Comparative evaluation of the Brucella Coombs gel test in laboratory diagnosis of human brucellosis

ABSTRACT

Brucellosis is widespread among humans and animals. Diagnosis of brucellosis mostly depends on serological methods. Serological tests are preferred over time-consuming and hazardous bacterial cultures in routine laboratory practice. However, these tests are somehow challenging due to ‘incomplete/blocking antibodies’ that prevent agglutination. Brucella Coombs gel test (BCGT) is newly developed test that contains Coombs antibodies (anti-human IgG) in gel microtubes and depends on gel centrifugation methods for the serological diagnosis of brucellosis. Here, performance of the BCGT is compared with standard serum tube agglutination (STA), standard serum tube agglutination with Coombs (C-STA) and immune capture agglutination (Brucellacapt). In total, 78 positive samples for study group and 16 samples for the control group were enrolled in the study. The samples were tested at dilutions of 1:40–1:5120. Titres at 1:160 and above were considered positive for brucellosis, whereas those lower than 1:160 were considered negative. Excellent agreement levels were determined between BCGT test and C-STA (κ, 0.894; p < 0.001), and BCGT and Brucellacapt (κ, 0.802; p < 0.001), when the diagnostic titre was accepted as 1:160. BCGT is easy to apply and interpret and provides reliable titre results in less than 2 h. It is also advantageous for screening.

KEYWORDS:

• Bacterial infection
• diagnosis
• equipment
• brucellosis

Introduction

Brucellosis is a common bacterial zoonotic disease in Turkey and around the world. Despite its very low incidence in developed countries, brucellosis remains an important public health issue in developing countries. In addition, 500,000 new cases are reported worldwide each year.[1–7] Brucellosis is widespread among humans and animals, especially in Mediterranean countries, the Arabian Peninsula, India, Africa and South America.[3,5,8,9] Transmission occurs by consumption of unpasteurized milk and dairy products (e.g. raw milk, soft cheese, yogurt, ice cream) or food products, such as undercooked meat from infected animals. Typically, professionals who come into contact with infected animals or elicitation of infected animals via respiratory, conjunctival, or mucous membranes are at risk for brucellosis. In addition, laboratory professionals who handle infected materials and travellers visiting the endemic areas are at risk of infection.[1–4,6]

Brucella species are Gram-negative, facultative intracellular bacteria. They lack capsules, flagellae, endospores and native plasmids.[10,11] Brucella abortus, Brucella melitensis and Brucella suis are the most common species of this genus responsible for human infection.[3,4] In humans, Brucella infections emerge with heterogeneous clinical presentations as acute, subacute and chronic infections, but are rarely fatal. Symptoms of infection are often non-specific, and long-term recurrent fever (undulating fever) is the most common clinical manifestation. Acute brucellosis can occur with a variety of symptoms or as insidious disease, whereas chronic illness is usually insidious and may continue for years without obvious symptoms. Therefore, diagnosis may not depend only on clinical manifestations; microbiological testing is required for an accurate diagnosis. Microbiological diagnosis depends on several serological tests, molecular methods and isolation of the bacteria. Isolation from clinical specimens, such as blood or bone marrow, is the gold standard of testing. Although automated culture systems shorten the isolation period, they are time-consuming, unavailable in several laboratories, and their sensitivity may decrease as the number of circulating bacteria increases. They also constitute a risk for laboratory-acquired infections. Therefore, serological tests predominate at diagnosis.[3,11,12] The most commonly used serological tests are the Rose Bengal (RB) plate test and the standard serum tube agglutination (STA) test.[3,11] However, these commonly used tests cannot detect so-called ‘incomplete/blocking antibodies’ that prevent agglutination, especially in chronic patients. In other words, these tests may lead to false-negative results. Therefore, to overcome this problem, anti-human globulin (Coombs antibody) was added to the STA (C-STA) and immune capture agglutination (ICA-Brucellacapt) tests, in which the surfaces of the microplate wells were coated with Coombs antibodies.[1,3–5,10–12] The sensitivities and specificities of these two tests are similar to each other.[10]

The Brucella Coombs gel test (BCGT) is a newly developed test that depends on gel centrifugation methods for the serological diagnosis of brucellosis. In this study, we evaluated the diagnostic performance of the BCGT and compared to other serological methods.

Materials and methods

A total of 78 serum samples from patients with a preliminary clinical diagnosis as brucellosis were enrolled in this study between June 2013 and December 2014. Samples were first evaluated with the RB test (Seromed, Istanbul, Turkey), STA test (Seromed, Istanbul, Turkey) and ICA-Brucellacapt® (Vircell Company, Santa Fé, Granada, Spain).

Any patient's sera, which were positive with any of the mentioned tests, were included into the study group. When all the tests were as negative, such samples were included in the control group. In total, 78 positive samples for the study group and 16 samples for the control group were selected and then the samples were stored at −80 °C. In addition to the previous tests, C-STA (anti-human IgG, Tulip, Mumbai, India) and BCGT (ODAK Brucella Coombs gel test, Toprak Medikal, Istanbul, Turkey) were performed simultaneously. The samples were tested at dilutions of 1:40–1:5120. All of the tests were performed according to the manufacturer's recommendations. All the negative and positive controls, which were recommended by the manufacturer, within the individual test, were tested prior to the testing with patients’ sera. Titres at 1:160 and above were considered positive for brucellosis, whereas those lower than 1:160 were considered negative.[1,4,5,13]

Brucella Coombs gel test

In this method, serum dilutions were performed in microplate wells. The first well was filled with 5 μL serum and 100 μL diluent, and other wells were filled with 50 μL diluent. All of the wells were serially diluted, and 50 μL pink brucella antigen suspension was added to each well. The manufacturer recommends the use of at least four wells ranging from 1:40 to 1:320 dilutions for routine laboratory practices. When needed, these ranges can be expanded. In this study, we implemented dilution ranges of 1:40–1:5120 for comparison purposes. After dilution, prepared samples (50 μL) were transferred to gel matrix microtubes containing Coombs antibodies. Then, gel matrix cards were centrifuged for 20 min, as recommended by the manufacturer. Results were assessed visually. In the absence of brucella antibodies, pink brucella antigens collapse to the bottom of the tube, and the result is negative. When antibodies are present, a pink antigen-antibody complex layer is visible on the top of the microtubes, and the result is positive (Figure 1).

Figure 1. Negative and positive results in the Brucella Coombs gel test (www.toprakmedikal.com/documents/Brucella.ppt).

For the screening purposes, 5 μL serum and 100 μL diluent were suspended in microplate wells, and 50 μL of the mixture was discharged. Next, 50 μL pink brucella antigen suspension was added. The mixture was transferred to gel matrix microtubes. The remainder of the test was performed as described above.

Statistical analysis

Descriptive analyses were performed to provide information on the general characteristics of the study population. Categorical variables are presented as both count and percentage. Cohen's kappa (κ) coefficient was used to determine concordance between BCGT and other methods.[14] The sensitivity, specificity and positive and negative predictive values were calculated to determine the diagnostic performance of BCGT relative to Brucellacapt, STA and C-STA. A p value < 0.05 was taken to indicate statistical significance. All of the analyses were performed using IBM SPSS Statistics version 22.0 software (Armonk, NY: IBM Corp.; MedCalc Statistical Software version 15.6.1, MedCalc Software bvba, Ostend, Belgium). The kappa coefficient was interpreted as follows: 0.01–0.20, slight agreement; 0.21–0.40, fair agreement; 0.41–0.60, moderate agreement; 0.61–0.80, substantial agreement; and 0.81–0.99, excellent/almost perfect agreement.[15]

Results and discussion

In total, 28 of 94 patients (29.8%) were male, and 66 (70.2%) were female. The mean age of patients was 42.48 ± 15.72 years.

Agreement analysis of STA, C-STA and Brucellacapt with BCGT is listed in Table 1. Diagnostic performance of Brucella Coombs gel test and other assays is presented in Table 2.

Table 1. Agreement analysis between the Brucella Coombs gel test and immune capture agglutination, standard tube agglutination and standard tube agglutination with Coombs.

		Standard tube agglutination
		NEG	1/20	1/40	1/80	1/160	1/320	1/640	1/1280	1/5120	TOTAL
Coombs gel test	NEG	16	–	–	–	–	–	–	–	–	16
	1/40	6	8	2	–	–	–	–	–	–	16
	1/80	2	4	4	–	–	–	–	–	–	10
	1/160		2	10	2	12	–	–	–	–	26
	1/320	–	–	6	2	10	4	–	–	–	22
	1/640	–	–	–	–	–	–	2	–	–	2
	1/1280	–	–	–	–	–	–	–	–	–	0
	1/2560	–	–	–	–	–	–	–	–	–	0
	1/5120	–	–	–	–	–	–	–	–	2	2
κ: 0.230, p < 0.001
		Standard tube agglutination with Coombs
		NEG	1/20	1/40	1/80	1/160	1/320	1/640	1/1280	1/5120	TOTAL
Coombs gel test	NEG	16	–	–	–	–	–	–	–	–	16
	1/40	–	–	16	–	–	–	–	–	–	16
	1/80	–	–	2	8	–	–	–	–	–	10
	1/160		–	–	4	22	–	–	–	–	26
	1/320	–	–	–	–	2	20	–	–	–	22
	1/640	–	–	–	–	–	–	2	–	–	2
	1/1280	–	–	–	–	–	–	–	–	–	0
	1/2560	–	–	–	–	–	–	–	–	–	0
	1/5120	–	–	–	–	–	–	–	–	2	2
κ: 0.894, p < 0.001
		Immune capture agglutination
		NEG	1/40	1/80	1/160	1/320	1/640	1/1280	1/2560	1/5120	TOTAL
Coombs gel test	NEG	14	2	–	–	–	–	–	–	–	16
	1/40	1	15	–	–	–	–	–	–	–	16
	1/80	–	2	8	–	–	–	–	–	–	10
	1/160	–	–	1	21	4	–	–	–	–	26
	1/320	–	–	–	3	17	2	–	–	–	22
	1/640	–	–	–	–	–	2	–	–	–	2
	1/1280	–	–	–	–	–	–	–	–	–	0
	1/2560	–	–	–	–	–	–	–	–	–	0
	1/5120	–	–	–	–	–	–	–	–	2	2
κ: 0.802, p < 0.001

Table 2. Diagnostic performance of Brucella Coombs gel test and other assays.

		Brucella Coombs gel test
Methods	Titers	≤1/80	≥1/160	Sensitivity	Specifity	PPV^a	NPV^b	κ	p
Immune capture agglutination	≤1/80	42	0	100.0	0.977	0.981	100.0	0.979	<0.001
	≥1/160	1	51
Standard tube agglutination	≤1/80	42	0	100.0	0.656	0.577	100.0	0.549	<0.001
	≥1/160	22	30
STA with Coombs	≤1/80	42	0	100.0	0.913	0.923	100.0	0.915	<0.001
	≥1/160	4	48

^aPositive predictive value; ^bnegative predictive value.

Results according to RB

RB slide agglutination tests were positive in 68 serum samples. Within these positive samples, STA, Brucellacapt, BCGT and C-STA positively detected 28 (32.3%), 48 (70.5%), 50 (70.3%) and 46 (67.6%) samples, respectively. RB tests were negative in 26 samples. STA, Brucellacapt, BCGT and C-STA methods indicated 24 of the 26 (92.3%) negative samples were indeed negative.

Results according to STA

When diagnostic titre (≥1:160) was considered, 30 samples tested positive with STA. All of these 30 samples (30, 100%) were detected as positive with C-STA, Brucellacapt and BCGT. Sixty four samples were negative (0–1:80) with STA. Of these 64 negative samples, 46(71.8%), 43 (67%) and 42 (65.6%) were considered negative with C-STA, Brucellacapt and BCGT.

Results according to C-STA

When diagnostic titre (≥1:160) was considered, 48 samples tested positive with C-STA. Of these 48 samples, 30 (62.5%), 47 (98%) and 48 (100%) were proven to be positive with STA, Brucellacapt and BCGT, respectively. Forty-six samples were negative (0–1:80) with C-STA. Of these 46 negative samples, 46 (100%), 42 (91%) and 42 (91%) were considered negative with STA, Brucellacapt and BCGT, respectively.

Results according to Brucellacapt

When diagnostic titre (≥1:160) was considered, 51 samples were considered positive with Brucellacapt. Of these, 30 (59%), 47 (92%) and 51 (100%) were positive with STA, C-STA and BCGT, respectively. Forty-three samples were negative (0–1:80) with Brucellacapt. Of these, 42 (97.6%) were considered negative with STA, C-STA and BCGT.

Clinical signs and symptoms are insufficient for a dependable diagnosis of human brucellosis. Therefore, microbiological methods are required. Serological tests are preferred in routine laboratory practice, because they are faster, convenient and highly sensitive compared to time-consuming and hazardous bacterial cultures. Several serological tests have been developed for the diagnosis of brucellosis, including STA, C-STA, Brucellacapt, indirect fluorescence antibody and enzyme-linked immunosorbent assay (ELISA).

During the first week of infection, IgM antibodies against the bacterial lipopolysaccharide antigen emerge in the serum, followed by IgG antibodies after two weeks. Both types of antibodies (IgM and IgG) reach their highest levels in four weeks.[10,13] Agglutination titres > 1:160 or a four-fold rise of titres in follow-up sera are indicative of an active infection. Serological tests are preferred at the initial diagnosis, as well as during follow-up treatment.[13] Diagnostic levels of antibody titres can be detected in successfully treated blood culture-negative individuals for months or even years after acute brucellosis. In addition, due to continuous exposure to the bacteria/bacterial antigens of the population in endemic areas, high and persistent antibody levels can be determined. Therefore, diagnostic antibody titre levels may be elevated in such regions.[4,13]

BCGT is a recently developed, manufactured and commercialized method in our country. Very few studies concerning BCGT are documented in the literature (to the best of our knowledge, there is just one article and one poster presentation).[10,16] Of these reports, only one examined the diagnostic performance of BCGT relative to other methods. In addition to these publications, a review article stated that BCGT was as reliable as C-STA.[17] However, it is not appropriate to draw a conclusion based on a few studies. Therefore, we compared the diagnostic performance of BCGT with other serological tests.

RB is a qualitative, highly specific screening test. It is easy to use (preferred in endemic areas) and provides results within minutes.[3,5,13,18] Some authors highlight its low sensitivity. Controversially, other authors emphasize that RB is highly sensitive.[3,5,13] Test medium pH and other factors related to laboratory personnel may also adversely affect the test results. Moreover, when RB is positive, additional quantitative serological tests are required for confirmation and to determine antibody titres. Therefore, we did not compare the diagnostic performances and concordance of RB with BCGT in this study.

The agreement value between BCGT and STA was low for all dilutions ranging from 1:40 to 1:5120 (κ, 0.230, fair agreement; p < 0.001). Low-level agreement values were also determined between STA/C-STA (κ, 0.256, p < 0.001) and STA/Brucellacapt (κ, 0.205; p < 0.001). When the diagnostic titre 1:160 was considered, a moderate agreement between BCGT and STA was detected (κ, 0.549; p < 0.001) (Table 2). However, even at a diagnostic titre, there was fairly low agreement between STA/C-STA (κ, 0.370; p < 0.001) and STA/Brucellacapt (κ, 0.260; p < 0.001).

STA is unreliable during the early phase of infection. Moreover, the presence of high levels of antibodies and/or incomplete/blocking antibodies may cause false-negative results. In addition, cross-reactivity caused by other Gram-negative bacilli may lead to false-positive reactions.[3–5,10,13,19,20] Such disadvantages related to STA could lead to low concordance between STA and BCGT. To overcome these problems, Coombs antibody was added to STA. Alternatively, a one-step immune capture assay (Brucellacapt), in which microplate wells are coated with Coombs antibodies, is also widely used. These two methods have higher sensitivity and specificity than STA alone, and many authors have reported high-level agreement.[10,19–22] In some studies, Brucellacapt has been found to be more sensitive than C-STA.[19,23] In this study, a substantial level of agreement was determined between Brucellacapt and C-STA (κ, 0.723; p < 0.001). However, some studies have reported higher agreement rates.[10] Despite the useful and superior aspects of these two methods, there are also some disadvantages. First, they are labour-intensive and time-consuming (Brucellacapt needs 18–24 h incubation and C-STA needs up to 48 h). In addition, they are not practical or convenient for screening populations.[3,5,13,19,22]

BCGT does not require incubation; all of the processes can be completed in approximately 2 h after the blood sample is drawn.[10] Results can be assessed with the naked eye and doing so does not require experience. Therefore, with respect to the laboratory process, BCGT is easier than the two other methods containing the Coombs antibody. Because similar methods, such as gel centrifugation systems, are still being used in blood banking and transfusion medicine practices and are fully automated, BCGT may also be automated. However, despite these advantages, the test requires only 5 μL patient serum. Due to this disadvantage, pipetting should be carefully performed. In addition, using 5 μL serum may adversely affect its sensitivity. Rarely interpretation of the BCGT results cannot be clear as seen as that of Figure 1.

The manufacturer indicates that BCGT may also be used for screening purposes. Due to its speed, the test has been proposed to replace RB in the future.[16] When RB and BCGT results are compared, BCGT can be suggested as a screening test for the diagnosis of brucellosis, since there is an excellent agreement between BCGT/C-STA and BCGT/Brucellacapt results. However, it should not go unnoticed that, according to the sales price in our country, the cost of BCGT is higher than that of RB. In addition, upon dilution, Brucellacapt and BCGT are similar in price. When positive results are obtained with both tests, the costs increase as expected. Burden and cost effectiveness are issues that should be addressed in a detailed manner. However, we did not evaluate financial issues, because they were not the main focus of this study.

When the results of all titres were evaluated, we determined an excellent level of agreement between BCGT and C-STA (κ, 0.894; p < 0.001), and a substantial level of agreement between BCGT and Brucellacapt (κ, 0.802; p < 0.001) (Table 1). Excellent levels of agreement were detected between BCGT and C-STA (κ, 0.915; p < 0.001), and between BCGT and Brucellacapt (κ, 0.979; p < 0.001) when diagnostic titres were assessed (Table 2). Compared to other methods, the diagnostic performance (sensitivity, specificity, negative predictive value and positive predictive value) of BCGT was quite high, as seen in Table 2. Similar diagnostic performance and kappa values were determined in two comprehensive studies that evaluated the performance of BCGT, in which Irvem et al. and Borsa et al. found similar sensitivity and specificity values for BCGT compared to C-STA and Brucellacapt.[10,16]

In microbiology laboratories, it is advised that, in addition to the STA test, at least one agglutination test or ELISA test for Brucella IgG and IgM be used for reliable serological results.[22] After sufficient data are accumulated for BCGT, this test may also be added to test combinations.

The most important limitation of our study was that we did not evaluate the performance of BCGT on patients with the exact brucellosis diagnosis authenticated by molecular methods or by bacterial isolation. Second, we did not include a cost analysis.

We found excellent agreement between the newly developed BCGT test and C-STA, and BCGT and Brucellacapt, for the serological diagnosis of brucellosis. BCGT is easy to apply and interpret and provides reliable titre results in less than 2 h. It is also advantageous for screening. However, its diagnostic performance should be evaluated in more comprehensive studies that compare it to gold standard methods.

Conclusion

BCGT is easy to apply and interpret and provides reliable titre results in less than 2 h. It is also advantageous for screening. However, its diagnostic performance should be evaluated in more comprehensive studies that compare it to gold standard methods.

Disclosure statement

The authors declare that they have no conflict of interest.