Serum Immunoglobulin G Avidity Test As a Tool for Improving Diagnosis of Cat Scratch Disease

Abstract

Cat scratch disease (CSD) is a common cause of subacute infectious regional lymphadenitis, caused by Bartonella henselae. Presently, detection of anti-B. henselae antibodies by immunofluorescence antibody assay or enzyme immunoassay (EIA) is the most widely used diagnostic test for CSD, but both are limited in establishing the timing of infection with B. henselae. In the present work we developed an avidity test for anti-B. henselae immunoglobulin G (IgG) based on EIA to distinguish recent from past CSD. We used 101 serum samples from 79 CSD patients with positive anti-B. henselae IgG as verified by EIA, and systematically developed an avidity assay using various detergent (urea) concentrations and incubation settings to optimize the test conditions to differentiate early CSD (less than 12 weeks) from late CSD (12 weeks or more). After serial experiments, the optimal conditions for performing the avidity test included incubation for 10 min at room temperature with 8 M urea at pH 7.4, and these parameters were used in the study. Our experiments showed that while the avidity indexes (AIs) of the early CSD samples were widely distributed, all the late CSD sera samples had AIs above 43, indicating that an AI <43 can serve as evidence of early CSD. The results of this study indicate that the avidity test can be useful in the serodiagnosis of CSD, particularly when anti-B. henselae IgM antibodies are not detected.

Introduction

C at scratch disease (CSD), is a common cause of subacute infectious regional lymphadenitis, and is caused by Bartonella henselae (Anderson and Neuman 1997). The disease is often associated with systemic symptoms such as fever and malaise, and approximately 10% of patients may develop atypical manifestations, such as Parinaud's oculoglandular syndrome, neuroretinitis, encephalitis, endocarditis, granulomatous hepatitis, and osteomyelitis (Anderson and Neuman 1997). CSD lymphadenopathy lasts for an average of 7 weeks (range 1–87 weeks) (Metzkor-Cotter et al. 2003), and may resemble a malignant process, producing concerns that often leads to unnecessary, sometimes invasive, diagnostic procedures (Wong et al. 1996). Timely and accurate diagnosis is therefore important.

The diagnosis of CSD can be complicated. The traditional CSD skin test is not licensed for routine use. Cultures for B. henselae from affected lymph nodes require invasive procedures such as fine-needle aspiration or excision biopsy, and are almost always negative. Other diagnostic tests that can be performed on lymph node aspirates or biopsies include histopathology, Warthin-Starry silver stain, or immunohistochemistry, all of which have variable sensitivity and specificity, or polymerase chain reaction (PCR), which although considered the most accurate, is still not universally available for routine diagnostic testing.

Detection of anti-B. henselae antibodies, particularly by immunofluorescence antibody (IFA) assay, has become the most widely used diagnostic test for CSD. This methodology, however, suffers from inter-observer variation, and is not suitable for screening large numbers of specimens. In addition, the majority of the IFA assays, including the one that is performed at the Centers of Disease Control and Prevention in the U.S., detects only immunoglobulin (Ig) G and not IgM, and thus is limited in differentiating acute from past infection (Dalton et al. 1995).

An enzyme immunoassay (EIA) for anti-B. henselae IgG and IgM antibodies has been described recently by our group in a large cohort of patients with established CSD. This study determined the EIA sensitivity to be 75% for anti-B. henselae IgG alone, 48% for IgM alone, and overall 85% for both IgG and IgM. The specificity for IgG and IgM was high, 98% and 99%, respectively (Giladi et al. 2001). Monitoring the kinetics of anti-B. henselae antibodies has demonstrated that patients remained IgM-positive for up to 3 months after disease onset, whereas 25% of the patients had anti-B. henselae IgG for more than a year (Metzkor-Cotter et al. 2003). Since the IgM response is short-lived, it is often missed because the majority of CSD patients are usually tested later in the course of the disease. Therefore, additional diagnostic tools are clearly indicated to corroborate the diagnosis of acute CSD in IgG-positive IgM-negative patients.

The IgG avidity test has been shown to be a powerful tool for diagnosing early infections in various clinical settings, especially in situations where timing and differentiation of primary and secondary infections is important (Bodeus et al. 2002; Andrews 2004). The basis for this assay is the observation that during acute infection, IgG antibodies bind antigen weakly (i.e., have low avidity), whereas chronically-infected patients have stronger binding (high avidity) antibodies (Eisen and Siskind 1964). Tests for IgG avidity assays use a detergent (mostly urea solution) which disrupts the antigen-antibody reaction differentially, affecting low-avidity antibodies more than high-avidity antibodies, thus distinguishing recent primary from distant (high-avidity) infection. The IgG avidity tests have been developed and applied successfully for the laboratory diagnosis of infections caused by a variety of viral, bacterial, or parasitic pathogens (Guigno et al. 1992; Jenum et al. 1997; Baccard-Longere et al. 2001; Bodeus et al. 2002). The purpose of the present study was to develop an IgG avidity test to assist in distinguishing recent from past CSD.

Materials and Methods

Patients and sera

A total of 101 frozen serum samples from 79 patients with CSD were retrospectively selected for this study. Patient demographic, epidemiologic, and clinical data, including the time of serum sampling during the disease course, were collected as part of a national CSD surveillance study as previously described (Giladi et al. 2001). For the purpose of this study a case of CSD was defined as a patient with regional lymphadenopathy, history of recent cat contact, and positive EIA for anti-B. henselae IgG, all in the absence of another diagnosis.

B. henselae IgG avidity assay

The avidity assay was constructed based on the EIA we previously developed for anti-B. henselae antibodies (Giladi et al. 2001; Metzkor-Cotter et al. 2003). Briefly, a sacrosyl-insoluble, outer-membrane protein extract of B. henselae 87-66 strain (ATCC 49793) was used to coat microtiter plates. Serum samples were diluted 1:100 in PBS with 0.1% Tween 20 (PBST), and incubated in triplicates with the antigen for 60 min at 35°C, with and without urea. To optimize the test conditions we compared various urea concentrations (4.5, 6, 7, 8, and 9 M), durations of incubation (5, 10, and 30 min), temperature (28°C and room temperature), and pH of the urea solution (8.1 and 7.4). Following incubation, the plates were washed three times with PBST and bound IgG was detected colorometrically by alkaline phosphatase-conjugated goat anti-human IgG (Sigma-Aldrich, St. Louis, MO), and the substrate (p-nitrophenyl phosphate; Sigma-Aldrich). Optical density (OD) was measured at 405 nm by an automatic ELISA reader. Serum from a previously identified PCR-confirmed individual with CSD, diluted 1:300 to give a reading of 1 OD, was chosen as a control serum for all assays. The binding of anti-B. henselae IgG in test samples was expressed as arbitrary units (AU), calculated according to the following equation: (mean OD of tested sample − background/mean OD of a positive control serum − background)×100.

Samples having 60 AU or more were considered positive for the presence of anti-B. henselae IgG antibodies, whereas samples with less than 60 AU were considered negative. For the avidity assay, samples tested by the EIA were also incubated in parallel in the presence of urea. The results were expressed as the avidity index (AI), calculated as follows: (AU in the presence of urea/AU in the absence of urea)×100.

For the purpose of this study, samples obtained at less than 12 weeks of disease onset were defined as early CSD, whereas samples drawn after 12 weeks or more of disease onset were considered late CSD.

Statistical analysis

Differences between groups were analyzed with the chi-square test or Fisher's exact test for categorical variables, and with Student's t-test or the Mann-Whitney U test for continuous variables. A two-sided p value of <0.05 was considered significant. Receiver operating characteristic (ROC) analysis was used to assess the performance of the avidity test.

Results

Characteristics of the 79 patients selected for the study are shown in Table 1. As mentioned above, all the patients presented with positive anti-B. henselae IgG, while 70 of the 79 patients (89%) had additional confirmatory tests, 61 (77%) had a positive EIA for anti-B. henselae IgM, 22 (28%) had positive PCR, and 13 (16%) had both positive IgM and PCR.

Table 1.

Characteristics of the 79 Cat Scratch Disease Study Patients

Variable	No. of patients	(%)
Age (years)
<20	31	(39)
20–39	29	(37)
≥40	19	(24)
Cat contact	79	(100)
Lymphadenopathy	79	(100)
Location of affected lymph nodes
Axillary	23	(29)
Inguinal	17	(22)
Cervical	8	(10)
Epitrochlear+axillary	8	(10)
Other	8	(10)
Unknown	15	(19)
Fever	48	(61)
Primary lesion	24	(30)
Time from onset of symptoms to serodiagnosis^a
<12 weeks	56	(71)
≥12 weeks	30	(38)
Positive serology for B. henselae
Anti-B. henselae IgG	79	(100)
Anti-B. henselae IgM	61	(77)
Positive PCR results for B. henselae	22	(28)

Seven patients had serum samples of both early (<12 weeks) and late (≥12 weeks) disease.

PCR, polymerase chain reaction; IgG, immunoglobulin G; IgM, immunoglobulin M.

Experiments were carried out to determine the optimal conditions by which the avidity assay could best differentiate between early and late CSD samples.

Table 2 shows the results of avidity tests carried out at various urea concentrations, using serum samples obtained from the same PCR-positive CSD patient, at weeks 4, 37, 46, and 59 after disease onset. The results show that AI was considerably lower in the early CSD sample (sample 1) compared with the late CSD samples (samples 2–4), at the urea concentrations that were tested. While no statistically significant differences were observed between the results achieved with different urea concentrations, those of 6 and 8 M at pH 7.4 appeared to differentiate early from late CSD better than 4.5 M testing, and thus were selected for further experimentation.

Table 2.

Avidity Testing at Various Urea Concentrations

Sample no.	Time from onset of disease (weeks)	Sample classification	Avidity index (AI)
			4.5 M urea	6 M urea	8 M urea
1	4	Early disease^a	48	33	19
2	37	Late disease^b	72	63	45
3	46	Late disease	70	62	46
4	59	Late disease	73	63	44

<12 weeks from disease onset.

≥12 weeks from disease onset.

Table 3 shows results of the avidity test performed on samples of 6 CSD patients, of which 3 samples were from early CSD patients (drawn at weeks 1, 2, and 3 after disease onset), and 3 samples were from late CSD patients (drawn at weeks 18, 23, and 46 after disease onset). A urea concentration of 8 M at pH 7.4 was superior to that of 6 and 8 M at pH 8.0. Avidity test conditions, including incubation for 10 min at room temperature with 8 M urea at pH 7.4, were selected for the testing of multiple samples of early and late CSD, as these were found to be slightly superior to the other combinations used for the distinction between early and late samples.

Table 3.

Avidity Testing at Various pH Conditions

			8 M urea pH 7.4		8 M urea pH 8.0		6 M urea pH 8.0
Patient no.	Time from onset of disease (weeks)	Sample classification	AI	Mean AI	AI	Mean AI	AI	Mean AI
1	1	Early disease^a	34	34	34	34	48	45
2	2		32		33		38
3	3		36		35		49
4	18	Late disease^b	72	73	75	70	81	78
5	23		66		62		69
6	46		80		74		85

<12 weeks from onset of cat scratch disease.

≥12 weeks from onset of cat scratch disease.

AI, avidity index.

Figure 1 shows the results of avidity testing of 101 serum samples from 79 CSD patients. Sixty-five of these samples were classified as early CSD, and 36 samples were defined as late CSD. The AIs of the early CSD samples were widely distributed, ranging between 16 and 82, whereas all the late CSD sera samples had AIs of 43 or above, ranging between 43 and 91. ROC analysis showed that the area under the curve is 0.76 (95% CI 0.67–0.85; p<0.001). The threshold of AI<43 provided sensitivity of 51% with specificity of 100% for defining early CSD; shifting the threshold to AI <47 resulted in sensitivity of 57%, with specificity decreasing to 90%. Further changes in the AI cutoff resulted in a significant drop in specificity, with only borderline improvement in sensitivity, since a considerable number of patients had high AI during early disease. Therefore, to address the clinical goal of defining early CSD with maximal specificity, the threshold of AI <43 was selected as an appropriate cut-off.

FIG. 1.

Avidity testing of serum samples of early and late CSD patients. The dotted line designates AI of 43. Serum samples that had AI below this line were all from early CSD patients.

Since it was recently reported elsewhere that a high titer of IgG antibodies in sera of acute patients can reduce the dissociation effect of urea, resulting in high AI (Dangel et al. 2006), CSD samples in the present study having more than 100 AU were diluted to contain 60–100 AU, and retested by the avidity assay. Altogether a total of 37 samples (21 with early and 16 with late CSD) were diluted and retested. No significant change in AI was seen in any of the late CSD samples after titration. Nevertheless, in 3 of the early CSD samples, having AIs of 78, 53, and 73, the avidity dropped below the 43 cut-off value, to 27, 36, and 42, respectively.

Analysis of the possible influence of the demographic, clinical, and laboratory characteristics (age, sex, history of fever or a primary lesion, and the presence of positive IgM on EIA or PCR for B. henselae) of the CSD patients on the avidity index showed no significant difference between groups of individuals with high (late or early disease) versus low (early disease) IgG AI (p>0.5 for all comparisons).

Discussion

Patients with CSD often seek medical help weeks or even months after disease onset (Giladi et al. 2001). When diagnosed, it is usually difficult to differentiate acute from past B. henselae infection. The aim of the present study was to develop an IgG avidity test to improve our ability to differentiate between early and late CSD. For this purpose we retrospectively selected sera of patients with established CSD, strictly defined by clinical and laboratory criteria, and divided them into early and late CSD samples based on the information that was collected from the patients and their physicians. We classified them as early CSD (12 weeks or less of disease onset) and late CSD (more than 12 weeks of disease onset), based on our previous findings that in sera of acute CSD patients anti-B. henselae IgM antibodies remain an average of up to 3 months after disease onset (Metzkor-Cotter et al. 2003).

The avidity tests are currently widely and effectively implemented in the diagnosis of cytomegalovirus (CMV) and Toxoplasma gondii infections when the knowledge of timing of exposure to these pathogens is essential, as it is in pregnant women (Jenum et al. 1997; Andrews 2004). Several different avidity assays are available for these diseases, and each one of them has been properly validated (Jenum et al. 1997; Baccard-Longere et al. 2001). The avidity test results are occasionally non-linearly distributed over time, with a rather broad range of AI values seen at the early phase after exposure to CMV or T. gondii. This limitation is resolved by defining precise diagnostic thresholds for each of these assays, taking into consideration the distribution of the avidity results at different stages after seroconversion (Jenum et al. 1997; Baccard-Longere et al. 2001; Bodeus et al. 2002; Prince and Leber 2002). This enables physicians to use the avidity tests as important laboratory tools to guide further and more invasive diagnostic steps (such as amniocentesis), in clinical scenarios of CMV or T. gondii infection in pregnant women. We tested 101 serum samples from 79 CSD patients. All patients with low AIs (<43) had disease for less than 12 weeks, whereas for patients with high AIs (≥43), it was impossible to distinguish early from late disease, since a significant number of these patients had symptoms for less than 12 weeks. It is possible, however, that CSD patients, who usually seek medical help weeks or even months after disease onset, may have been imprecise in determining the onset and duration of the disease, which may partially explain the high AIs (>43) we found in the early CSD patients. However, it has been shown previously with other infections, such as CMV and T. gondii, that high avidity can also be found occasionally in acutely-infected patients (Jenum et al 1997; Lazzarotto et al. 1997; Baccard-Longere et al. 2001; Bodeus et al. 2002).

While detection of anti-B. henselae antibodies is the most widespread diagnostic method for CSD, the serology has significant limitations in distinguishing early from late infection for several reasons: (1) both EIA and IFA assays have relatively low sensitivity for the detection of anti-B. henselae IgM; (2) the anti-B. henselae IgM response is often short-lived and may disappear by the time the serological diagnosis is attempted; and (3) although the level of anti-B. henselae IgG decreases over time, the rate of the IgG decay is usually too slow and thus precludes the use of the dynamics of anti-B. henselae IgG titers as a diagnostic tool for follow-up after acute infection (Szelc-Kelly et al. 1997; Metzkor-Cotter et al. 2003). The IgG avidity test developed in the present study can serve as an additional test to avoid the aforementioned limitations.

In summary, the results of this study indicate that the avidity test can play an important role in the serodiagnosis of early CSD. This is particularly true when anti-B. henselae IgM antibodies are not detected. Although high (≥43) AIs do not distinguish early from late disease, low AIs (<43) support the diagnosis of recently acquired CSD.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

References

Anderson

, Neuman

. Bartonella spp. as emerging human pathogens. Clin Microbiol Rev, 1997; 10:203–219.

Andrews

. Diagnosis of fetal infections. Curr Opin Obstet Gynecol, 2004; 16:163–166.

Eisen

, Siskind

. Variations in affinities of antibodies during the immune response. Biochemistry, 1964; 3:996–1008.

Baccard-Longere

, Freymuth

, Cointe

et al. Multicenter evaluation of a rapid and convenient method for determination of cytomegalovirus immunoglobulin G avidity. Clin Diagn Lab Immunol, 2001; 8:429–431.

Bodeus

, Van Ranst

, Bernard

et al. Anticytomegalovirus IgG avidity in pregnancy: a 2-year prospective study. Fetal Diagn Ther, 2002; 17:362–366.

Dalton

, Robinson

, Cooper

et al. Use of Bartonella antigens for serologic diagnosis of cat-scratch disease at a national referral center. Arch Intern Med, 1995; 155:1670–1676.

Dangel

, Bader

, Enders

. Improvement of cytomegalovirus avidity testing by adjusting the concentration of CMV-specific IgG in test samples. J Clin Virol, 2006; 35:303–309; Epub 2005 Nov 18.

Giladi

, Kletter

, Avidor

et al. Enzyme immunoassay for the diagnosis of cat-scratch disease defined by polymerase chain reaction. Clin Infect Dis, 2001; 33:1852–1858; Epub 2001 Oct 23.

Guigno

, Coupland

, Smith

et al. Primary humoral antibody response to Coxiella burnetii, the causative agent of Q fever. J Clin Microbiol, 1992; 30:1958–1967.

10.

Jenum

, Stray-Pedersen

, Gundersen

. Improved diagnosis of primary Toxoplasma gondii infection in early pregnancy by determination of antitoxoplasma immunoglobulin G avidity. J Clin Microbiol, 1997; 35:1972–1977.

11.

Lazzarotto

, Spezzacatena

, Pradelli

et al. Avidity of immunoglobulin G directed against human cytomegalovirus during primary and secondary infections in immunocompetent and immunocompromised subjects. Clin Diagn Lab Immunol, 1997; 4:469–473.

12.

Metzkor-Cotter

, Kletter

, Avidor

et al. Long-term serological analysis and clinical follow-up of patients with cat scratch disease. Clin Infect Dis, 2003; 37:1149–1154; Epub 2003 Oct 7.

13.

Prince HE Leber

. Validation of an in-house assay for cytomegalovirus immunoglobulin G (CMV IgG) avidity and relationship of avidity to CMV IgM levels. Clin Diagn Lab Immunol, 2002; 9:824–827.

14.

Szelc-Kelly

, Goral

, Perez-Perez

et al. Serologic responses to Bartonella and Afipia antigens in patients with cat scratch disease. Pediatrics, 1995; 96:1137–1142.

15.

Wong

, Kruskal

, Kane

et al. Cat-scratch disease simulating lymphoma. J Comput Assist Tomogr, 1996; 20:165–166.