DIAGN MICROBIOLINFECTDIS 1992;15:587-593
Performance Characteristics of a Commercial Antibody-Capture Enzyme Immunoassay for Detection of Toxoplasma-Specific IgM Antibodies David R. Gretch, John J. Warren, Rose Mary Bacina, Eric D. Stefansson, and Thomas R. Fritsche
Antibody-capture enzyme immunoassay (EIA) for the detection of Toxoplasma-specific IgM has been shown to provide significantly higher specificity than the indirect IgM EIA. A new commerically available antibody-capture EIA (PLATELIA Toxo IgM EIA) converted 99 out of 100 false-positive Toxoplasma IgM determinations to true negative. Experiments using Toxoplasma IgM calibrators demonstrated the antibodycapture EIA is approximately eightfold more sensitive than a new automated microparticle EIA for Toxoplasma IgM antibodies (IMX Toxo IgM EIA), and approximately equal in sensitivity to the indirect immunofluorescence assay. Precision studies of the antibody capture EIA using low, medium, and
high calibrators gave coefficients of variation ranging from 3.0%-4.5% for within-run and 5.2%-11.4% for run-to-run variation experiments. Interference from high levels of bilirubin, albumin, hemoglobin, and lipid was not detected. Sera from patients with inflammatory or infectious disorders were tested for interference in the antibody-capture EIA. False-positive Toxoplasma IgM results were not observed, but lowlevel negative interference was detectable when patient sera was mixed with Toxoplasma-positive sera. Preparations of purified human IgM also produced negative interference in the antibody-capture EIA for Toxoplasma IgM.
USA is seropositive for Toxoplasma antibodies, based on early studies (Feldman and Miller, 1956). Most Toxoplasma infections are mild or asymptomatic in otherwise healthy individuals, but the protozoan is an important pathogen in predisposed populations. Acute infection in pregnant w o m e n can lead to birth defects (Desmonts and Couvreur, 1974), and in patients with l y m p h a d e n o p a t h y syndromes, Toxoplasma infection must be distinguished from lymphoproliferative disorders (McCabe et al., 1987). Infection is most severe in immunocompromised patients where the disease can be fatal. Whether i m m unocom prom i se d because of chemotherapy, organ transplantation, or infection with the h u m a n immunodeficiency virus
Toxoplasma gondii is an intestinal coccidian parasite of felines capable of producing extraintestinal lesions in a wide variety of warm-blooded hosts. From 30% to 40% of the adult population in the From the Department of LaboratoryMedicine, Universityof Washington Medical Center, Seattle, Washington, USA. Address reprint requests to Dr. T R. Fritsche, Department of LaboratoryMedicine, SB-10, Universityof Washington Medical Center, Seattle, WA 98195, USA. Received 16 September 1991; revised and accepted 9 December 1991. © 1992 Elsevier SciencePublishing Co., Inc. 655 Avenue of the Americas, New York, NY 10010 0732-8893/92/$5.00
(HIV), this group now represents the majority of patients being tested for evidence of Toxoplasmainfection. Serology has become the preferred method of diagnosis because of difficulties in obtaining and processing specimens for culture and in interpreting histologic findings (Karin and Ludlam, 1975; Brooks et al., 1987). The Sabin-Feldman Dye Test described in 1948 remains the gold standard of Toxoplasma serologies (Sabin and Feldman, 1948). It is limited in its usefulness, however, because of the requirement for live infective tachyzoites and the labor and expertise required in its performance. Other assays detecting total antibody, including complement fixation (CF) and indirect hemagglutination (IHA), also have certain shortcomings (Welch et al., 1980). High CF titers may persist for long periods (>12 months) in some patients, and seroconversion is delayed with IHA (4-10 weeks). Indirect immunofluorescent antibody tests (IFA) and enzyme immunosorbant assays (EIA) are popular methodologies for Toxoplasma serodiagnosis in the clinical laboratory because they combine improved accuracy with ease of performance. Whereas IgG assays generally perform well by these methods, tests for Toxoplasma-specificIgM antibodies are problematic. IFA tests for IgM antibodies can be falsely positive due to the presence of antinuclear antibodies and rheumatoid factor (Araujo et al., 1971), and falsely negative in the presence of Toxoplasma-specific IgG (Pyndiah et al., 1979; Filice et al., 1980; McCabe et al., 1987). Problems with false-positive Toxoplasma IgM results have also been reported by investigators using indirect EIA methods (van Loon et al., 1983; Rotmans et al., 1988). Many of these problems have been overcome with the application of the antibody-capture principle to EIA testing for Toxoplasma-specific IgM (Noat et al., 1980; Wielaard et al., 1983). Few commercial antibody-capture IgM EIAs have been marketed, possibly because of difficulty in performing and controlling the test and high cost. Recent studies at our institution have revealed an unacceptably high number (18% of total requests) of false positive Toxoplasma IgM results using a commercially available indirect EIA. Many of these results have occurred with transplant or other immunocompromised patients at high risk for toxoplasmosis, causing considerable concern. This situation has lead us to reexamine our testing strategy and consider changing to an assay with improved specificity. Two new assays recently marketed in the United States with FDA approval for Toxoplasma serodiagnosis include an antibody-capture EIA manufactured by Sanofi Diagnostics-Pasteur and an automated microparticle EIA manufactured by Abbott Laboratories, the latter receiving ~Ycpll~nt reviews in a clinical evaluation (Schaefer
D.R. Gretch et al.
et al., 1989). This report demonstrates the clinical utility of these two ToxoplasmaIgM assays with our patient population and presents additional performance characteristics (precision and interference studies) of the antibody-capture EIA.
MATERIALS AND METHODS Serodiagnostic Kits Commerical serodiagnostic kits evaluated for the detection of Toxoplasma IgM were PLATELIAToxo IgM EIA (Sanofi Diagnostics Pasteur, Marnes-la-Coquette, France), IMX Toxo IgM EIA (Abbott Laboratories, Abbott Park, IL), SIA Toxo M EIA (Sigma, St. Louis, MO), and an IgM indirect immunofluorescence assay (Electro-Nucleonics, Columbia, MD). The PLATELIA Toxo IgM kit utilizes the antibodycapture (double sandwich) EIA principle. Briefly, patient sera is added to microtiter wells coated with an anti-human IgM antibody. Captured Toxoplasmaspecific IgM is detected using a purified Toxoplasma antigen (P30) followed by an anti-P30 monoclonal antibody conjugate. The IMX Toxo IgM kit is a microparficle EIA (modified indirect EIA). For this assay, microparticle beads coated with Toxoplasmaantigen lysate are mixed with patient sera. Bound Toxoplasma-specificIgM is detected with an anti-human IgM conjugate, and the rate of fluorescent product formation is measured (Fiore et al., 1988). The SIA Toxo M EIA is a standard indirect EIA, employing antigen-coated microtiter plates. The indirect immunofluorescence assay (IFA) utilizes whole Toxoplasmaorganisms fixed on microscopic slides and reacted with patient sera, followed by incubation with fluorescein-conjugated goat anti-human IgM with Evans Blue counterstain. All assays were performed as recommended by the manufacturer. Toxoplasma IgG determinations were performed using the Sigma SIA Toxo G EIA kit.
Specimens All sera were obtained from the clinical laboratories of the University of Washington Medical Center and originated from patients either at risk for toxoplasmosis or with symptoms suggestive of that infection. Of 815 requests for Toxoplasma IgM determination, 158 were positive or equivocal by the SIA solid-phase EIA. A designation of equivocal is the manufacturer's interpretation for EIA results falling between positive and negative cutoff values. All positive and equivocal specimens received confirmatory testing in our laboratory and/or a reference laboratory (Palo Alto Medical Research Foundation, Palo Alto, CA), and all patients were evaluated clinically. Specimens were judged to be false-positive for ToxoplasmaIgM if they met the following criteria: (a) any optical den-
Antibody-Capture EIA for
sity greater than the negative cutoff value in the SIA Toxo M EIA, (b) no significant changes in Toxoplasma IgG titers after a minimum of three monthly determinations, (c) negative results in a confirmatory test for Toxoplasma IgM, and (d) lack of clinical criteria for toxoplasmosis such as lymphadenopathy, persistent fever, and/or myalgia. Confirmatory tests included biopsy when appropriate, the Sabin-Feldman dye test, and IFA, the latter demonstrating >95% agreement with the dye test (Walton et al., 1966). Adequate volume for multiple testing was found in 100 specimens with false-positive ToxoplasmaIgM results, and the specimens were stored at - 20°C (constant temperature) until needed. Of the 100 falsepositive specimens analyzed, 23 were positive and 77 were equivocal in the SIA solid-phase EIA for ToxoplasmaIgM; 71 false-positive specimens were obtained from 27 different heart transplant recipients, making this patient population the predominant source of false-positive results. The maximum number of specimens from one individual patient was ten; 18 false-positive specimens were obtained from 14 different obstetrical patients. One false-positive specimen was obtained from each of eight different patients with acquired immunodefficiency syndrome and three patients with fever of unknown origin. Repeat specimens were obtained at intervals required by transplant protocols (usually 1 month), or at 2-4 week intervals to detect seroconversion. Other specimens, stored in a similar fashion, included convalescent sera from patients with cytomegalovirus infection, rheumatoid arthritis, Sjogren's syndrome, systemic lupus, infectious mononucleosis, Lyme disease, and syphillis. Commercially prepared ToxoplasmaIgG and IgM calibrators were obtained from Abbott Laboratories and Sanofi Diagnostics Pasteur and stored as recommended by the manufacturers.
Analytical Sensitivity Studies Analytical sensitivities of the antibody-capture EIA, microparticle EIA, and indirect immunofluorescence assays were compared as follows: commercially prepared ToxoplasmaIgM calibrators or sera from a patient with acute toxoplasmosis were subjected to serial twofold dilutions with a buffer appropriate for each assay. Each dilution was then treated as a separate mock clinical sample for the various assays. The samples were immediately analyzed and results scored as positive or negative.
ation was determined using 20 repetitive samples assayed per plate per experiment. Run-to-run variation was determined using calibrators that were aliquoted and stored at -20°C. Samples with medium and high IgM levels were analyzed in 20 different assays over a 6-week period. The low calibrator (IgM cutoff standard) was assayed ten times during the same interval.
Interference Studies Interference studies were performed on the antibody-capture EIA using Toxoplasma IgM calibrators mixed one-to-one with either phosphate-buffered saline (PBS); normal sera (seronegative for Toxoplasma IgG and IgM), sera from patients seropositive for Toxoplasma IgG but seronegative for Toxoplasma IgM, convalescent sera from patients with a variety of inflammatory or infectious diseases, purified human IgM (seronegative for Toxoplasma), or other physiologic solutions. ToxoplasmaIgG and IgM levels were determined for the specimens before and after mixing. Purified human IgM was obtained from Cal Biochem (San Diego, CA). Other reagents included human albumin (Sigma), bilirubin (National Bureau Standard 916a, Gaithersburg, MD), Interlipid (Cutter Medical, Berkeley, CA), and hemoglobin (obtained by lysis of human erythrocytes), kindly provided by the clinical laboratories of the University of Washington Medical Center.
RESULTS IgM Testing on False-Positive Clinical Samples A total of 100 specimens repeatedly found to be false positive for ToxoplasmaIgM antibodies by using the SIA solid phase EIA were retested with the antibodycapture EIA and the microparticle EIA. The results of this retesting are summarized in Table 1. Both the antibody-capture and the microparticle EIAs eliminated a majority of the false-positive results obtained with the solid-phase assay. The antibody-capture EIA converted 99% of the false-positive results to true negatives, whereas the microparticle EIA converted 70% of false positives to true negatives. Thus, both the antibody-capture and microparticle EIAs displayed superior specificity than the solid-phase EIA using the problem specimens encountered in our laboratory. Overall, the antibody-capture EIA outperformed the microparticle EIA with these specimens.
Precision Studies Precision studies were performed on the antibodycapture EIA using commercially prepared Toxoplasma IgM calibrators diluted to appropriate levels within the linear range of the assay. Within-run vari-
Analytical Sensitivity of Commercial EIAs versus Indirect Immunofluorescence Clinical utility of the antibody-capture and microparticle EIAs was further investigated by examining
D.R. Gretch et al.
TABLE 1 Analysis of 100 False-Positive a Toxoplasma IgM Specimens by PLATELIA Antibody-Capture EIA or IMX Microparticle EIA Test Result
Converted to true negative Remained false positive Equivocal
99/100 (99%) 0/100 (0%) 1/100 (1%)
70/100 (70%) 20/100 (20%) 10/100 (10%)
aFalse positives using SIA Toxo-Msolid-phase EIA.
their ability to detect low levels of Toxoplasma IgM when compared with the indirect immunofluorescence assay. The importance of this analysis is that low levels of Toxoplasma IgM may be clinically significant in certain situations, including congenital disease, early seroconversion, or immunodeficiency states. For these experiments, commercial Toxoplasma IgM calibrators and patient serum were subjected to twofold serial dilutions with appropriate buffers and immediately assayed by the three methods. The results are summarized in Table 2. In all assays, the antibody-capture EIA compared favorably with the IFA in terms of analytical sensitivity. Both assays were 8- to 16-fold more sensitive in detecting Toxoplasma IgM than the microparticle EIA.
identical samples were assayed in 20 separate experiments to examine consistency among different lots of EIA kits. Coefficients of variation (CVs) ranged from 3.0% to 4.5% for within-run and 5.2% to 11.4% for run-to-run variation experiments. A low positive IgM calibrator at the threshold of positivity for the antibody-capture EIA was read as positive in 10 of 10 separate assays, with a CV of 11.4%, slightly higher than other calibrator preparations (5.2% to 6.6%).
The above studies suggest that the antibody-capture EIA has definite advantages in terms of clinical specificity and analytical sensitivity for the detection of Toxoplasma IgM. Therefore, the antibody-capture EIA kit was evaluated further. Precision of this test was evaluated in a series of replication experiments using commercially prepared Toxoplasma IgM calibrators adjusted to low, medium, and high levels within the linear range of the assay. For within-run variation studies, 20 or more identically prepared samples were assayed per experiment to examine uniformity of the microtiter plates. For run-to-run variation studies,
Interference experiments were performed to investigate the performance of the antibody-capture EIA under abnormal physiologic states. Albumin (10 g/dl), bilirubin (10 mg %), lipid (10%), and hemoglobin (150 mg/dl) had no appreciable effect on the performance of the antibody-capture EIA in the presence or absence of commercial ToxoplasmaIgM calibrators (data not shown). In addition, multiple sera from patients with very high Toxoplasma IgG titers were tested for interference in the Toxoplasma IgM antibody-capture assay. These sera produced no falsepositive Toxoplasma IgM results and no detectable negative interference in the antibody-capture EIA (data not shown). Other interference studies were performed to assess the detection of ToxoplasmaIgM calibrators in the presence of convalescent sera from patients with a variety of inflammatory or infectious diseases, including lupus, rheumatoid arthritis, Sjo-
Analytical Sensitivity of Three Methods for Determination a
Patient serum b TOXOIgM calibrator A Toxo IgM calibrator B
aResultsexpressed as the highest dilution of calibrator IgM or patient sera giving a positive result. bA 29-year-oldman with acute toxoplasmosis. CNotdone.
Antibody-Capture EIA for Toxoplasma IgM
gren's syndrome, Lyme disease, infectious mononucleosis, syphillis, and cytomegalovirus infection. None of the convalescent sera tested gave false-positive results in the antibody-capture EIA. However, a detectable level of negative interference was observed when the Toxoplasma IgM calibrators were assayed following dilution with sera from patients with several different inflammatory or infectious disorders and results compared with controls consisting of calibrators diluted with either buffer (PBS) or normal patient sera (data not shown). The degree of interference was consistent through several experiments using different preparations and concentrations of sera and calibrators.
Human IgM Interference in the AntibodyCapture EIA The antibody-capture EIA method is theoretically vulnerable to interference by increased levels of nonspecific IgM competing for a limited number of IgM binding sites on the solid phase. This competition may explain the interference observed when testing sera from patients with inflammatory or infectious disorders. To test this hypothesis, a Toxoplasma IgM calibrator and sera from a patient with acute toxoplasmosis were assayed in the presence or absence of purified h u m a n IgM lacking Toxoplasmareactivity, as shown in Figure 1. Negative interference was observed when the undiluted patient sera was assayed after adding nonspecific human IgM to a final concentration of 100 mg/dl or greater. Interference was
more pronounced when the Toxoplasma IgM calibrator and dilutions of calibrator or patient sera were assayed in the presence of purified human IgM. Test results with these specimens became negative at nonspecific IgM final concentrations of 10 to 100 mg/dl (see Figure 1).
DISCUSSION For many clinical laboratories, false-positive IgM determinations continue to confound the serological diagnosis of infectious disorders, including acute toxoplasmosis. False-positive Toxoplasma IgM determinations potentially have serious consequences, including inappropriate termination of pregnancy in cases of suspected congenital toxoplasmosis. The recent emergence of opportunistic infections secondary to organ transplantation, clinical immunosuppression, and AIDS has necessitated the evaluation of new methodologies for accurate Toxoplasma serodiagnosis. The current study of two new commercial EIAs for the detection of Toxoplasma IgM confirms previous reports (Noat et al., 1980; Rotmans et al., 1988) that the antibody-capture EIA offers improved specificity compared to solid-phase EIAs. Differences in specificity of the microparticle EIA and SIA-EIA (both solid-phase assays) were also observed and may be due to differences in effective surface area of the respective solid phases (Fiore et al., 1988), different wash stringency, or differences in solid-phase an-
1.6 1.4 1.2
~- Patient sera -~ Pt. sera at t:20 dilution
1 OD 492nm
~>- IgM calibrator
-x-IgM calibrator at 1:20 dilution
0.2 0 0
mg/dL human IgM
FIGURE 1 Effects of nonspecific human IgM on the antibody-capture EIA for Toxoplasma IgM. Patient serum positive for ToxoplasmaIgM and ToxoplasmaIgM calibrators were tested directly or at 1:20 dilutions with increasing concentrations of purified human IgM.
tigen. A second finding in our studies is improved analytical sensitivity for the antibody-capture EIA relative to the modified solid-phase EIA. Finally, the commercial antibody-capture EIA kit we evaluated performed very well in precision and interference studies, with the exception being negative interference by elevated nonspecific human IgM. Solid-phase Toxoplasma IgM EIAs employ antigencoated microtiter (or microparticle) surfaces, which are capable of binding specific IgG as well as IgM, leading to interference by IgG. This interference may occur when IgG is b o u n d by rheumatoid factor, or when IgG competes for antigen binding sites. The IgM antibody-capture EIA design has two fundamental advantages that reduce interference: (a) total IgM, but not IgG, is captured in the initial step, which reduces the likelihood of IgG-mediated interference; and (b) the IgM detection step is antigen specific, thereby eliminating nonspecific IgM (such as rheumatoid factor). In contrast, both specifically bound IgM and, occasionally, nonspecifically bound IgM are detected in conventional solid-phase EIAs, which explains the occurrence of some false-positive results. Using an in vitro analytical approach, we demonstrate that negative interference by nonspecific human IgM can occur with the antibody-capture EIA. However, a very high level of human IgM (1000 mg/dl) was required to convert a positive result to negative when sera from a patient with documented acute toxoplasmosis was studied. Lower levels of
D.R. Gretch et al.
nonspecific human IgM (100-500 mg/dl) were also able to interfere with the detection of low level Toxoplasma IgM in our experiments. Considering the increased analytical sensitivity of the antibody-capture EIA, negative interference by nonspecific human IgM is likely to be insignificant except perhaps in unusual cases where total IgM is markedly elevated (for example, Waldenstrom's macroglobulinemia). This potential limitation of the antibody-capture EIA is most likely due to competition for IgM binding sites during the initial IgM capture step of the assay and should not be ignored. In summary, the current study confirms the utility of the antibody-capture EIA in the serodiagnosis of acute toxoplasmosis. The commercial antibody-capture EIA kit manufactured by Sanofi Diagnostic-Pasteur performed exceptionally well in our evaluation of false-positive Toxoplasrna IgM specimens. In addition, the kit displayed excellent precision characteristics and appears well suited for large-volume clinical diagnostic testing. The test procedure is slightly more difficult to perform than conventional EIAs and the test is more expensive. The decreased need for repeat and confirmatory testing may balance the increased costs, however. Future clinical studies of this methodology in comparison with other new technologies, such as Toxoplasma antigen-detection assays (Asai et al., 1987) and polymerase chain reaction (Burg et al., 1989) are required to improve further our ability to diagnose toxoplasmosis.
REFERENCES Araujo FG, Barnett FV, Gentry LO, Remington JS (1971) False Positive anti-Toxoplasma fluorescent antibody test in patients with antinuclear antibodies. Appl Microbiol 22:270-275. Asai T, Kim T, Kobayashi M, Kojima S (1987) Detection of nucleoside triphosphate hydrolase as a circulating antigen in sera of mice infected with Toxoplasma gondii. Infect Immun 55:1332-1335. Brooks RG, McCabe RE, Remington JS (1987) Role of serology in the diagnosis of toxoplasmic lymphadenopathy. Rev Infect Dis 9:1055-1062. Burg JL, Gover CM, Pouletty P, Boothroyd JC (1989) Direct and sensitive detection of a pathogenic protozoan, Toxoplasma gondii, by polymerase chain reaction, l Clin Microbiol 27:1787-1792. Desmonts G, Couvreur J (1974) Congenital toxoplasmosis: a prospective study of 378 pregnancies. N Engl J Med 290:1110-1116. Feldman AA, Miller LT (1956) Serologic study of toxoplasmosis prevalence. Am ] Hyg 64:320-335. Filice GA, Yeager AS, Remington JS (1980) Diagnostic significance of immunoglobulin M antibodies to Toxoplasma gondii detected after separation of immunoglob-
ulin M from immunoglobulin G antibodies. J Clin Microbiol 12:336-342. Fiore M, Mitchell J, Doan T, Nelson R, Winter G, Grandone C, Zeng K, Haraden R, Smith J, Harris K, Leszczynski J, Berry D, Safford S, Barnes G, Scholnick A, Ludington K (1988) The Abbott IMx automated benchtop immunochemistry analyzer system. Clin Chem 34:1726-1732. Karin KA, Ludlam GB (1975) The relationship and significance of antibody titers as determined by various serologic methods in glandular and ocular toxoplasmosis. J Clin Pathol 28:42-49. McCabe RE, Brooks RG, Dorfman RF, Remington JS (1987) Clinical spectrum in 107 cases of toxoplasmic lymphadenopathy. Rev Infect Dis 9:754-774. Naot Y, Remington JS (1980) An enzymeqinked immunosorbent assay for detection of IgM antibodies to Toxoplasma gondii: use for diagnosis of acute acquired toxoplasmosis. J Infect Dis 142;757-766. Pyndiah N, Kreech U, Price P, Wilhelm J (1979) Simplified chromatographic separation of immunoglobulin M from immunoglobulin G and its application for Toxoplasma indirect immunofluorescence. J Clin Microbiol 9:170-174. Rotmans J, Duchenne W, Linschooten C, In't Veld N,
A n t i b o d y - C a p t u r e EIA for Toxoplasma IgM
Polderman A (1988) Comparative study of three immunoassays for detection of immunoglobulin M antibodies against Toxoplasma gondii. Eur J Clin Microbiol Infect Dis 7:535-538. Sabin AB, Feldman HA (1948) Dyes as microchemical indications of a new immunity phenomenon affecting a protozoan parasite (Toxoplasma). Science 108:660-663. Schaefer LE, Dyke JW, Meglio FD, Murray PR, Crafts W, Niles AC (1989) Evaluation of microparticle enzyme immunoassays for immunoglobulins G and M to rubella and Toxoplasma gondii on the Abbott IMx automated analyzer. ] Clin Microbiol 27:2410-2413. Van Loon AM, van der Logt JTM, Heesen FWA, van der Veen J (1983) Enzyme-linked immunosorbent assay that uses labeled antigen for detection of immunoglobulin M and A antibodies to toxoplasmosis: comparison with
indirect immunofluorescence and double-sandwich enzyme-linked immunosorbent assay. J Clin Microbiol 17:997-1004. Walton BC, Benchoff BM, Brooks WH (1966) Comparison of the indirect fluorescent antibody test and methylene blue dye test for detection of antibodies to T. gondii. Am ] Trop Med Hyg 15:149-152. Welch PC, Masur H, Jones TC, Remington JS (1980) Serologic diagnosis of acute lymphadenopathic toxoplasmosis. ] Infect Dis 142:256-264. Wielaard F, van Gruijthuijsen H, Duermeyer W, Joss AW, Skinner L, Williams H, van Elven EH (1983) Diagnosis of acute toxoplasmosis by an enzyme immunoassay for specific immunoglobulin M antibodies. J Clin Microbiol 17:981-987.