confidence interval; IgM-SAPA test, immunoglobulin M shed acute-phase antigen test; IgM-TESA blot, Western blot with trypomastigote excretory-secretory antigen for the detection of immunoglobulin M antibodies; NP, not performed; qPCR, quantitative polymerase chain reaction.

DISCUSSION

This study shows that the IgM-SAPA test can be used for the early diagnosis of CChD and has similar performance to the improved version of qPCR. This test utilizes an ELISA format that can be easily implemented in endemic areas.

IgM detection is used for the diagnosis of Zika, dengue, syphilis, cytomegalovirus, toxoplasmosis, and other infections [13–17]. In Chagas disease, previous publications have shown contradictory results [18, 19]. Differences in the performance of IgM tests could be attributed to the source of antigen and the test used as a gold standard. When a lysate antigen of the trypomastigote stage was used, the sensitivity and specificity were 82.9% and 29.4%, respectively [19]. The low specificity of previous IgM tests can be also explained because uninfected fetuses from chronically infected mothers can develop an adaptive response [20], so the selection of antigens of acute infection is essential. In previous studies, higher specificity was obtained with a recombinant SAPA at birth (sensitivity and specificity: 83.3% [10/12] and 100% [12/12], respectively) [18]. The IgM-TESA blot has good performance for the early diagnosis of CChD, but is challenging to implement in endemic areas because it is based on a crude antigen, and electrophoretic separation is needed to discriminate SAPA reactions [21]. Besides, there is no commercially available TESA blot. Previous studies evaluated the detection of IgG antibodies by ELISA to recombinant SAPA in the early diagnosis of CChD by subtracting the OD value of the mother to the OD value of the infant and found a sensitivity of 90.5% [22]. However, this procedure requires evaluation of maternal and infant samples at the same time, and factors such as hemolysis and high bilirubin and lipid content in one of the samples may affect OD subtractions.

Differences in the sensitivity of diagnostic tests were found between the 2 cohorts. In the training cohort, the IgM-SAPA test was conducted with samples that were stored for >5 years. In some of these samples, protein degradation was observed: When the IgM-TESA blot results were repeated, they had converted to negative even though they had been positive on initial testing (data not shown). An improved version of the qPCR was used on the validation cohort [11]. Because of these differences, the sensitivity of the qPCR and the IgM-SAPA test were lower in the training cohort.

Low sensitivity of all diagnostic tests at birth could be attributed to the time that vertical transmission occurs, parasite strain, and host response [23, 24]. Although vector transmission cannot be excluded, most infected infants that were detected at 6 or 9 months were not living in a rural area and were not exposed to the vector.

This study constitutes a robust systematic evaluation of diagnostic tests with 2 different sizeable cohorts of infected and at-risk infants in Bolivia, the country with the highest prevalence of Chagas disease in the world and where the discrete typing unit (DTU) V of T. cruzi is predominant in human infections [25–29]. The utility of the IgM test will need to be assessed in other geographic areas with different T. cruzi DTUs [21]. However, bioinformatic analysis of the SAPA sequence used in this study and previous literature suggests that this is a conserved protein expressed by all T. cruzi DTUs [30]. SAPA constitutes an immunodominant and tandemly repeated amino acid motif that is expressed mainly by the trypomastigote state [30, 31]. Based on that, we hypothesized that antibodies against SAPA could be detected in infections produced by all T. cruzi DTUs primarily in the acute phase, which correspond to CChD. Another limitation was the high percentage of loss to follow-up of infants born of seropositive mothers. This could produce a selection of positive individuals based on the tests that were used for early diagnosis. However, our composite standard reference that was used during the first month consisted of the detection of 2 different forms of the parasite (microscopic observation of trypomastigotes and T. cruzi DNA) and immune response to a specific acute phase protein (IgM to SAPA). The use of this composite may provide a diverse group of positive individuals. Finally, biotin levels in serum may interfere with the results of our biotin-streptavidin system [32]. Future studies may determine the degree of this interference that, if found, should be reported to stakeholders. However, mild biotin deficiency can occur during pregnancy [33, 34], and the use of high-dose biotin supplements (up to 20 mg) that are marketed for hair and nail care may be rare in endemic countries.

Implementation of the IgM-SAPA test will facilitate early diagnosis of CChD and could provide the opportunity to administer antiparasitic treatment early in life when treatment is highly effective and safe [3].

Notes

Acknowledgments. The authors acknowledge work by health and administrative professionals at Hospital Universitario Japones, Universidad Catolica Boliviana, and Municipal Women’s Hospital Dr Percy Boland Rodriguez in the department of Santa Cruz, Bolivia, as well as at Hospital Santa Rosa and Universidad Peruana Cayetano Heredia in Piura and Lima, Peru. The authors also express their gratitude to InBios International, Inc, for the provision of all the rapid tests of Chagas disease for maternal screening.

Financial support. The work was supported by Fondo Nacional de Desarrollo Científico, Tecnológico y de Innovación Tecnológica, FONDECYT, Perú (N084-2016) to Y. E. C.-S.; the National Institutes of Health (R01-AI87776 and D43-TW010074 to R. H. G. and D43 TW007393 to A. G. L.); and InBios International, Inc to R. H. G. through the Johns Hopkins Bloomberg School of Public Health.

Potential conflicts of interest. Y. E. C.-S. reports nonfinancial support from InBios International Inc during the conduct of the study and outside the submitted work. Y. E. C.-S. is also planning to patent the IgM-SAPA test for congenital Chagas disease. C. B. reports grants from Mundo Sano Foundation and personal fees from UpToDate, both outside the submitted work. K. H. is an employee of InBios International Inc. R. H. G. reports grants, nonfinancial support, and other from InBios International during the conduct of the study. All other authors report no potential conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

Author notes

Chagas Working Group in Peru and Bolivia: Manuela Verastegui, Maritza Calderon, Clarisa Chávez, Jean Karla Leigue, Edith Hinojosa, Federico Urquizu, Mirko Gorena, Victoria Serrudo, Lilia Cabrera, Yomara K. Romero.

Presented in part: 68th Annual Meeting of the American Society of Tropical Medicine and Hygiene, National Harbor, Maryland, 20 November 2019.