| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 University of Utah, Associated Regional and University Pathologists (ARUP) Institute for Clinical and Experimental Pathology, Salt Lake City, Utah, USA
2 ,3 Department of Medicine, Division of Cardiology2 and Departments of Pathology, Pediatrics and Medicine3 , University of Utah School of Medicine, Salt Lake City, Utah, USA
Correspondence
Christine M. Litwin
christine.litwin{at}path.utah.edu
Received 6 May 2005
Accepted 30 September 2005
Abbreviations: ECG, electrocardiogram; EIA, enzyme immunoassay; FDA, Food and Drug Administration; IFA, immunofluorescent assay; SLE, systemic lupus erythematosus.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
After 2–3 months, the infection enters an indeterminate stage that can last for decades. Eventually, 10–30 % of infected individuals develop a chronic infection (WHO, 1991). Cardiomyopathy, megacolon and megaesophagus (Oelemann et al., 1998) are frequent complications, and lead to the demise of approximately 50 000 people per year (WHO, 1991).
Although vector to human transmission is the best understood, oral (Shikanai-Yasuda et al., 1991) and congenital transmission (Freilij & Altcheh, 1995), and infections via blood transfusion (Schmunis, 1991) and organ transplantation (CDC, 2002) have also been described. Transfusion-related infection is also a concern in the United States, where an estimated 50 000–100 000 Latin American immigrants are infected with T. cruzi (Kirchhoff, 1993). Transfusion infections in the United States are well documented (Kirchhoff, 1996). Seroepidemiology studies have shown that 1·1 % of blood donors in Los Angeles, California had antibodies against T. cruzi (Kerndet et al., 1991), while Leiby et al. (1997) estimated that 2·5 % of blood donors nationwide could have potentially been exposed to T. cruzi. Out of concern over the risk of future transfusion-related infections, the American Red Cross is preparing to screen all donated blood for T. cruzi, pending a Food and Drug Administration (FDA) approval of a test specific for screening blood donors. It is predicted that such a test will be approved in 2–3 years (McCarthy, 2003).
In order to assist physicians, who may treat either blood donors with positive test results or patients from T. cruzi endemic areas with chronic Chagas' disease, reference laboratories should prepare a testing algorithm for the diagnosis of Chagas' disease. Numerous studies addressing the performance of antibody assays have been published, most of which were carried out in Latin America (Avila et al., 1993; Ferreira et al., 2001; Leiby et al., 2000; Mendes et al., 1997; Oelemann et al., 1998; Reiche et al., 1998). Unfortunately, the assays that were previously evaluated either are no longer being produced by the respective manufacturer or are not available in the United States. In the present study, we evaluated the performance characteristics of three IgG ELISAs, and IgG and IgM indirect immunofluorescent assays (IFAs) that are available in the United States. We studied true T. cruzi antibody positive sera from patients at a Brazilian Chagas' disease clinic and control sera collected at our reference laboratory in Salt Lake City, Utah.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
Group I.
This group contained 30 samples from patients in Recife, Brazil with clinical evidence of T. cruzi infection [i.e. heart failure, syncope or arrhythmias, characteristic electrocardiogram (ECG) abnormalities, residence in a region endemic for T. cruzi or having a family member known to be infected; Table 1
]. Frozen specimens from Oswaldo Cruz University hospital in Recife, Brazil, were shipped frozen and stored at –20 °C until ELISA or IFA testing commenced. Specimens were then stored at 4 °C while the evaluations were performed. Sixteen of the patients were asymptomatic, but resided in areas where Chagas' disease is endemic. Nine patients showed evidence of congestive heart failure. Three patients experienced palpitations and two patients were referred for further evaluation due to syncope (Table 1).
|
Group III. This group contained 53 samples from patients throughout the United States with serological evidence of brucellosis, syphilis, infectious mononucleosis, rheumatoid arthritis, systemic lupus erythematosus (SLE) and rheumatic fever. Ten samples from each disease category were included in the group, except for brucellosis (n=3).
Group IV. This group contained 100 samples collected from random healthy donors in the Salt Lake City, Utah area in 2002.
IgG and IgM IFA.
All samples were tested for IgG and IgM antibodies against T. cruzi using an IFA slide supplied by MarDx Diagnostics. The antigen preparation consisted of epimastigotes of T. cruzi (Corpus Cristi strain), which had been cultured in a modification of a medium originally developed for the culture of Leishmania sp. A standardized amount of washed parasites were placed into the wells of specially prepared microscope slides. In the IgG assay, samples were diluted 1 : 16 in PBS diluent [0·5 % fetal bovine serum, 0·1 % thimerosal (Sigma) in PBS (Bio-Rad)] and added to IFA slides coated with a fixed antigen preparation of T. cruzi epimastigotes. The slides were incubated in a moist chamber at room temperature for 30 min and then washed for 5 min in PBS. Fluorescein-labelled IgG conjugate, supplied by MarDx, was added to each reaction well and incubated at room temperature for 30 min. After this second incubation, the slides were washed in PBS, coverslips were added and the slides examined at x400 magnification (numeric aperture 0·85 mm) by fluorescent microscopy. Results were determined based on the fluorescence of the organisms in each sample well. Fluorescence was graded as recommended by the manufacturer, where a sample was reported positive if a 2+ homogeneous surface fluorescence was observed in >80 % of the parasites in the well. Positive samples were serially diluted and titrated to end point. The reference interval for a positive result was 
1 : 16.
Samples were assayed for IgM antibodies against T. cruzi in the same manner as described above, except that the initial dilution was performed in goat anti-human IgG diluent (Martins et al., 1995) and fluorescein-labelled IgM conjugate supplied by MarDx was added to each reaction well in the conjugation step. Samples were assayed three times on three consecutive days.
Commercial ELISA test systems. Three commercial ELISAs were evaluated in this study: the CeLLabs T. cruzi IgG ELISA (distributed by PANBIO), the FDA-cleared Hemagen Chagas' kit [enzyme immunoassay (EIA) method] (Columbia) and the IVD Research Chagas (T. cruzi) serum microwell ELISA (distributed by Biotech Trading Partners). The Hemagen EIA uses antigens from cultured T. cruzi epimastigotes. The IVD Research and CeLLabs EIAs use unspecified T. cruzi antigens. Testing was performed according to the manufacturers' instructions. Test results were analysed as recommended in the corresponding technical information for each assay. Samples that were discrepant between the assays were repeated in duplicate and reported as the mean A450 value. Two positive samples and one negative sample were assayed three times in three consecutive days to evaluate reproducibility.
Statistical analysis. The clinical sensitivity of each assay was determined as the percentage of positive results with each test kit for the 30 known positive samples from Brazil. The clinical specificity of each assay was determined as the percentage of negative results with each test kit for the group IV samples (random healthy donors representing true negatives, n=100). The cross-reactivity of each assay was examined by testing sera from patients with other parasitic, bacterial and viral infections as well as a variety of autoimmune diseases (groups II and III). The 95 % confidence intervals for the clinical sensitivity and specificity were also determined for each assay.
Reproducibility of T. cruzi assays. The reproducibility of each assay was examined using serologically positive and negative sera repeatedly tested over 3 days. For the IgG IFA, one negative and two positive samples were assayed three times in duplicate. For the IgM IFA, one positive and two negative samples were assayed three times in duplicate. For the commercial ELISA test systems, two positive samples were assayed three times in replicates of five on three consecutive days. Where possible, inter-assay and intra-assay precision was determined.
|
RESULTS AND DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
. The ages of the patients ranged from 25 to 71 (mean 49·1) years. Fifty-three per cent of the patients (16) were female. Surface ECG revealed right bundle branch block in eight patients and left anterior fascicular block was identified in two patients, one of whom also had a right bundle branch block. Findings of ventricular tachycardia, third degree atrioventricular block, bigeminy and atrial fibrillation were all identified in one patient. Fifteen patients had normal ECGs.
Among patients aged 49 years and younger, 14 % (2/14) had abnormal ECG findings or clinical symptoms of Chagas' disease, while 81 % (13/16) of patients aged 50 years and older had abnormal ECG findings and 75 % (12/16) showed clinical symptoms of Chagas' disease. Eighty-eight per cent (14/16) of asymptomatic patients had normal ECG findings (Table 1).
Clinical sensitivity of the IgG assays
The three commercial ELISA test systems each reported all 30 samples of group I as positive, resulting in a calculated sensitivity of 100·0 % for each assay. The MarDx IgG IFA reported 28/30 samples as positive, resulting in a calculated sensitivity of 93·3 % (Table 2).
|
Clinical specificity
Clinical specificity was determined using normal sera from individuals in a non-endemic area. The determination of specificity did not include patients with infections or immunological abnormalities that might affect the true clinical specificity. Both the CeLLabs and the Hemagen ELISAs reported all 100 samples of group IV as negative, resulting in a specificity of 100·0 %. The IVD Research ELISA reported 7 samples as positive and 93 samples as negative, resulting in a specificity of 93·0 %. The MarDx IgG IFA reported 1 sample as positive and 99 samples as negative, resulting in a calculated specificity of 99·0 % (Table 2). Although the MarDx IgM IFA had a calculated specificity of 100·0 % using normal sera, the true clinical sensitivity for the MarDx IgM IFA could not be determined with the samples used in this study.
Cross-reactivity studies
The CeLLabs ELISA showed the least cross-reactivity with sera positive for antibodies directed against other parasites, infectious agents or autoantigens. Two samples previously found to be positive for antibodies against Leishmania were reported as positive by the CeLLabs T. cruzi ELISA, at just above the positive A450 cutoff of 0·30 (A450 0·36 and 0·31). Both Leishmania-serology positive samples and three of the ten SLE-serology positive samples had positive results by the Hemagen ELISA. One of the two Leishmania-serology positive, one of the ten Toxoplasma-serology positive and one of the ten SLE-serology positive samples had positive results with the IVD Research ELISA. The one sample positive for Leishmania antibodies that was reported as positive by the IVD Research EIA was just above the positive A450 cutoff of 0·20 (A450 0·21). The MarDx IgG IFA showed the most cross-reactivity with other antibody-positive samples. One malaria, one Schistosoma, two Leishmania, two Toxoplasma, one rheumatoid factor, one SLE- and one DNase B- (rheumatic fever) serology positive sample had positive results in the MarDx IgG IFA (Table 3). No cross-reactivity was observed with any of the test systems from the following infection categories: Cysticercosis (n=7), Strongyloides (n=3), Brucella (n=3), Toxocara (n=2), E. histolytica (n=3), Leptospira (n=1), Trichinella (n=1), syphilis (n=10) or infectious mononucleosis (n=10).
|
Comparison of the ELISA test systems
Quantitative ELISA results (A450 readings or calculated index values) were subdivided into groups I–IV as defined previously (Fig. 1). The IVD Research ELISA showed the greatest range of A450 results, from 0·0 to 3·16. The CeLLabs ELISA showed the second greatest A450 range, from 0·0 to 2·04. The Hemagen ELISA showed the narrowest range of results, with index values from 0·0 to 1·07. When the manufacturers' recommended cut-off values were employed, the CeLLabs assay showed the best overall results. The results for group I samples were separated from the results of groups II to IV samples, with the exception of the two Leishmania samples that were above the positive result cut-off.
|
). Skeiky et al. (1992) pointed out similarities between proteins produced by patients infected with T. cruzi and SLE, and cross-reactivity between T. cruzi and Leishmania samples in antibody assays is well documented (Chiller et al., 1990; Kirchhoff, 1996; Reiche et al., 1998; Solana et al., 1995). Due to the low specificity in EIAs, Latin American laboratories are instructed to test samples from chagasic patients using at least two methodologies (Mocayo, 1992; Schmunis, 1991). By coupling several assays with acceptable sensitivities and specificities, it is hoped that any antibody-positive and thus T. cruzi-positive patients will be accurately identified, and thus precluded from donating blood. IFA is commonly used in Latin American laboratories as one of the two tests for Chagas' disease serological diagnosis (Oelemann et al., 1998). However, it is generally less sensitive and more subjective than ELISA tests. A variety of fluorescence patterns from cross-reacting antibodies make IFA interpretation very challenging and can result in false-positive results.
The IgM IFA is the only commercial assay that is available for detection of IgM-specific T. cruzi antibodies. A positive IgM IFA result is indicative of acute infection (Camargo & Amato Neto, 1974; Reyes et al., 1990; Umezawa et al., 1996a, b, c), and hence an IgM assay could be an important diagnostic tool. But it is suspected that the antigens used in most IgG assays may not be specific enough for IgM assays, which would lower the overall sensitivity (Umezawa et al., 1996c). A negative result by the MarDx IgM IFA could mean the IgM antibody response has faded over time, the IFA is not sensitive enough to detect the low amount of IgM that may be present, or the assay is not manufactured using the best combination of antigens. As discussed by Umezawa et al. (1996c), purified surface trypomastigote antigens used in ELISAs may be both sensitive and specific enough to diagnose acute T. cruzi infections. Additionally, clinical diagnosis of acute infection is uncommon due to the asymptomatic course of the disease, and by the time chronic infection is evident, IgM antibody has been absent for years or decades. An exception has been in cases of transfusion-related infection. When immunosuppressed patients receive T. cruzi-contaminated blood products, a fulminant illness usually results (Kirchhoff, 1996) and symptoms are easily recognized. Although IgM testing has its place in laboratory diagnosis of acute T. cruzi infection and is useful for the diagnosis of congenital infection, a better overall approach is in testing paired acute and convalescent samples for IgG antibodies. The majority of T. cruzi antibody/antigen relationship research has been performed using IgG antibodies, thus IgG assays have been extensively developed and generally perform better. Some patients remain seronegative all their lives (Camargo, 1992), however, so for diagnosis, immunological assays should always be coupled with ECG data and clinical symptoms. Further research into IgM assays should be pursued so that the best possible product can be made available.
Confirmation of a positive serology result for T. cruzi is still difficult. Xenodiagnosis and haemoculture are highly impractical (Leiby et al., 2000). The radioimmunoprecipitation assay appears to be very promising and specific (Leiby et al., 2000), as does Western blotting (Reiche et al., 1998; Solana et al., 1995; Umezawa et al., 1996c), although these assays are not available commercially. PCR is more sensitive than xenodiagnosis (Avila et al., 1993) and microscopic examination (Kirchhoff, 1996). However, a negative PCR result does not necessarily exclude infection. Because the parasitaemia is very low in chronic infection, it is unlikely that PCR will be useful in the chronic stage.
Unlike the Hemagen assay, which is FDA approved for diagnostic purposes only (not for blood donor screening), the IVD IgG ELISA and the CeLLabs IgG ELISA are labelled as analyte-specific reagents (ASR). The clinical reference ranges of ASR assays may be adjusted after assessment of the performance characteristics according to the FDA (Jaros, 1992; Litwin, 2000; Paxton, 1998). Therefore, if cut-off values are adjusted based on the results collected for the samples in groups I to IV, the IVD Research ELISA outperformed the Hemagen and CeLLabs ELISAs. The IVD Research ELISA showed the greatest separation between positive and negative results; samples from group I had results that ranged from A450 1·71 to 3·16, while samples from groups II to IV had results that ranged from A450 0·0 to 1·0. The Hemagen ELISA showed the highest values for cross-reactivity. If an index value of greater than 0·50 was obtained in the Hemagen ELISA, it would not be possible to confidently report the result as positive for T. cruzi antibody versus a cross-reactive antibody. If an A450 of greater than 1·20 was obtained in the IVD Research ELISA, however, the sample had a high likelihood of being a true positive T. cruzi result (Fig. 1). Therefore, if the cut-off value for this ELISA is changed to 1·20, the agreement and specificity of the assay would change from 94·6 % and 93·0 %, respectively, to 100 % for both.
It is possible that the manufacturer's cut-off value of 0·20 for the IVD IgG ELISA may have been based on studies that included weakly positive serum samples from patients with Chagas' disease, whereas our dataset may have included more strongly positive sera. Taken together, we think it might be useful to include an equivocal A450 range of 0·2–1·19. Further studies with a higher number of previously classified sera (high, medium and low antibody titre) will be needed to refine this range. An equivocal result would indicate to the clinician the possibility of low level Chagas' antibodies, cross-reactivity with another parasite or a non-specific interfering substance.
All of the assays presently available in the United States have an adequate ability to discriminate positive and negative tests. The IVD Research ELISA showed the best separation between positive and negative results, and the least cross-reactivity, using an adjusted cut-off value. Based on the clear separation between positive and negative serum samples, an equivocal range encompassing results between the recommended manufacturer's cut-off and the adjusted cut-off suggested would more likely detect antibodies that are cross-reactive with T. cruzi, such as seen in leishmaniasis, rather than true positives. Overall the CeLLabs assay had the best performance of the three kits when following the manufacturer's instructions and established cut-offs.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTS AND DISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  Z. C. Caballero, O. E. Sousa, W. P. Marques, A. Saez-Alquezar, and E. S. Umezawa Evaluation of Serological Tests To Identify Trypanosoma cruzi Infection in Humans and Determine Cross-Reactivity with Trypanosoma rangeli and Leishmania spp. Clin. Vaccine Immunol., August 1, 2007; 14(8): 1045 - 1049. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
, Group I; 
, group II; 
, group III; 
, group IV.