MAJOR ARTICLE Diagnosis of Coccidioidomycosis with Use of the Coccidioides Antigen Enzyme Immunoassay Michelle Durkin, 1 Patricia Connolly, 1 Tim Kuberski, a Robert Myers, 2 Bernard M. Kubak, 3 David Bruckner, 4 David Pegues, 3 and L. Joseph Wheat 1 1 MiraVista Diagnostics, Indianapolis, Indiana; 2 Maricopa Medical Center, Phoenix, Arizona; and 3 Division of Infectious Diseases and 4 Department of Pathology & Laboratory Medicine, David Geffen School of Medicine at the University of California at Los Angeles, Los Angeles Background. We have previously shown antigenuria in patients with coccidioidomycosis through use of the Histoplasma antigen enzyme immunoassay (EIA), and now we have developed a specific Coccidioides antigen EIA. Methods. The Coccidioides EIA uses antibodies to Coccidioides galactomannan. The sensitivity of the Coccidioides and Histoplasma EIAs was evaluated in patients with more-severe coccidioidomycosis, and the specificity of these EIAs was evaluated in patients with nonfungal infections, in patients with other endemic mycoses, and in healthy individuals. Results. Among patients in the present study, antigenuria was detected in 70.8% of patients with coccidioidomycosis with use of the Coccidioides EIA and in 58.3% of patients with use of the Histoplasma EIA. Antigenuria was absent in 99.4% of healthy individuals, patients with nonfungal infections, and patients with noninfectious conditions. Cross-reactions with other endemic mycoses were observed in 10.7% of patients. Conclusions. The Coccidioides EIA has potential to be useful in the rapid diagnosis of more-severe forms of coccidioidomycosis. Coccidioidomycosis is a fungal infection caused by the dimorphic fungi Coccidioides immitis and Coccidioides posadasii, which are endemic to specific geographic locations in the Western hemisphere. The clinical manifestations of exposure to these fungi vary, but in symptomatic individuals, they are most commonly pulmonary in nature. Severe pneumonia or disseminated disease may occur and is more frequent in those who are immunosuppressed, including recent solid-organ transplant recipients [1 3], those with AIDS [4], patients with chronic inflammatory conditions who receive immunosuppressive therapy [5, 6], pregnant women [7], black Americans, Filipinos, and individuals with certain other conditions [8, 9]. The diagnosis of coccidioidomycosis may be difficult and often is delayed. For most patients with pneumonia, culture specimens obtained through broncho- Received 11 June 2008; accepted 15 July 2008; electronically published 9 September 2008. a Private practice, Phoenix, Arizona. Reprints or correspondence: Dr. L. Joseph Wheat, MiraVista Diagnostics, 4444 Decatur Blvd., Ste. 300, Indianapolis, IN 46241 (jwheat@miravistalabs.com). Clinical Infectious Diseases 2008; 47:e69 73 2008 by the Infectious Diseases Society of America. All rights reserved. 1058-4838/2008/4708-00E1$15.00 DOI: 10.1086/592073 alveolar lavage have positive results, and the mean time for isolation is nearly 1 week [10]. Rapid diagnosis by finding spherules in bronchoalveolar lavage on direct examination is possible in approximately two-thirds of cases. Others have reported demonstration of spherules in bronchoalveolar lavage in 30% 40% of patients with Coccidioides pneumonia [11]. Thus, although demonstration of spherules in respiratory secretions may provide a rapid diagnosis, at least one-third of cases may be missed by direct examination. Serological testing for Coccidioides can be useful but may have negative results early in the infection and in immunosuppressed patients. The immunodiffusion test had positive results in 53% of immunosuppressed patients, compared with 73% of nonimmunosuppressed patients, and the complement fixation test had positive results in 67% and 75%, respectively [12]. The sensitivity of the immunodiffusion test may be enhanced by concentration of the serum [13]. In nonimmunosuppressed patients with acute Coccidioides pneumonia, the immunodiffusion test for IgM antibodies was positive in 1 of 10 patients, but results were positive in 9 of 10 patients after serum concentration [14]. Others noted seropositivity in approximately one-half of patients with Coccidioides pneumonia [10, 11]. Coccidioides Antigen Immunoassay CID 2008:47 (15 October) e69
Figure 1. Titration of Histoplasma and Coccidioides galactomannan (GM) in the Histoplasma EIA (Hc EIA) and the Coccidioides EIA (Ci EIA). OD, optical density. Antigen detection could be useful for the early diagnosis of coccidioidomycosis. Antigenemia was noted in 56% 78% of patients with coccidioidomycosis in early studies [15 17] and in a more recent report [18]. We reported detection, using the Histoplasma antigen EIA, of a cross-reactive galactomannan antigen in the urine of 58% of patients with more-severe coccidioidomycosis [19]. We have endeavored to develop a specific Coccidioides antigen detection test and now evaluate its potential usefulness for the diagnosis of coccidioidomycosis. PATIENTS AND METHODS Vaccine preparation and immunization. Six patient isolates of C. immitis and C. posadasii were grown in brain-heart infusion broth at 37 C on a gyratory shaker (New Brunswick Scientific) for 72 h. Hyphae were removed by centrifugation, and the supernatant was decanted and combined with 3 volumes of absolute ethanol. Galactomannan was purified by immunoaffinity chromatography with use of concanavalin A, as described for preparation of Histoplasma galactomannan [20]. New Zealand white rabbits were immunized with galactomannan and formalin-killed mold, as described for production of antibodies to Histoplasma capsulatum [20]. Calibration standards. Calibration standards were prepared from urine containing high levels of Coccidioides antigen, as described elsewhere [21]. Multiple dilutions of the standard were prepared, and the antigen content of each was determined by comparison with known concentrations of the purified galactomannan. EIA. The MVista Coccidioides antigen EIA (MiraVista Diagnostics) was performed as reported elsewhere for the MVista Histoplasma antigen EIA (MiraVista Diagnostics) [20, 21], except that microplates coated with rabbit anti-coccidioides antibodies were used. Galactomannan attached to the capture antibody was detected with biotinylated rabbit anti-coccidioides detector antibody. Clinical specimens. Urine specimens were available from 18 patients described elsewhere [19] and 6 additional patients with culture-proven cases of coccidioidomycosis from the clinical practices of the investigators. Controls included samples from patients with histoplasmosis (10 patients) or blastomycosis (10), which were left over from clinical testing; patients with paracoccidioidomycosis (8), provided by A. Restrepo, Colombia; patients with nonfungal infectious or noninfectious conditions for which Histoplasma antigen testing was ordered (50); patients with legionnaires disease (10), obtained from the Centers for Disease Control and Prevention; patients with Mycoplasma pneumonia (25), provided by J. Summersgill, University of Louisville; and healthy volunteers (75), purchased from SeraCare. Statistical analysis. The cutoff for positivity was determined using receiver operating characteristic analysis. RESULTS Coccidioides and Histoplasma galactomannans were titrated using the Histoplasma and Coccidioides EIAs (figure 1). The lower limit of detection for Coccidioides galactomannan was 0.03 ng/ ml in the Coccidioides EIA, compared with 0.5 ng/ml in the Histoplasma antigen assay. The lower limit of detection for the Histoplasma galactomannan was 0.125 ng/ml in the Histoplasma EIA, compared with 8 ng/ml in the Coccidioides EIA. Thus, low concentrations of Coccidioides galactomannan (0.5 ng/ml) could be detected by the Histoplasma antigen EIA, which explains the high rate of cross-reactivity (58%) in patients with coccidioidomycosis [19]. Conversely, high concentrations of Histoplasma galactomannan (8 ng/ml) were required for cross-reactivity in the Coccidioides antigen EIA. e70 CID 2008:47 (15 October) Durkin et al.
Among 24 patients with coccidioidomycosis, underlying immunosuppressive conditions were present in 19 (79.2%), including AIDS in 15 (62.5%), organ transplantation in 2 (8.3%), and corticosteroid use in 2 (8.3%) (table 1). Infection was acquired in Arizona by 21 patients and in California by 3 patients. Clinical manifestations included pneumonia (13 patients [54.2%]), pneumonia with extrapulmonary dissemination (9 [37.5%]), and chronic disseminated disease alone (2 [8.3%]). Seven patients (29.2%) died of the infection. The diagnosis was based on positive culture results in 21 patients (87.5%), including 8 (33.3%) with positive results of culture samples obtained from extrapulmonary sites. Spherules were seen in respiratory specimens in 6 patients (25%). Serological test results were positive in 13 (54.2%) patients and were the basis for diagnosis in 3 (12.5%). Twenty-two urine specimens from 11 date were available for 7 patients, for a total of 46 specimens. The cutoff for positivity was determined by testing 46 specimens from patients with coccidioidomycosis and 160 specimens from control individuals without systemic fungal infection. The area under the receiver operating characteristic curve was 0.940 (SE, 0.052; 95% CI, 0.90 0.97; P!.001). With use Table 1. Clinical and laboratory findings in coccidioidomycosis cases. Patient Underlying condition Presentation Positive culture Serological test 1 Cirrhosis, treated with corticosteroids Antigen concentration, ng/ml Coccidioides Histoplasma Chronic pneumonia, fatal BAL IgM +, IgG Negative Negative 2 HIV infection Acute pneumonia a Blood, BAL, urine IgM, IgG 3.61 6.52 3 HIV infection Acute pneumonia, fatal BAL IgM, IgG + 0.38!0.6 4 Diabetes Chronic pneumonia a None IgM +, IgG + Negative!0.6 5 Diabetes, treated with corticosteroids Acute pneumonia Spt IgM +, IgG + 0.73 Negative 6 Pregnancy Chronic pneumonia None CF 1:2 Negative Negative 7 HIV infection Acute pneumonia Spt IgM, IgG 2.78 4.69 8 None Chronic pneumonia None IgM +, IgG Negative Negative 9 HIV infection Acute pneumonia, a fatal BAL, Spt, Blood IgM, IgG, CF 3.09 2.31 10 Cirrhosis Acute pneumonia, fatal Spt IgM +, IgG +, CF 1:32 0.65 3.18 11 HIV infection Acute pneumonia a BAL, b blood IgM, IgG, CF 2.78 3.67 12 HIV infection Acute pneumonia a BAL b IgM +, IgG + 1.51 Negative 13 HIV infection Acute pneumonia None IgM +, IgG +, CF 1:16 Negative Negative 14 HIV infection Acute pneumonia a BAL b IgM, IgG, CF 2.70 Negative 15 HIV infection Acute pneumonia BAL IgM +, IgG 0.09 Negative 16 HIV infection Acute pneumonia BAL, b pleura CF 1:64 2.03!0.6 17 Heart transplantation Acute pneumonia, a fatal Blood IgM, IgG 3.42 4.16 18 Kidney transplantation Acute pneumonia, a fatal Blood, BM, BAL IgM, IgG, CF!1:2 3.88 1.23 19 None Chronic disseminated coccidioidomycosis Soft tissue, bone Not done 0.55 Negative 20 HIV infection Acute pneumonia, a fatal Blood CF 1:32 Negative 1.23 21 HIV infection Acute pneumonia BAL IgM +, IgG +, CF 1:16 0.31 2.66 22 HIV infection Chronic pneumonia BAL 2 Not done 1.67!0.6 23 None Chronic disseminated coccidioidomycosis Soft tissue, bone Not done 2.07 Negative 24 HIV infection Chronic pneumonia BAL 2 Not done Negative Negative of that cutoff, antigenuria was detected in 17 (70.8%; 95% CI, 50.8% 85.1%) of 24 patients with coccidioidomycosis and 1 (0.6%; 95% CI, 0.1% 3.4%) of 160 control individuals without fungal infections (figure 2). Under the assumption of a 5% prevalence of coccidioidomycosis among patients tested for Coccidioides antigen, assumed on the basis of histoplasmosis test results obtained at MiraVista Diagnostics, the positive predictive value is 85.7% and the negative predictive value is 98.5% among patients with more-severe forms of coccidioidomycosis. Antigenuria was detected by the Histoplasma EIA in 14 (58.3%) of the patients with coccidioidomycosis (95% CI, 38.3% 75.5%), 2 of whom had negative results of the Coccidioides EIA. The first patient with a false-negative Coccidioides result was not immunocompromised and presented with pulmonary coccidioidomycosis, as determined by a positive serological test result; the Histoplasma antigen was weakly positive, at!0.6 ng/ml. The second patient was immunocompromised and had Coccidioides fungemia; the Histoplasma antigen concentration was 1.23 ng/ml. Neither patient had culture results positive for H. capsulatum or clinical findings suggestive of histoplasmosis. Overall, antigenuria was detected in 19 pa- NOTE. BAL, bronchoalveolar lavage; BM, bone marrow; CF, complement fixation; Spt, sputum; +, positive;, negative. a Concurrent extrapulmonary dissemination. b Spherules seen by direct examination. Coccidioides Antigen Immunoassay CID 2008:47 (15 October) e71
Figure 2. Coccidioides antigenuria in patients with coccidioidomycosis and control individuals, as determined by the Coccidioides EIA. The broken horizontal line represents the cutoff for positivity. The numbers in the boxes represent the number of patients with negative results. tients (79.2%; 95% CI, 59.6% 90.8%) with use of either the Coccidioides or Histoplasma EIA. Antigenuria was detected in 3 (10.7%) of 28 control individuals with other endemic mycoses (95% CI, 3.7% 27.2%), 2 of whom had histoplasmosis and 1 of whom had paracoccidioidomycosis (detected at 0.12 and 0.14 ng/ml and at 0.15 ng/ml, respectively). The Histoplasma antigen concentration was 139 ng/ml in 2 individuals and was 28.82 ng/ml in the third control individual, who had a positive Coccidioides EIA result. Interassay reproducibility was determined by testing specimens from 5 patients 5 times. The results, expressed as mean SD (coefficient of variation), were as follows: 7.97 1.02 ng/ml (12.7%), 2.62 0.19 ng/ml (7.3%), 0.67 0.04 ng/ml (6.4%), 0.48 0.06 ng/ml (12.3%), and 0.13 0.01 ng/ml (9.5%). Twenty positive specimens and 30 negative specimens were tested with 2 separate assays and were reproducible in all 50 tests. DISCUSSION The lower limit of detection for Coccidioides galactomannan was 0.03 ng/ml in the Coccidioides EIA, compared with 0.5 ng/ ml in the Histoplasma EIA, showing that the Coccidioides EIA was more sensitive for detection of antigenuria in patients with coccidioidomycosis than was the Histoplasma EIA. The sensitivity in patients with more-severe forms of coccidioidomycosis was 71% for the Coccidioides EIA, compared with 58% for the Histoplasma EIA. An important caveat is that these findings should be applied only to patients with more-severe forms of coccidioidomycosis. More than three-quarters (79%) of the study patients were immunocompromised, nearly one-half (46%) had evidence of disseminated disease, almost one-third (30%) had fatal cases, and 87% had culture-positive specimens. Two patients with Coccidioides infection had positive results with the Histoplasma EIA but negative results with the Coccidioides EIA. This suggests that the antigenic makeup of individual Coccidioides strains may vary and that testing with both assays may be required to achieve the highest sensitivity. Coccidioides species related specificity is unlikely, however, because the vaccine contained antigens from both C. immitis and C. posadasii, and antigenuria was demonstrated in patients from both Arizona and California. The Coccidioides EIA is more specific for coccidioidomycosis than is the Histoplasma EIA. For example, Histoplasma galactomannan was detected only at concentrations of 8 ng/ml with use of the Coccidioides antigen EIA. Also, cross-reactions were noted in only 11% of specimens from patients with other endemic mycoses, which occurred in specimens containing very high Histoplasma antigen concentrations. Thus, cross-reactivity could be determined by testing the specimens with both EIAs. Because of the favorable performance characteristics of the Coccidioides EIA described in this study, this test may be useful for rapid diagnosis of more-severe forms of coccidioidomycosis. These findings should not be applied to nonimmunocompromised patients with primary coccidioidomycosis or other forms of coccidioidomycosis that have not been evaluated. The turnaround time for the antigen test is 1 day in most cases. Although the EIA is highly specific, cross-reactions are possible in patients with other endemic mycoses. Concurrent testing with the Histoplasma EIA may detect some false-negative cases and identify cross-reactions caused by other endemic mycoses. Additional studies are needed to define the usefulness of the test in the various clinical forms of coccidioidomycosis, in patients with a variety of underlying diseases or no underlying disease, and in specimen types other than urine. Acknowledgments Financial support. The cost of assay development and testing was a research expenditure of MiraBella Technologies. Potential conflicts of interest. M.D., P.C., and L.J.W. are employees of MiraVista Diagnostics. All other authors: no conflicts. References 1. Braddy CM, Heilman RL, Blair JE. Coccidioidomycosis after renal transplantation in an endemic area. Am J Transplant 2006; 6:340 5. 2. Blair JE, Logan JL. Coccidioidomycosis in solid organ transplantation. Clin Infect Dis 2001; 33:1536 44. 3. Kubak B. Coccidioidomycosis in heart and lung transplant recipients. In: Lynch J, ed. New York: Taylor and Francis, 2006:527 47. 4. Ampel NM. Coccidioidomycosis in persons infected with HIV-1. Ann N Y Acad Sci 2007; 1111:336 42. e72 CID 2008:47 (15 October) Durkin et al.
5. Mertz LE, Blair JE. Coccidioidomycosis in rheumatology patients: incidence and potential risk factors. Ann N Y Acad Sci 2007; 1111:343 57. 6. Bergstrom L, Yocum DE, Ampel NM, et al. Increased risk of coccidioidomycosis in patients treated with tumor necrosis factor a antagonists. Arthritis Rheum 2004; 50:1959 66. 7. Arnold CA, Rakheja D, Arnold MA, et al. Unsuspected, disseminated coccidioidomycosis without maternofetal morbidity diagnosed by placental examination: case report and review of the literature. Clin Infect Dis 2008; 46:e119 23. 8. Shubitz LF, Galgiani JN, Tian ZQ, Zhong Z, Timmermans P, Katz L. Efficacy of ambruticin analogs in a murine model of coccidioidomycosis. Antimicrob Agents Chemother 2006; 50:3467 9. 9. Rosenstein NE, Emery KW, Werner SB, et al. Risk factors for severe pulmonary and disseminated coccidioidomycosis: Kern County, California, 1995 1996. Clin Infect Dis 2001; 32:708 15. 10. Sarosi GA, Lawrence JP, Smith DK, Thomas A, Hobohm DW, Kelley PC. Rapid diagnostic evaluation of bronchial washings in patients with suspected coccidioidomycosis. Semin Respir Infect 2001; 16:238 41. 11. DiTomasso JP, Ampel NM, Sobonya RE, Bloom JW. Bronchoscopic diagnosis of pulmonary coccidioidomycosis: comparison of cytology, culture, and transbronchial biopsy. Diagn Microbiol Infect Dis 1994; 18:83 7. 12. Blair JE, Coakley B, Santelli AC, Hentz JG, Wengenack NL. Serologic testing for symptomatic coccidioidomycosis in immunocompetent and immunosuppressed hosts. Mycopathologia 2006; 162:317 24. 13. Pappagianis D. Serologic studies in coccidioidomycosis. Semin Respir Infect 2001; 16:242 50. 14. Petersen LR, Marshall SL, Barton-Dickson C, et al. Coccidioidomycosis among workers at an archeological site, northeastern Utah. Emerg Infect Dis 2004; 10:637 42. 15. Weiner MH. Antigenemia detected in human coccidioidomycosis. J Clin Microbiol 1983; 18:136 42. 16. Galgiani JN, Grace GM, Lundergan LL. New serologic tests for early detection of coccidioidomycosis. J Infect Dis 1991; 163:671 4. 17. Galgiani JN, Dugger KO, Ito JI, Wieden MA. Antigenemia in primary coccidioidomycosis. Am J Trop Med Hyg 1984; 33:645 9. 18. Miller CN, Johnson SM, Pappagianis D. Detection of circulating coccidioidal antigen using an inhibition ELISA [abstract F-038]. In: Proceedings of the 51st Annual Meeting of the Coccidioidomycosis Study Group (Tempe, AZ). Washington, DC: American Society for Microbiology, 2007. 19. Kuberski T, Myers R, Wheat LJ, Kubak BM, Bruckner D, Peuges D. Diagnosis of coccidioidomycosis by antigen detection using cross-reaction with a Histoplasma antigen. Clin Infect Dis 2007; 44:e50 4. 20. Wheat LJ, inventor. MiraVista Diagnostics. Wheat LJ, assignee. Fungal antigen immunoassay. US patent 20070020711. 21. Connolly PA, Durkin MM, LeMonte AM, Hackett EJ, Wheat LJ. Detection of histoplasma antigen by a quantitative enzyme immunoassay. Clin Vaccine Immunol 2007; 14:1587 91. Coccidioides Antigen Immunoassay CID 2008:47 (15 October) e73