New Criteria for Immunofluorescence Assay for Q Fever Diagnosis in Japan
http://www.100md.com
微生物临床杂志 2005年第11期
Laboratory of Rickettsia and Chlamydia
Department of Virology I, National Institute of Infectious Diseases, 1-23-1 Toyama, Shinjuku-Ku, Tokyo 162-8640, Japan
Laboratory of Pathology, Department of Parasitology and Pathology, Faculty of Veterinary Medicine, Bogor Agricultural University, Jl. Agatis, Kampus IPB Darmaga Bogor 16680, Indonesia
ABSTRACT
A study was made to evaluate the cutoff value of indirect immunofluorescent-antibody (IFA) test for Q fever diagnosis in Japan. We used 346 sera, including 16 from confirmed Q fever cases, 304 from Japanese pneumonia patients, and 26 from negative cases. Thirteen sera from the confirmed Q fever cases with an immunoglobulin M (IgM) titer of 1:128 and/or IgG titer of 1:256 by the IFA test were positive by both enzyme-linked immunosorbent assay (ELISA) and Western blotting assay (WBA), whereas 298 sera from pneumonia patients and 26 negative sera with an IgM titer of 1:16 and an IgG titer of 1:32 by the IFA test were negative by both ELISA and WBA. In the proposed "equivocal area," with an IgM titer of 1:32 and 1:64 and/or an IgG titer of 1:64 and 1:128, we found 9 sera, 3 from confirmed Q fever cases and 6 from Japanese pneumonia patients, by the IFA test. Three sera from the confirmed Q fever cases and one of the sera from pneumonia patients were IgM and/or IgG positive by both ELISA and WBA. These results suggest that a single cutoff value for the IFA test may cause false-positive and false-negative results. In conclusion, this study showed that an "equivocal area" should be used for the IFA test rather than a single cutoff value and that sera in the equivocal area should be tested by additional serological assays for confirmation.
INTRODUCTION
Q fever is a zoonosis caused by Coxiella burnetii, an obligate intracellular rickettsial organism. The clinical manifestations of Q fever are readily divided into acute and chronic forms. The most common clinical presentation of acute Q fever in human is an influenza-like illness, often accompanied by pneumonia. The chronic Q fever form, particularly endocarditis, may appear several years after the primary episode (1, 2, 6, 21). Because the clinical presentation of the infection is not specific, serological confirmation is required for the diagnosis of Q fever. In Japan, the currently used serological method is the indirect immunofluorescent-antibody (IFA) test, and more recently, some attempts have been made to use other methods, such as the enzyme-linked immunosorbent assay (ELISA).
The unique characteristic of C. burnetii is its antigenic phase variation (17). Virulent phase I can be isolated from natural infection of humans or from laboratory infections of animals. Phase II develops during serial passage in an immunologically incompetent host, such as cell cultures or fertilized eggs (1). Serologically, anti-phase I antibodies are normally found at high levels only during the chronic form of the disease, whereas specific anti-phase II antibodies predominate primarily in acute Q fever (14).
The IFA test has previously been used to detect immunoglobulin M (IgM) antibodies in the sera of Q fever patients within the first 2 weeks of illness (8, 12). The estimation of anti-C. burnetii IgM antibody using the IFA test in a single serum sample has been proven useful in confirming acute infection in humans (8). The IFA method is more sensitive and specific than the complement fixation test; however, it is less sensitive than the ELISA (15). Therefore, the validation of the immunofluorescent threshold value for Q fever serology should be important for the establishment of the diagnosis.
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by immunoblotting assay, has been used to identify the biologically important antigen in a complex mixture of proteins (9). The outer membrane-associated protein of C. burnetii is believed to be the antigenic target for the detection of antibody in clinical serum samples. This protein has been well characterized, with a molecular mass of approximately 27 kDa (18), and its usefulness as a immunodiagnostic reagent had also been evaluated (23).
In a preliminary study, we found some unsure results in the Q fever diagnosis by our IFA test and then compared them with ELISA results (data not published). Several samples reacted to phase II C. burnetii antigen in the IFA test but were negative by ELISA. On the other hand, a few of these sera were positive by ELISA but nonreactive in the IFA test. This phenomenon led us to reevaluate the criteria for the IFA test for the diagnosis of Q fever.
MATERIALS AND METHODS
Serum samples. A total of 346 human serum samples were included in this study. The specimens consisted of 16 serum samples from patients diagnosed with Q fever in past examinations by the IFA test, ELISA, and complement fixation test, which also served as a positive control, including three imported cases (13) and others kindly provided by Werner Slenczka (Institut fur Virologie, Marburg, Germany), Barrie P. Marmion (Institute of Medical and Veterinary Science, Adelaide, Australia), and E. Kovacova (Institute of Virology, Slovak Academy of Sciences, Slovak Republic), 150 paired and 4 single sera from Japanese pneumonia patients in Okayama Prefecture with unknown fever and no clinical information regarding the microbes in those pneumonia cases, and 26 negative sera from healthy donors, including 1 serum sample from Slovakia and 25 sera from Japan.
C. burnetii antigen. A total of 107 C. burnetii phase II strain Nine Mile (ATCC VR615) bacteria with a high passage number were purified by differential centrifugation and formalin treatment as described previously (20), with minor modifications. Briefly, after 3 to 5 days of inoculation of C. burnetii to Vero cell lines, the medium was discarded and the cells were collected with a cell scraper. The cells were then resuspended by Dounce homogenization in 0.02% formalin-phosphate-buffered saline (PBS) (pH 7.2) 20 times. The cell solution was centrifuged at 1,300 x g for 5 min, and the supernatant was collected. Subsequently, the collected supernatant was filtered through a 5.0-μm-pore-size filter (Millipore Corp., Bedford, MA) to remove soluble cell culture debris and centrifuged at 13,000 x g at 4°C for 15 min. The supernatant was then discarded, and the pellets were washed with PBS twice. The concentration of C. burnetii antigen was measured with a Bio-Rad protein assay kit (Bio-Rad Lab., Hercules, CA). After microscopic examination of impression smears checked by the IFA test, the partially purified C. burnetii was pooled and stored at –80°C until use.
IFA test. The IFA test was performed by using prepared C.burnetii antigen with twofold dilutions of serum from 1:16 to 1:2,048 in PBS. In brief, 4 μl of antigen was dotted in triplicate onto a clean 15-well multitest slide (ICN Biomedicals, Inc., Aurora, Ohio) and allowed to air dry. Once dry, the slides were fixed in acetone for 15 min at room temperature. The diluted serum samples were then overlaid onto the antigen spots and incubated at 37°C in a humidified chamber for 1 h. After one wash with distilled water, two washes with 0.05% PBS-Tween, and one more wash with distilled water, 8 μl of fluorescein isothiocyanate-labeled anti-human IgM or IgG (Biosource, Camarillo, CA) diluted 1:200 in 0.001% Evans blue solution was added to each antigen spot and the slides were incubated for 1 h at 37°C. Finally, they were washed as before, dried in air, and mounted in 50% glycerol-PBS (pH 8.6). The slides were examined with a 40x objective (x400 magnification) using a fluorescence microscope (Axioskop 2 plus; Zeiss) equipped for visualizing brilliant green staining of C. burnetii microorganisms.
ELISA. A commercial ELISA kit (PanBio Co Ltd., Windsor, Queensland, Australia) was used as recommended by the manufacturer. The bound conjugates were detected by using tetramethyl benzidine as a substrate, and the color change was assessed in a microplate reader at a test wavelength of 450 nm. The ELISA index can be obtained by calculating the ratio of the cutoff absorbance to the sample absorbance and multiplying by 10. The serum samples were considered positive if they had an index of more than 11, equivocal if the index was between 9 and 11, and negative if the index was less than 9.
WBA. The outer membrane complex of C. burnetii was extracted by trichloroacetic acid and the temperature treatment method as described elsewhere (22). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was carried out with a 12% polyacrylamide gel as a separating gel (12), and a Western blotting assay (WBA) was performed using prepared outer membrane complex with 1:400 dilutions of the sera. The assay result was considered positive if the sera recognized the approximately 27-kDa protein of C. burnetii.
RESULTS
Serological analysis. Three hundred forty-six human sera, including 16 sera from confirmed Q fever cases, 304 sera from pneumonia patients, and 26 negative sera, were tested by the IFA test (Fig. 1), ELISA, and WBA (Table 1). Thirteen sera with an IgM titer of 1:128 and/or an IgG titer of 1:256 by the IFA test were positive by both ELISA and WBA, whereas 302 sera with an IgM titer of 1:16 and an IgG titer of 1:32 by the IFA test were negative by both ELISA and WBA. In the case of an IgM titer of 1:32 and 1:64 and/or an IgG titer of 1:64 and 1:128 by the IFA test, 4 sera were IgM and/or IgG positive by both ELISA and WBA, whereas 2 sera were negative by ELISA but positive by WBA. In this study, we labeled this area the "equivocal area" and divided the results of the IFA test into 3 groups, as shown in Table 1. The detailed results of the 3 groups are discussed in the next section.
Groups. Group I included 13 sera from confirmed Q fever cases. Four of these sera were IgG positive by both ELISA and WBA, indicating convalescent or chronic cases, whereas 9 of these sera were IgM and IgG positive by both ELISA and WBA, suggesting acute or subacute cases (Table 2). Group II included 298 sera from pneumonia patients and 26 confirmed negative sera that were negative by ELISA and WBA. Group III, which was proposed as the "equivocal area" in this study, included 3 sera from confirmed Q fever cases and 6 sera from pneumonia patients. The 3 confirmed Q fever sera and one serum from a pneumonia patient were IgM and/or IgG positive by both ELISA and WBA (Table 3). However, 2 of the sera from pneumonia patients were negative by ELISA but positive by WBA, and the other 3 sera were negative by both ELISA and WBA.
DISCUSSION
We evaluated new criteria for the IFA test in the diagnosis of Q fever in Japan. We used 346 human sera, including 16 from confirmed Q fever cases, 304 from Japanese pneumonia patients, and 26 that were negative. Thirteen sera from the confirmed Q fever cases were positive by the IFA test, ELISA, and WBA, whereas 298 sera from pneumonia patients and 26 negative sera were negative by the IFA test, ELISA, and WBA. In the proposed "equivocal area," we found 9 sera with various results by the IFA test, ELISA, and WBA for each serum sample.
The high cutoff value should emphasize the predictive value of a positive result with a high probability. In our temporary criteria, the combination of a phase II IgM titer of 1:128 and/or a phase II IgG titer of 1:256 gave a positive result for 13 sera from confirmed Q fever patients. Under these conditions, the diagnosis can be made even with only a single serum sample. Additionally, the serum samples recognized an approximately 27-kDa protein of C. burnetii by WBA, which other workers suggested as an immunodominant component in certain acute cases of the disease (18, 19), and were positive by ELISA. These criteria can be considered more reliable than those of a recent study that defined the high sensitivity of the IFA test at a cutoff titer of 1:400 (16).
The low cutoff value for either IgM or IgG should give a high predictive value of a negative result; thus, diagnosis cannot be made below this titer with a high probability. We may consider phase II IgM and phase II IgG titers of 1:16 and 1:32, respectively, as the low cutoff values (Fig. 1). Most samples under the low cutoff values, including 149 paired and 26 negative-control sera, were negative and also did not recognize the specific protein by WBA, although a few of them had a significant index by ELISA (Table 4). This is in accordance with a previous study showing that ELISA is suitable for use as a screening assay for Q fever diagnosis, with the IFA test used to confirm negative results (3, 4). We may explain this difference by the fact that a nonspecific reaction by ELISA may still occur due to cellular debris in the antigen preparation from culture. Although a recent study showed that LightCycler nested PCR can also be applied as a secondary tool in the diagnostic strategy for the early diagnosis of acute Q fever (5), the result of this study showed a good correlation between IFA titers and ELISA index values. A higher IFA titer correlated with a higher index in the ELISA result. Based on this clarification, confirmation with another serological test might not be required for samples categorized as negative by the IFA test.
We proposed the titer equivocal area and found 9 sera, including 3 from confirmed Q fever cases and 6 from Japanese pneumonia patients in the area, by the IFA test. Three sera from the confirmed Q fever cases and one serum from a pneumonia patient were positive by ELISA and WBA. However, 2 sera from pneumonia patients were negative by ELISA but positive by WBA, which left 3 sera negative by both ELISA and WBA. These results suggest that a single cutoff value for the IFA test may cause false-positive and false-negative results. In addition, only the IgM titer was positive, and the titer was very low in the positive case of pneumonia. This result suggests that serological assay with paired sera should be done for confirmation.
The results presented here illustrate the new criteria for the IFA test for Q fever. We recommended that an "equivocal area" should be used for the IFA test, rather than a single cutoff value and that sera in the equivocal area should be tested by additional serological assays to eliminate false-positive and false-negative results.
ACKNOWLEDGMENTS
We thank all the physicians who kindly provided the clinical samples for this study and Werner Slenczka, Barrie P. Marmion, and Elena Kovacova for kindly providing the control human sera.
This work was financially supported by The Japan Human Sciences Foundation.
REFERENCES
Baca, O. G., and D. Paretsky. 1983. Q fever and Coxiella burnetii: a model for host-parasite interactions. Microbiol. Rev. 47:127-149.
Brouqui, P., H. Tissot Dupont, M. Drancourt, Y. Berlan, J. Etienne, C. Leport, F. Goldstein, P. Massip, M. Micoud, A. Bertrand, and D. Raoult. 1993. Chronic Q fever: 92 cases from France including 27 cases without endocarditis. Arch. Intern. Med. 153:642-648.
Field, P. R., J. L. Mitchell, A. Santiago, D. J. Dickeson, S.-W. Chan, D. W. T. Ho, A. M. Murphy, A. J. Cuzzubbo, and P. L. Devine. 2000. Comparison of a commercial enzyme-linked immunosorbent assay with immunofluorescence and complement fixation tests for detection of Coxiella burnetii (Q fever) immunoglobulin M. J. Clin. Microbiol. 38:1645-1647.
Field, P. R., A. Santiago, S.-W. Chan, D. B. Patel, D. Dickeson, J. L. Mitchell, P. L. Devine, and A. M. Murphy. 2002. Evaluation of a novel commercial enzyme-linked immunosorbent assay detecting Coxiella burnetii-specific immunoglobulin G for Q fever prevaccination screening and diagnosis. J. Clin. Microbiol. 40:3526-3529.
Fournier, P. E., and D. Raoult. 2003. Comparison of PCR and serology assays for early diagnosis of acute Q fever. J. Clin. Microbiol. 41:5094-5098.
Haldane, E. V., T. J. Marrie, R. S. Faulkner, S. H. Lee, J. H. Cooper, D. D. MacPherson, and T. J. Montague. 1983. Endocarditis due to Q fever in Nova Scotia: experience with five patients in 1981-1982. J. Infect. Dis. 148:978-985.
Hendrix, L. R., L. P. Mallavia, and J. E. Samuel. 1993. Cloning and sequencing of Coxiella burnetii outer membrane protein gene com1. Infect. Immun. 61:470-477.
Hunt, J. G., P. R. Field, and A. M. Murphy. 1983. Immunoglobulin responses to Coxiella burnetii (Q fever): single-serum diagnosis of acute infection, using an immunofluorescence technique. Infect. Immun. 39:977-981.
Kenneth, O., and K. M. Yamada. 1977. Direct detection of antigens in sodium dodecyl sulfate-polyacrylamide gels. Anal. Biochem. 78:483-490.
Kovacova, E., J. Gallo, S. Schramek, J. Kazar, and R. Brezina. 1987. Coxiella burnetii antigens for detection of Q fever antibodies by ELISA in human sera. Acta Virol. 31:254-259.
Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685.
Murphy, A. M., and L. Magro. 1980. IgM globulin response in Q fever (Coxiella burnetii) infections. Pathology 12:391-396.
Ogawa, M., T. Kawamoto, A. Kawamoto, T. Yamashita, Y. Uchida, K. Kato, A. Setiyono, S. Shiga, and T. Kishimoto. 2003. Time course of the levels of antibodies to Coxiella burnetii and detection of C. burnetii-DNA in three imported cases of acute Q fever. Kansenshogaku Zasshi 77:127-132. (In Japanese.)
Peacock, M. G., R. N. Philip, J. C. Williams, and R. S. Faulkner. 1983. Serological evaluation of Q fever in humans: enhanced phase I titers of immunoglobulins G and A are diagnostic for Q fever endocarditis. Infect. Immun. 41:1089-1098.
Peter, O., G. Dupuis, W. Burgdorfer, and M. Peacock. 1985. Evaluation of complement fixation and indirect immunofluorescence tests in the early diagnosis of primary Q fever. Eur. J. Clin. Microbiol. 4:394-396.
Rollain, J. M., C. Lecam, and D. Raoult. 2003. Simplified serological diagnosis of endocarditis due to Coxiella burnetii and Bartonella. Clin. Diagn. Lab. Immunol. 10:1147-1148.
Stocker, M. G. P., and P. Fiset. 1956. Phase variation of the Nine Mile and other strains of Rickettsia burnetii. Can. J. Microbiol. 2:310-321.
To, H., A. Hotta, G. Q. Zhang, S. V. Nguyen, M. Ogawa, T. Yamaguchi, H. Fukushi, K. Amano, and K. Hirai. 1997. Antigenic characteristics of polypeptides of Coxiella burnetii isolates. Microbiol. Immunol. 42:81-85.
To, H., A. Hotta, T. Yamaguchi, H. Fukushi, and K. Hirai. 1997. Antigenic characteristics of the lipopolysaccharides of Coxiella burnetii isolates. J. Vet. Med. Sci. 60:267-270.
To, H., K. K. Htwe, N. Yamasaki, G. Q. Zhang, M. Ogawa, T. Yamaguchi, H. Fukushi, and K. Hirai. 1995. Isolation of Coxiella burnetii from dairy cattle and ticks, and some characteristics of the isolates in Japan. Microbiol. Immunol. 39:663-671.
Turk, W. P., G. Howitt, L. A. Turnberg, H. Fox, M. Longson, M. B. Matthews, and R. DasGupta. 1976. Chronic Q fever. Q. J. Med. 45:193-217.
Vavrekova, M., F. Ciampor, and M. Lukacova. 1992. Electron-microscopic study of the effect of various extractants on the morphology of Coxiella burnetii. Folia Microbiol. 37:87-92.
Zhang, G. Q., A. Hotta, T. Ho, T. Yamaguchi, H. Fukushi, and K. Hirai. 1998. Evaluation of a recombinant 27-kDa outer membrane protein of Coxiella burnetii as an immunodiagnostic reagent. Microbiol. Immunol. 42:423-428.(A. Setiyono, M. Ogawa, Y.)
Department of Virology I, National Institute of Infectious Diseases, 1-23-1 Toyama, Shinjuku-Ku, Tokyo 162-8640, Japan
Laboratory of Pathology, Department of Parasitology and Pathology, Faculty of Veterinary Medicine, Bogor Agricultural University, Jl. Agatis, Kampus IPB Darmaga Bogor 16680, Indonesia
ABSTRACT
A study was made to evaluate the cutoff value of indirect immunofluorescent-antibody (IFA) test for Q fever diagnosis in Japan. We used 346 sera, including 16 from confirmed Q fever cases, 304 from Japanese pneumonia patients, and 26 from negative cases. Thirteen sera from the confirmed Q fever cases with an immunoglobulin M (IgM) titer of 1:128 and/or IgG titer of 1:256 by the IFA test were positive by both enzyme-linked immunosorbent assay (ELISA) and Western blotting assay (WBA), whereas 298 sera from pneumonia patients and 26 negative sera with an IgM titer of 1:16 and an IgG titer of 1:32 by the IFA test were negative by both ELISA and WBA. In the proposed "equivocal area," with an IgM titer of 1:32 and 1:64 and/or an IgG titer of 1:64 and 1:128, we found 9 sera, 3 from confirmed Q fever cases and 6 from Japanese pneumonia patients, by the IFA test. Three sera from the confirmed Q fever cases and one of the sera from pneumonia patients were IgM and/or IgG positive by both ELISA and WBA. These results suggest that a single cutoff value for the IFA test may cause false-positive and false-negative results. In conclusion, this study showed that an "equivocal area" should be used for the IFA test rather than a single cutoff value and that sera in the equivocal area should be tested by additional serological assays for confirmation.
INTRODUCTION
Q fever is a zoonosis caused by Coxiella burnetii, an obligate intracellular rickettsial organism. The clinical manifestations of Q fever are readily divided into acute and chronic forms. The most common clinical presentation of acute Q fever in human is an influenza-like illness, often accompanied by pneumonia. The chronic Q fever form, particularly endocarditis, may appear several years after the primary episode (1, 2, 6, 21). Because the clinical presentation of the infection is not specific, serological confirmation is required for the diagnosis of Q fever. In Japan, the currently used serological method is the indirect immunofluorescent-antibody (IFA) test, and more recently, some attempts have been made to use other methods, such as the enzyme-linked immunosorbent assay (ELISA).
The unique characteristic of C. burnetii is its antigenic phase variation (17). Virulent phase I can be isolated from natural infection of humans or from laboratory infections of animals. Phase II develops during serial passage in an immunologically incompetent host, such as cell cultures or fertilized eggs (1). Serologically, anti-phase I antibodies are normally found at high levels only during the chronic form of the disease, whereas specific anti-phase II antibodies predominate primarily in acute Q fever (14).
The IFA test has previously been used to detect immunoglobulin M (IgM) antibodies in the sera of Q fever patients within the first 2 weeks of illness (8, 12). The estimation of anti-C. burnetii IgM antibody using the IFA test in a single serum sample has been proven useful in confirming acute infection in humans (8). The IFA method is more sensitive and specific than the complement fixation test; however, it is less sensitive than the ELISA (15). Therefore, the validation of the immunofluorescent threshold value for Q fever serology should be important for the establishment of the diagnosis.
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by immunoblotting assay, has been used to identify the biologically important antigen in a complex mixture of proteins (9). The outer membrane-associated protein of C. burnetii is believed to be the antigenic target for the detection of antibody in clinical serum samples. This protein has been well characterized, with a molecular mass of approximately 27 kDa (18), and its usefulness as a immunodiagnostic reagent had also been evaluated (23).
In a preliminary study, we found some unsure results in the Q fever diagnosis by our IFA test and then compared them with ELISA results (data not published). Several samples reacted to phase II C. burnetii antigen in the IFA test but were negative by ELISA. On the other hand, a few of these sera were positive by ELISA but nonreactive in the IFA test. This phenomenon led us to reevaluate the criteria for the IFA test for the diagnosis of Q fever.
MATERIALS AND METHODS
Serum samples. A total of 346 human serum samples were included in this study. The specimens consisted of 16 serum samples from patients diagnosed with Q fever in past examinations by the IFA test, ELISA, and complement fixation test, which also served as a positive control, including three imported cases (13) and others kindly provided by Werner Slenczka (Institut fur Virologie, Marburg, Germany), Barrie P. Marmion (Institute of Medical and Veterinary Science, Adelaide, Australia), and E. Kovacova (Institute of Virology, Slovak Academy of Sciences, Slovak Republic), 150 paired and 4 single sera from Japanese pneumonia patients in Okayama Prefecture with unknown fever and no clinical information regarding the microbes in those pneumonia cases, and 26 negative sera from healthy donors, including 1 serum sample from Slovakia and 25 sera from Japan.
C. burnetii antigen. A total of 107 C. burnetii phase II strain Nine Mile (ATCC VR615) bacteria with a high passage number were purified by differential centrifugation and formalin treatment as described previously (20), with minor modifications. Briefly, after 3 to 5 days of inoculation of C. burnetii to Vero cell lines, the medium was discarded and the cells were collected with a cell scraper. The cells were then resuspended by Dounce homogenization in 0.02% formalin-phosphate-buffered saline (PBS) (pH 7.2) 20 times. The cell solution was centrifuged at 1,300 x g for 5 min, and the supernatant was collected. Subsequently, the collected supernatant was filtered through a 5.0-μm-pore-size filter (Millipore Corp., Bedford, MA) to remove soluble cell culture debris and centrifuged at 13,000 x g at 4°C for 15 min. The supernatant was then discarded, and the pellets were washed with PBS twice. The concentration of C. burnetii antigen was measured with a Bio-Rad protein assay kit (Bio-Rad Lab., Hercules, CA). After microscopic examination of impression smears checked by the IFA test, the partially purified C. burnetii was pooled and stored at –80°C until use.
IFA test. The IFA test was performed by using prepared C.burnetii antigen with twofold dilutions of serum from 1:16 to 1:2,048 in PBS. In brief, 4 μl of antigen was dotted in triplicate onto a clean 15-well multitest slide (ICN Biomedicals, Inc., Aurora, Ohio) and allowed to air dry. Once dry, the slides were fixed in acetone for 15 min at room temperature. The diluted serum samples were then overlaid onto the antigen spots and incubated at 37°C in a humidified chamber for 1 h. After one wash with distilled water, two washes with 0.05% PBS-Tween, and one more wash with distilled water, 8 μl of fluorescein isothiocyanate-labeled anti-human IgM or IgG (Biosource, Camarillo, CA) diluted 1:200 in 0.001% Evans blue solution was added to each antigen spot and the slides were incubated for 1 h at 37°C. Finally, they were washed as before, dried in air, and mounted in 50% glycerol-PBS (pH 8.6). The slides were examined with a 40x objective (x400 magnification) using a fluorescence microscope (Axioskop 2 plus; Zeiss) equipped for visualizing brilliant green staining of C. burnetii microorganisms.
ELISA. A commercial ELISA kit (PanBio Co Ltd., Windsor, Queensland, Australia) was used as recommended by the manufacturer. The bound conjugates were detected by using tetramethyl benzidine as a substrate, and the color change was assessed in a microplate reader at a test wavelength of 450 nm. The ELISA index can be obtained by calculating the ratio of the cutoff absorbance to the sample absorbance and multiplying by 10. The serum samples were considered positive if they had an index of more than 11, equivocal if the index was between 9 and 11, and negative if the index was less than 9.
WBA. The outer membrane complex of C. burnetii was extracted by trichloroacetic acid and the temperature treatment method as described elsewhere (22). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was carried out with a 12% polyacrylamide gel as a separating gel (12), and a Western blotting assay (WBA) was performed using prepared outer membrane complex with 1:400 dilutions of the sera. The assay result was considered positive if the sera recognized the approximately 27-kDa protein of C. burnetii.
RESULTS
Serological analysis. Three hundred forty-six human sera, including 16 sera from confirmed Q fever cases, 304 sera from pneumonia patients, and 26 negative sera, were tested by the IFA test (Fig. 1), ELISA, and WBA (Table 1). Thirteen sera with an IgM titer of 1:128 and/or an IgG titer of 1:256 by the IFA test were positive by both ELISA and WBA, whereas 302 sera with an IgM titer of 1:16 and an IgG titer of 1:32 by the IFA test were negative by both ELISA and WBA. In the case of an IgM titer of 1:32 and 1:64 and/or an IgG titer of 1:64 and 1:128 by the IFA test, 4 sera were IgM and/or IgG positive by both ELISA and WBA, whereas 2 sera were negative by ELISA but positive by WBA. In this study, we labeled this area the "equivocal area" and divided the results of the IFA test into 3 groups, as shown in Table 1. The detailed results of the 3 groups are discussed in the next section.
Groups. Group I included 13 sera from confirmed Q fever cases. Four of these sera were IgG positive by both ELISA and WBA, indicating convalescent or chronic cases, whereas 9 of these sera were IgM and IgG positive by both ELISA and WBA, suggesting acute or subacute cases (Table 2). Group II included 298 sera from pneumonia patients and 26 confirmed negative sera that were negative by ELISA and WBA. Group III, which was proposed as the "equivocal area" in this study, included 3 sera from confirmed Q fever cases and 6 sera from pneumonia patients. The 3 confirmed Q fever sera and one serum from a pneumonia patient were IgM and/or IgG positive by both ELISA and WBA (Table 3). However, 2 of the sera from pneumonia patients were negative by ELISA but positive by WBA, and the other 3 sera were negative by both ELISA and WBA.
DISCUSSION
We evaluated new criteria for the IFA test in the diagnosis of Q fever in Japan. We used 346 human sera, including 16 from confirmed Q fever cases, 304 from Japanese pneumonia patients, and 26 that were negative. Thirteen sera from the confirmed Q fever cases were positive by the IFA test, ELISA, and WBA, whereas 298 sera from pneumonia patients and 26 negative sera were negative by the IFA test, ELISA, and WBA. In the proposed "equivocal area," we found 9 sera with various results by the IFA test, ELISA, and WBA for each serum sample.
The high cutoff value should emphasize the predictive value of a positive result with a high probability. In our temporary criteria, the combination of a phase II IgM titer of 1:128 and/or a phase II IgG titer of 1:256 gave a positive result for 13 sera from confirmed Q fever patients. Under these conditions, the diagnosis can be made even with only a single serum sample. Additionally, the serum samples recognized an approximately 27-kDa protein of C. burnetii by WBA, which other workers suggested as an immunodominant component in certain acute cases of the disease (18, 19), and were positive by ELISA. These criteria can be considered more reliable than those of a recent study that defined the high sensitivity of the IFA test at a cutoff titer of 1:400 (16).
The low cutoff value for either IgM or IgG should give a high predictive value of a negative result; thus, diagnosis cannot be made below this titer with a high probability. We may consider phase II IgM and phase II IgG titers of 1:16 and 1:32, respectively, as the low cutoff values (Fig. 1). Most samples under the low cutoff values, including 149 paired and 26 negative-control sera, were negative and also did not recognize the specific protein by WBA, although a few of them had a significant index by ELISA (Table 4). This is in accordance with a previous study showing that ELISA is suitable for use as a screening assay for Q fever diagnosis, with the IFA test used to confirm negative results (3, 4). We may explain this difference by the fact that a nonspecific reaction by ELISA may still occur due to cellular debris in the antigen preparation from culture. Although a recent study showed that LightCycler nested PCR can also be applied as a secondary tool in the diagnostic strategy for the early diagnosis of acute Q fever (5), the result of this study showed a good correlation between IFA titers and ELISA index values. A higher IFA titer correlated with a higher index in the ELISA result. Based on this clarification, confirmation with another serological test might not be required for samples categorized as negative by the IFA test.
We proposed the titer equivocal area and found 9 sera, including 3 from confirmed Q fever cases and 6 from Japanese pneumonia patients in the area, by the IFA test. Three sera from the confirmed Q fever cases and one serum from a pneumonia patient were positive by ELISA and WBA. However, 2 sera from pneumonia patients were negative by ELISA but positive by WBA, which left 3 sera negative by both ELISA and WBA. These results suggest that a single cutoff value for the IFA test may cause false-positive and false-negative results. In addition, only the IgM titer was positive, and the titer was very low in the positive case of pneumonia. This result suggests that serological assay with paired sera should be done for confirmation.
The results presented here illustrate the new criteria for the IFA test for Q fever. We recommended that an "equivocal area" should be used for the IFA test, rather than a single cutoff value and that sera in the equivocal area should be tested by additional serological assays to eliminate false-positive and false-negative results.
ACKNOWLEDGMENTS
We thank all the physicians who kindly provided the clinical samples for this study and Werner Slenczka, Barrie P. Marmion, and Elena Kovacova for kindly providing the control human sera.
This work was financially supported by The Japan Human Sciences Foundation.
REFERENCES
Baca, O. G., and D. Paretsky. 1983. Q fever and Coxiella burnetii: a model for host-parasite interactions. Microbiol. Rev. 47:127-149.
Brouqui, P., H. Tissot Dupont, M. Drancourt, Y. Berlan, J. Etienne, C. Leport, F. Goldstein, P. Massip, M. Micoud, A. Bertrand, and D. Raoult. 1993. Chronic Q fever: 92 cases from France including 27 cases without endocarditis. Arch. Intern. Med. 153:642-648.
Field, P. R., J. L. Mitchell, A. Santiago, D. J. Dickeson, S.-W. Chan, D. W. T. Ho, A. M. Murphy, A. J. Cuzzubbo, and P. L. Devine. 2000. Comparison of a commercial enzyme-linked immunosorbent assay with immunofluorescence and complement fixation tests for detection of Coxiella burnetii (Q fever) immunoglobulin M. J. Clin. Microbiol. 38:1645-1647.
Field, P. R., A. Santiago, S.-W. Chan, D. B. Patel, D. Dickeson, J. L. Mitchell, P. L. Devine, and A. M. Murphy. 2002. Evaluation of a novel commercial enzyme-linked immunosorbent assay detecting Coxiella burnetii-specific immunoglobulin G for Q fever prevaccination screening and diagnosis. J. Clin. Microbiol. 40:3526-3529.
Fournier, P. E., and D. Raoult. 2003. Comparison of PCR and serology assays for early diagnosis of acute Q fever. J. Clin. Microbiol. 41:5094-5098.
Haldane, E. V., T. J. Marrie, R. S. Faulkner, S. H. Lee, J. H. Cooper, D. D. MacPherson, and T. J. Montague. 1983. Endocarditis due to Q fever in Nova Scotia: experience with five patients in 1981-1982. J. Infect. Dis. 148:978-985.
Hendrix, L. R., L. P. Mallavia, and J. E. Samuel. 1993. Cloning and sequencing of Coxiella burnetii outer membrane protein gene com1. Infect. Immun. 61:470-477.
Hunt, J. G., P. R. Field, and A. M. Murphy. 1983. Immunoglobulin responses to Coxiella burnetii (Q fever): single-serum diagnosis of acute infection, using an immunofluorescence technique. Infect. Immun. 39:977-981.
Kenneth, O., and K. M. Yamada. 1977. Direct detection of antigens in sodium dodecyl sulfate-polyacrylamide gels. Anal. Biochem. 78:483-490.
Kovacova, E., J. Gallo, S. Schramek, J. Kazar, and R. Brezina. 1987. Coxiella burnetii antigens for detection of Q fever antibodies by ELISA in human sera. Acta Virol. 31:254-259.
Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680-685.
Murphy, A. M., and L. Magro. 1980. IgM globulin response in Q fever (Coxiella burnetii) infections. Pathology 12:391-396.
Ogawa, M., T. Kawamoto, A. Kawamoto, T. Yamashita, Y. Uchida, K. Kato, A. Setiyono, S. Shiga, and T. Kishimoto. 2003. Time course of the levels of antibodies to Coxiella burnetii and detection of C. burnetii-DNA in three imported cases of acute Q fever. Kansenshogaku Zasshi 77:127-132. (In Japanese.)
Peacock, M. G., R. N. Philip, J. C. Williams, and R. S. Faulkner. 1983. Serological evaluation of Q fever in humans: enhanced phase I titers of immunoglobulins G and A are diagnostic for Q fever endocarditis. Infect. Immun. 41:1089-1098.
Peter, O., G. Dupuis, W. Burgdorfer, and M. Peacock. 1985. Evaluation of complement fixation and indirect immunofluorescence tests in the early diagnosis of primary Q fever. Eur. J. Clin. Microbiol. 4:394-396.
Rollain, J. M., C. Lecam, and D. Raoult. 2003. Simplified serological diagnosis of endocarditis due to Coxiella burnetii and Bartonella. Clin. Diagn. Lab. Immunol. 10:1147-1148.
Stocker, M. G. P., and P. Fiset. 1956. Phase variation of the Nine Mile and other strains of Rickettsia burnetii. Can. J. Microbiol. 2:310-321.
To, H., A. Hotta, G. Q. Zhang, S. V. Nguyen, M. Ogawa, T. Yamaguchi, H. Fukushi, K. Amano, and K. Hirai. 1997. Antigenic characteristics of polypeptides of Coxiella burnetii isolates. Microbiol. Immunol. 42:81-85.
To, H., A. Hotta, T. Yamaguchi, H. Fukushi, and K. Hirai. 1997. Antigenic characteristics of the lipopolysaccharides of Coxiella burnetii isolates. J. Vet. Med. Sci. 60:267-270.
To, H., K. K. Htwe, N. Yamasaki, G. Q. Zhang, M. Ogawa, T. Yamaguchi, H. Fukushi, and K. Hirai. 1995. Isolation of Coxiella burnetii from dairy cattle and ticks, and some characteristics of the isolates in Japan. Microbiol. Immunol. 39:663-671.
Turk, W. P., G. Howitt, L. A. Turnberg, H. Fox, M. Longson, M. B. Matthews, and R. DasGupta. 1976. Chronic Q fever. Q. J. Med. 45:193-217.
Vavrekova, M., F. Ciampor, and M. Lukacova. 1992. Electron-microscopic study of the effect of various extractants on the morphology of Coxiella burnetii. Folia Microbiol. 37:87-92.
Zhang, G. Q., A. Hotta, T. Ho, T. Yamaguchi, H. Fukushi, and K. Hirai. 1998. Evaluation of a recombinant 27-kDa outer membrane protein of Coxiella burnetii as an immunodiagnostic reagent. Microbiol. Immunol. 42:423-428.(A. Setiyono, M. Ogawa, Y.)