Detection of Candida antigen and antibody in serum from patients with invasive candidiasis

Original

Report

Detection of Can&da Antigen and Antibody Patients with Invasive Candidiasis

in Serum from

Koji Takaki, MD;* Nobuyuki Shimono, MD; Toshiyuki Ishimaru, MD; Kaoru Okada, MD; Yoshiro Sawae, MD; and Yoshiyuki Niho, MD

ABSTRACT Objective: To compare the sensitivity and specificity different assays for diagnosis of invasive candidiasis.

of three

Methods: A passive hemagglutination assay (PHA), counterimmunoelectrophoresis assay (CIE), and Cand-tee were used to test sera from 125 patients with hematologic malignancies and 65 other hospitalized patients. The former group included 15 patients with invasive candidiasis, 38 patients with Candida colonization, and 72 patients without candidiasis. Results: Sensitivity/specificity of PHA, CIE, and Cand-tee were 87%/85%, 67%/98%, and 33%/97%, respectively. The measurement of antibody in paired sera, by PHA, was sensitive and specific; however, increased antibody titers usually occurred late in the disease. Conclusions: The combination of PHA and CIE, with a sensitivity of 67% and specificity of 98%, appeared to be the best assays for detection of invasive candidiasis in this cohort. The Cand-tee assay for Candida antigen had poor sensitivity for diagnosis of infection.

The serologic diagnosis of candidiasis has been explored by many investigators, with conflicting results. l-6 Detection of antibody has generally been unsuccessful, because these immunocompromised patients are incapable of mounting a brisk humoral response to infection. Detection of antigen may also be difficult since Cundida appears to disseminate only scant quantities of antigen into body fluids and many of these antigens, because they form complexes with antibodies or other serum proteins, do not reach detectable concentrations in serum. The sensitivity and specificity of three different serologic assays were compared for diagnosis of invasive candidiasis: a passive hemagglutination assay (PHA), counterimmunoelectrophoresis assay (CIIE), and Cand-tee (Ramco Laboratories, Houston, TX, USA). These assays were used to test serum from patients with and without candidiasis. MATERIAL Patients

Key Words:

Int J Infect

antibodies, antigens, Candida, counterimmunoelectrophoresis, leukemia, passive hemagglutination assay sen.sitivitL: specificity

Dis 1996;

1:78-82

Deep-seated mycosis remains a major infectious complication among patients with leukemia, malignant lymphoma, and other neoplasms, and among transplant recipients. Invasive candidiasis is the most common serious opportunistic fungal infection among these patients. Although there have been improvements in antifungal therapy, there is a need for more sensitive and reliable methods to diagnose invasive candidiasis.

*First Department of Internal Medicine, Faculty University, 3-1-1, Maidashi, Higashi-ku, Fukuoka, Received:

April

19, 1996;

Address correspondence nal Medicine, Faculty Higashi-ku, Fukuoka,

78

Accepted:

July

to Dr. Koji Takaki, of Medicine, Kyushu 812, Japan.

of Medicine, 812, Japan.

Kyushu

12, 1996. First Department University, 3-l-1,

of InterMaidashi,

AND

METHODS

and Controls

The cohort included 15 patients with hematologic disorders who had invasive candidiasis and 110 patients who did not. All had been admitted to the First Department of Internal Medicine, Kyushu University The diagnosis of invasive candidiasis was based on the following criteria: histopathologic proof or isolation of Cundida spp from autopsy specimens (cases 7, 12, and 13) or from sterile site biopsy (cases 8-l 1 and 14) and isolation of more than two positive blood cultures of Candida spp from specimens taken with at least a 24-hour interval from different sites (cases 1-6 and 15) (Table 1). Underlying diseases included three cases of myelodysplastic syndrome (MDS), three cases of malignant lymphoma (ML), two cases of acute myelocytic leukemia &ML), two cases of adult T-cell leukemia (ATL), and one case each of hairy cell leukemia, acute promyelocytic leukemia (APL), acute lymphocytic leukemia (ALL), acute monocytic leukemia (AMoL), and acute myelomonocytic leukemia (AMMoL). Mean age of the patients was 38 years

Detection of Candida

Table No.

Sex

1

M M M M M M M M M F M F F F F

2 3 4 5 6 7 a 9 10 11 12 13 14 15 MDS = myelodysplastic ALL = acute lymphocytic

1.

Invasive

Candidiasis Underlying

Age W 19 71 25 32 17 16 66 25 31 22 45 54 41 25 78

with

Hematologic

Antigen and Antibody / Takaki et al

Disorders

Disease

MDS Hairy cell leukemia MDS APL ALL ML ATL ML AML ML ATL AMoL AML AMMoL MDS

Candidiasis

Organism

Candidemia Candidemia Candidemia Candidemia Pulmonary Pulmonary Pulmonary GI-tract GI-tract GI-tract GI-tract Liver abscess Renal abscess Arthritis Urinary

C. parapsilosis C. tropicalis C. glabrata C. a/b/cans C. albicans C. albicans Candida spp C. albicans C. albicans C. albicans C. albicans Candida spp Candida spp C. albicans C. albicans

syndrome; ML = malignant lymphoma; AML = acute myelocytic leukemia; ATL = adult T-cell leukemia; APL = acute promyelocytic leukemia; AMoL = acute monocgic leukemia: AMMoL = acute myelomonocytic leukemia; GI-tract = gastroinWhx4 tract.

(range, 16-78 yr). There were ten male and five female patients (see Table 1). Sixty serum samples were collected from the patients with invasive candidiasis. Controls consisted of four groups: group 1, 80 serum samples obtained from 38 afebrile patients with hematologic disorders, who had Candida isolated from sputum, urine, or stool; group 2, 170 serum samples collected from 72 patients with hematologic disorders who had negative surveillance cultures for Candida; group 3, 42 serum samples from 42 healthy hospital staff members; and group 4, 23 serum samples from 23 febrile patients with shigellosis. The latter group was included to explore the effects of intestinal infection and mucosal injury on the interaction between Candida and the host. Antigen Detection Cand-tee was used to detect the Candida antigen. Tests were done according to the manufacturer’s instructions. Antibody Detection Passive Hemagglutination

Assay

The Candida albicans hemagglutination assay antigen test (Hoffmann-La Roche Diagnostica, Basel, Switzerland) was used to detect Candida antibody. When several difTable

2.

Serologic

Results Candidiasis,

among Patients Asymptomatic

(inters

Patients (n = 125) Candidiasis (n = 15) With Candida colonization (n = 38) Without Candida colonization (n = 72) Controls (n = 65) Hospital staff (n = 42) Shigellosis (n = 23) PHA = passive hemagglutination;

CIE = counterimmunoelectrophoresis.

PHA 2 7:32CJ) n (%)

leukemia;

ferent antibody titers were obtained from the same patient, the highest titer was used for the analysis. Counterimmunoelectropboresis

Assay

Two types of Candida antigens were used in the CIE assay. One was Candida antigen for the allergen test (Iorii Co., Tokyo, Japan), which consisted of non-mannan antigens, and the other was mannan prepared from a clinical isolate of Candida albicans, by the method of Reiss et al.’ The CIE assay was performed in a Gelman electrophoretic chamber (Gelman, Ann Arbor, MI, USA) with 8 X 10 cm glass slides: 0.8% agarose (Behringer, Germany) and 0.4% sodium citrate (Kant0 Kagaku, Japan) in a 0.07 M barbital buffer, pH 8.6 (Cosmo, Tokyo, Japan) was poured onto the glass and two wells measuring 3 mm in diameter were cut 3 mm apart. Antisera were loaded on the anodal side. Electrophoresis of samples was conducted for 60 minutes at room temperature, with a current of 20 mA, in 0.07 M of barbital buffer, pH 8.6. Precipitation lines were confirmed after cooling 16 hours at 4°C. The sensitivity, specificity, and predictive values were all calculated as described by Vecchio,* Feinsteiq9 and McNeill et al.1° with Hematologic Disorders Patients, and Controls Antibody

Group

79

and

Invasive

Test

Antigen C/E

Mannan

i 3 (87) 11 (29) 16 (22)

1 (7) 0 (0) 0 (0)

2 (4.7) 1 (4.3)

0 0 0 (0)

n (%j

Non-Mannan

n (%)

IO (67) 2 (5) 2 (3) 0 0

(0) (0)

(Titers

Test

Cand-Tee 2 1.4) n (%) 5 (33) 0 (0) 5 (7) 0 0

0 (0)

80

International

Journal

of Infectious

Diseases

/ Volume

RESULTS of Candida

Detection

Antibodies

By PHA, 13 (86.7%) of the 15 patients with invasive candidiasis had serum antibody titers of 1:320 or more, compared to 11 (28.9%) in group 1; I6 (22.2%) in group 2; 3 (4.7%) in group 3; and 1 (4.3%) in group 4 (Table 2). The sensitivity was 87%, and the specificity, positive predictive value (PVP), and negative predictive value (PVN) were 85%, 30%, and 99%, respectively (Table 3). By CIE, when the non-mannan antigen was used, the sera of 10 (66.7%) patients with invasive candidiasis yielded positive results, whereas 2 (5.0%) 2 (3.0%) and 0 patient samples were positive in groups 1, 2, and 3, respectively. The sensitivity was 67%, and the specificity, PVP and PVN were 98%, 71%, and 97%, respectively. When mannan antigen was used, only one (6.7%) patient with invasive candidiasis was positive (see Tables 2 and 3). of Candida

Detection

Antigenemia

A Cand-tee titer of 1:4 or greater was observed in 5 (33.3%) of the 15 patients with invasive candidiasis. On the other hand, 5 (6.9%) of the patients with hematologic disorders who lacked evidence of Candida infection or colonization (group 2) showed positive results, and none of the patients in groups 1, 3, and 4 were positive (see Table 2). The Cand-tee test showed a low sensitivity of 33.3%, but a specificity of 97%, PVP of 50%, and PVN of 95% (see Table 3). Other Observations Figure 1 shows Candida antibody titers by PHA or CIE with non-mannan antigen. When a positive result was scored only if both assays were positive, there were three false positives and five false negative cases. The sensitivity was 67%; the specificity, PVP, and PVN were 98%, 77%, and 97%, respectively (see Table 3). The sequential change of PHA titers in pairs of serum was also analyzed. This consisted mostly of serum specimens taken at admission and during or after a febrile episode (Figure 2). The antibody titers of 10 of 15 patients with invasive candidiasis were elevated twofold

1, Number

2, November

1996

or more, whereas only three cases among 57 patients without Candida infection showed antibody titer elevation. The results of the three antibody tests and the one antigen detection test in the 15 patients with invasive candidiasis are summarized in Table 4. In the detection of Candida antigen by Cand-tee, five (33%) of the patients were positive. Four patients with gastrointestinal candidiasis and three patients with pulmonary candidiasis were negative. In the antibody detection of candidiasis by CIE with non-mannan antigen, ten (67%) patients had positive results, and there were no correlations between the types of infection and positivity. When using mannan antigen, only one patient with candidiasis yielded a positive result, DISCUSSION The diagnosis of serious Candida infections remains problematic. Colonization of nonsterile sites may provide an opportunity for invasion; however, the detection of colonization does not identify those patients who will require systemic therapy. Antemortem blood cultures are negative in many patients with autopsy-proven disseminated Candida infections.

PHA titer 5120

e00

2560

00

0

0

1280

000

0

0

640

0

000

320

000

)00000

0000000

0

0

160

)0000000

000000C

)OOOO

0

80

)000000

000000

00000000

0

0

0000000

)OOOO

000000

)OOOO

C)oooooooc

000000000

)OOOO

0000000

000000000

0

00000 0000

40

00000 00000

20

Table

3.

Comparison of Invasive

of Assays Candidiasis

0

00

00000 ooooc

for Diagnosis 10

0

)OOOO

0000000

C/E 0

1oJ

PHA Sensitivity Specificity PVP (%) PVN (%)

(%) (%)

87 85 30 99

Non-Mannan 67 98 71 97

Mannan 7 100 100 93

PHA+CIE 67 98 77 97

PHA = passive hemagglutination; CIE = counterimmunoelectrophoresis; PVP = positive predictive value; PVN = negative predictive value.

Can&Tee 33 97 50 95

000

~-

Ci lndidiasis (ll = 15)

Asymptomatic with Candida colonies (ll = 38)

Asymptomatic without Candida colonies (n = 72)

Shigellosis (II = 23)

Hospital staff member (control) (n = 42)

0 = CIE (+); o = CIE (-)

PHA titer and CIE in the patient Figure 1. trol group (n = 65).

groups

(n = 125) and con-

Detection of Candida

PHA titer

1280

640

320

160

80

40

20

10

lo&

before

after

Candidask (n = 10)

Figure serum

(+)

2. The relation between specimens (n = 67).

before

afte1

Candidask (n = 57)

(-)

PHA titer and candidiasis

using

paired

In this study, tests for Candida antigen and antibodies were evaluated for diagnosis of invasive candidiasis among patients with hematologic disorders. Some Table

4.

Clinical

Features

Antigen and Antibody / Takaki et al

of these are commercially marketed tests that are commonly used to decide whether to treat patients for candidiasis. False negative results may have significant adverse consequences if they prompt physicians to withhold treatment of this life-threatening infection. False positive results may also be harmful as patients may needlessly receive toxic antifungal therapies. The PHA assay was sensitive: when titers in excess of 1:32O were defined as positive, 87% of patients with invasive candidiasis developed positive titers. Unfortunately, 4.5% of samples from healthy hospital workers and 29% and 22% of patients in control groups 1 and 2 also yielded positive results. Kostiala and Kostiala reported a false positive ratio of 3% using this method.” Bisbe et al showed that positive PHA tests with a titer in excess of 1:2560 strongly suggested a Candida infecti0n.l’ By this criteria, 6 (40%) of 15 patients with invasive candidiasis were positive, and two (1%) control patients without invasive candidiasis yielded false positive results. It has been suggested that a serologic diagnosis is difficult in those situations because the immunocompromised hosts cannot produce a sufficient amount of Candida antibody. 13s1*However, 9 of 15 patients with invasive candidiasis had a threefold or greater increase in antibody titers, whereas 3 of 57 asymptomatic patients showed antibody titer elevation of three times or more, and only one case had a titer of 1: 320 or higher. Thus the examination of PHA with paired serum samples has some diagnostic value. With the CIE method, non-mannan antigen or mannan antigen was used. The sensitivity and specificity with nonmannan antigen were 67% and 98%, whereas with mannan antigen they were 7% and lOO%, respectively Bisbe et al and Dee et al reported similar results with CIE.12J5 Antigen, rather than antibody detection has been advocated for early diagnosis of candidiasis. The Cand-tee test for Candida antigen was insensitive in this study. Its and Serologic

Results Antibody

Antigen NO.

invasive

Candidiasis

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Candidemia Candidemia Candidemia Candidemia Pulmonary Pulmonary Pulmonary GI-tract GI-tract GI-tract GI-tract Liver abscess Renal abscess Arthritis Urinary

CIE = counterimmunoelectrophoresis; - = negative result; + = positive

Test

Cand-Tee

Titer

I:8 I:4 1:4 I:2 1:2 I:2 1:2 1:4 I:4 PHA = passive result.

hemagglutination.

81

C/E Mannan + -

Non-Mannan + + f + + + + + + +

Test PHA Titer Peak 5120x 320x 1280x 1280x 1280x 2560x 10x 5102x 640x 5102x 20x 320x 320x 5120x 2560x

Rheumatoid Arthritis Factor +

82

International

Journal

of Infectious

Diseases / Volume

was 33%, and its specificity was 97% (PVp 50%; PVN, 95%). This suggests that reliance on this assay sensitivity might

result

in two cases missed

for every

case detected.

This low sensitivity makes the test an unacceptable guide for treatment decisions. Ness et al reported that the Cand-tee test had a sensitivity of 54.5%, a specificity of 29%, PVP of 17%, and PVN of 70.5%.16 Bougnoux et al also reported a sensitivity of 29% and a specificity of 97%.” Jones came to similar conclusions.18 If a 1:s or greater titer were considered to be positive, Fung et al reported a sensitivity of 71% and a specificity of 984o i9 but, using these criteria, data from the present study show a sensitivity of 7% and a specificity of 99%. By combining the PIIA and CIE with non-mannan antigen methods, higher diagnostic significance was obtained than with either method alone. The sensitivity of the combined tests was 67%, and the specificity was 98%. The antigens used in these tests were polysaccharides, including mannan in the PHA test and non-mannan proteins in the CIE test. The combination of two tests for Candida antibody was a sensitive and specific method for diagnosis of invasive candidiasis. These tests do not, however, provide an early diagnosis that reliably indicates when treatment should be initiated. With current tests, antigen detection misses most infected patients and antibody detection is too late. Neither test provides a reliable indicator of who should be treated and when treatment should start or stop. ACKNOWLEDGMENT The authors thank Kazuko Hosokawa

for secretarial

assistance.

REFERENCES 1. Kaufman L, Reiss E. Serodiangnosis of fungal diseases. In: Rose NR, et al, eds. Manual of clinical laboratory immunology. 3rd Ed. Washington: American Society for Microbiology 1986:446-466. 2. Kiehn TE, Bernard EM, Gold JWM, Armstrong D. Candidiasis: detection by gas-liquid chromatography of D-arabinitol, a fungal metabolite, in human serum. Science 1979; 206: 577-580. 3. Repentigny L, Kuykendall RJ, Chandler Fw, Broderson JR, Reiss E. Comparison of serum mannan, arabinitol, and

1, Number

2, November

1996

mannose in experimental disseminated candidiasis. J Clin Microbial 1984; 19:804-812. 4. Herent P Stynen D, Hernando E Fruit J, Poulain D. Retrospective evaluation of two latex agglutination tests for detection of circulating antigens during invasive candidosis. J Clin Microbial 1992; 30:2158-2164. 5. Gutierrez J, Maroto G, Piedrola G, Martin E, Perez JA. Circulating Candida antigens and antibodies: useful markers of candidemia. J Clin Microbial 1993; 31:2550-2552. 6. Kohno S, Mitsutake K, Yasuoka A, et al. An evaluation of serodiagnostic tests in patients with candidemia: beta-glucan, mannan, Candida antigen by Cand-Tee and n-arabinitol. Microbial Immunol 1993; 37:207-212. 7. Reiss E, Stones S, Hasenclever HE Serological and cellular immuno-activity of peptidoglucomannan fraction of Candida aZbicans cell walls. Infect Immun 1974; 9:881-890. 8. Vecchio TJ. Predictive value of a single diagnostic test in unselected population. N Engl J Med 1966; 274:1171-1173. 9. Feinstein AR. On the sensitivity, specificity, and discrimination of diagnostic tests, in clinical biostatistics. St.Louis: CV Mosby, 1977:214. 10. McNeill BJ, Keeler E, Adelstein SJ. Primer on certain elements of medical decision making. N Engl J Med 1975; 293: 211-215. 11. Kostiala AA, Kostiala I. Enzyme-linked immunosorbent assay (ELISA) for IgM, IgG, and IgA class antibodies against Candida albicans antigens: development and comparison with other methods. Sabouraudia 1981; 19:123-134. 12. Bisbe J, Miro JM, Torres LX, et al. Diagnostic value of serum antibody and antigen detection in heroin addicts with systemic candidiasis. Rev Infect Dis 1989; 11:310-315. 13. Filice GYB, Armstrong D. Immunodiffusion and agglutination tests for Candida in patients with neoplastic disease. J Infect Dis 1977; 135:349-357. 14. Renpentigny L. Serological techniques for diagnosis of fungal infection. Eur J Clin Microbial 1989; 8:362-375. 15. Dee TH, Johnson GM, Berger CS. Sensitivity, specificity, and predictive value of anti-Candida serum precipitin and agglutinin quantitation: comparison of counterimmunoelectrophoresis and latex agglutination. J Clin Microbial 1981; 13:750-753. 16. Ness MJ, Vaughan Wp Woods GL. Candida antigen latex test for detection of invasive candidiasis in immunocompromised patients. J Infect Dis 1989; 159:495-502. 17. Bougnoux ME, Hill C, Moissenet D, et al. Comparison of antibody antigen, and metabolite assays for hospitalized patients with disseminated or peripheral candidiasis. J Clin Microbial 1990; 28:905-909. 18. Jones JM. Laboratory diagnosis of invasive candidiasis. Clin Microbial Rev 1990; 3:32-45. 19. Fung JC, Donta ST, Tilton RC. Candida detection system (Cand-Tee) to differentiate between Candida albicans colonization and disease. J Clin Microbial 1986; 24:542-547.