Brucella Diagnostics in Endemic Areas: Evaluation of Point-of-Care Kits and the Need for Alternative Diagnostic Tests

Abstract

Brucellosis is a significant public health challenge in Kenya, particularly in pastoralist communities where the disease is endemic. Reliable and accurate point-of-care diagnostics are essential for timely case identification and effective disease management. The Febrile Brucella Agglutination Test (FBAT) is commonly used for diagnosis of brucellosis in Kenya, but concerns have been noted about its diagnostic accuracy, prompting an independent evaluation. The aim of this study was to compare the diagnostic performance of five FBAT kits with a commercial Enzyme-Linked Immunosorbent Assay (ELISA) as the reference standard, and to build local laboratory capacity for in-house kit validation for the Kajiado County laboratory staff. A total of 200 serum samples (100 ELISA-confirmed positives and 100 negatives) were tested using the FBAT kits. Each kit was evaluated for its ability to detect antibodies to both B. abortus and B. melitensis antigens. Diagnostic performance indicators, including sensitivity, specificity, and Cohen’s Kappa, were calculated, and McNemar’s test was applied to assess concordance with the ELISA results. Overall, none of the FBAT kits proved to have acceptable sensitivity and specificity compared to ELISA. We conclude that FBAT kits are not sufficient for the diagnosis of brucellosis and that alternative diagnostics should be considered.

1. Introduction

Brucellosis is a globally neglected zoonotic disease that disproportionately affects populations in low- and middle-income countries, particularly pastoralist communities where livestock farming is central to livelihoods [1,2]. In Kenya, brucellosis is among the top five priority zoonotic diseases that pose a significant public health challenge [3,4]. In this setting, the disease is caused primarily by Brucella abortus and Brucella melitensis, leading to economic losses and considerable health impacts [5,6]. Diagnosis is complicated by non-specific symptoms, which often mimic other febrile illnesses such as malaria and typhoid, making clinical identification difficult without laboratory support [7]. Ruminants, especially cattle, goats, and sheep, are the primary reservoirs of human brucellosis [8,9,10], and transmission to humans occurs through direct contact with infected animals or their secretions, particularly during birthing or handling of aborted materials, and through the consumption of unpasteurised dairy products [11,12]. Occupational exposure is also a significant risk factor, affecting farmers, veterinarians, slaughterhouse workers, and laboratory personnel [13]. In these environments, serological tests, especially Febrile Brucella Antigen Tests (FBATs), are commonly used due to their low cost and deployment simplicity [14]. However, the performance of these kits varies widely, raising concerns about their reliability [15,16]. Inaccurate results may result in misdiagnosis, inappropriate treatment, and missed opportunities for proper care [17,18]. This study assessed the diagnostic performance of five commercially available Brucella FBAT kits in comparison to ELISA testing. Furthermore, diagnostic capacity for staff of a laboratory health facility in Kajiado was strengthened through training on validation of kits. The dual approach of FBAT kit validation and workforce training is essential for improving disease surveillance and control in high-risk areas.

2. Materials and Methods

2.1. Ethical Considerations

This study was part of a broader initiative aimed at building laboratory capacity to mitigate zoonotic diseases in underserved areas of Kenya. Ethical approval for the secondary analysis of archived samples was granted by the Kenyatta National Hospital-University of Nairobi Ethics and Research Committee (KNH/ERC No. P880/11/2019) and the National Commission for Science and Technology (NACOSTI Permit No. NACOSTI/P/21/13647). The serum samples used in this study were originally collected under a separate, previously approved study conducted at the University of Nairobi’s Institute of Tropical and Infectious Diseases (UNITID). All the samples were anonymized prior to analysis, and no identifiable patient data were accessed. The retrospective analysis of these archived samples was conducted between October 2021 and January 2022.

2.2. Samples and Testing Procedures

A total of 200 serum samples that had been archived at the University of Nairobi’s Institute of Tropical and Infectious Diseases (UNITID) laboratory, Nairobi, Kenya, were used for evaluating the diagnostic performance of five commercial point-of-care (POC) FBAT diagnostic kits (Atlas Medical, Cambridge UK; Lab-Care Diagnostics, Sarigam INA, India; Expert Diagnostics, Sharja, UAE; Shubam Diagnostics, Kalyan Maharashtra, India; Fortress Diagnostics Ltd., Antrim, UK). These samples had been tested previously at UNITID for antibodies to Brucella using a validated, commercial indirect ELISA (PrioCHECK, Brucella Antibody 2.0, Thermo Fisher Scientific, Waltham, MA, USA); the sensitivity and specificity of this test system were not provided by the manufacturer, but the suggested validation criteria for positive and negative control samples were met. The ELISA served as the reference standard against which the FBAT kits were evaluated. Importantly, the ELISA did not differentiate between responses to B. abortus and B. melitensis. Of the 200 samples tested, 100 were ELISA-positive and 100 were ELISA-negative. The five commercial Brucella FBAT kits evaluated are described in

Table 1. Product details of the five commercial Brucella FBAT Kits evaluated.

Source of Kit	Product Description	Controls Provided	Diagnostic Performance Claims
Atlas Medical	Lot 19072705/Ref 8.01.15.1.0010	Yes	Sensitivity: 100%. Specificity: 98%
Accucare (Lab-Care Diagnostics)	Lot BAM 200101 Ref-Not Available	No	70% specificity and sensitivity
Expert Diagnostics	Lot 20062203 Ref FA-16.0010	Yes	Not indicated in the insert
Shubam Diagnostics	Lot BAM 200302 Ref-Not Available	No	Not indicated in the insert
Fortress Diagnostics Limited	Lot FC-1904-4 Ref FEBAMP05	Yes	Not indicated in the insert

. For each manufacturer, separate kits specific to both B. abortus and B. melitensis were evaluated.

2.3. Training and Capacity Building

This study was part of a broader initiative aimed at building laboratory capacity to mitigate zoonotic diseases in underserved areas of Kenya. The study was conducted at the UNITID laboratory by two laboratory staff from the Kajiado County referral hospital (AK and DW), who received hands-on instruction in performing in-house validation protocols for commercial diagnostic kits, including quality control, sample preparation, and interpretation of results from an experienced laboratory technician. This component was designed to empower the staff to critically evaluate and select appropriate diagnostic tools based on local needs and epidemiological context. A structured in-house validation protocol was implemented and taught to the Kajiado staff to ensure standardization and reliability. The protocol steps included the following:

• Sample selection and blinding: Use of 100 ELISA-positive and 100 ELISA-negative samples, blinded to the operator. The serum samples were allowed to thaw at room temperature before analysis, and the test kits’ reagents were removed from the refrigerator and allowed to attain room temperature. A plastic board with a white background was used because of the many samples to be tested (Figure 1).
• The routine qualitative procedure, as detailed in the kit’s inserts, was strictly adhered to and is summarized as follows:
  • Two areas of the demarcated board were labelled positive and negative. The other regions were labelled with the unique serum sample numbers, ensuring each sample had a pair representation, including the labels ‘A’ for abortus and ‘M’ for melitensis.
  • 50 μL of the paired serum samples were put in areas labelled A and B with their respective unique sample numbers.
  • A drop of the antigen was added to the serum samples and mixed using an applicator stick.
  • The whiteboard was then manually rocked and observed for agglutination over a period, as indicated in the manufacturer’s kit insert.

The results of FBAT were recorded as positive if agglutination was observed macroscopically and negative if agglutination was not present. The results of FBAT were classified as described in

Table 2. Result classification of FBAT compared to ELISA standard.

ELISA Test Result	FBAT Positive for Either B. abortus or B. melitensis	FBAT Negative for Both B. abortus and B. melitensis
Positive	True positive	False negative
Negative	False positive	True negative

.

2.4. Data Processing and Statistical Analyses

Statistical analysis was conducted using R software (version 4.2.0, R Core Team, 2023). Sensitivity and specificity for each kit relative to the ELISA results at 95% confidence intervals (CIs) were computed using the Wilson method. McNemar’s test was performed using the mcnemar.test() function in R, to assess paired categorical agreement between each diagnostic kit and the ELISA reference standard [19]. McNemar’s test is appropriate for diagnostic validation studies where the same subjects are tested with both index and reference tests, as it accounts for the paired nature of the data [20]. Cohen’s Kappa coefficient (a measure of agreement beyond chance) was calculated for each diagnostic kit to quantify agreement with the ELISA reference standard. To assess inter-rater agreement between each diagnostic kit and the ELISA reference standard, the Cohen’s Kappa coefficients were computed along with their corresponding 95% confidence intervals and p-values via the R irr package. Sensitivity and specificity with 95% confidence intervals were calculated using the R epiR package. The statistical methods selected follow established guidelines for diagnostic test validation studies recommended by the Standards for Reporting Diagnostic Accuracy Studies (STARD) guidelines [21]. The Wilson score interval method provides more accurate confidence intervals for small sample sizes and proportions and is used in contemporary diagnostic validation studies [22]. Cohen’s Kappa coefficient was selected as the primary measure of agreement beyond chance, as recommended by Landis and Koch [23] and supported by contemporary diagnostic accuracy literature [24].

3. Results

The sensitivity and specificity of five commercially available febrile antigen Brucella antigen test (FBAT) kits were evaluated in comparison to ELISA, which served as the reference standard. A summary of the diagnostic performance data: false negatives (FNs), false positives (FPs), true negatives (TNs), and true positives (TPs) is presented in

Table 3. Summary of sensitivity and specificity for validated FBAT kits against ELISA.

FBAT Kit *	FN	FP	TN	TP	Sensitivity	Specificity
Atlas	32	38	62	68	0.68	0.62
Expert	55	32	68	45	0.45	0.68
Fortress	38	77	23	62	0.62	0.23
Shubham	59	37	63	41	0.41	0.63
Labcare	49	45	55	51	0.51	0.55

* Manufacturer claims for sensitivity and specificity are described in Table 1.

. The sensitivity ranged from 0.41 (Shubham) to 0.68 (Atlas), while specificity varied widely, from 0.23 (Fortress) to 0.68 (Expert).

To further evaluate agreement between FBAT kits and the ELISA reference method, McNemar’s test for paired categorical data was performed (

Table 4. Diagnostic performance of five commercial Brucella point-of-care kits compared to ELISA.

FBAT Kit	Sensitivity (95% CI)	Specificity (95% CI)	McNemar p-Value
Atlas	0.68 (0.58–0.76)	0.62 (0.52–0.71)	0.550
Labcare	0.51 (0.41–0.61)	0.56 (0.46–0.66)	0.678
Expert	0.45 (0.35–0.55)	0.68 (0.58–0.77)	0.018 *
Shubham	0.41 (0.31–0.51)	0.63 (0.53–0.72)	0.032 *
Fortress	0.62 (0.52–0.71)	0.23 (0.15–0.33)	0.0004 *

p-values * indicate statistically significant disagreement with the ELISA reference standard (p < 0.05) based on McNemar’s test for paired categorical data [19].

). McNemar’s test revealed significant disagreement with ELISA for three kits: Expert (p = 0.018), Shubham (p = 0.032), and Fortress (p = 0.0004), indicating limited concordance with the ELISA reference standard.

The Atlas kit demonstrated the highest level of agreement with the ELISA results, with a Cohen’s Kappa coefficient of 0.30 (95% CI: 0.18–0.42; p = 0.0003), indicating statistically significant agreement beyond chance (

Table 5. Cohen’s Kappa coefficients for five commercial Brucella diagnostic kits compared to ELISA.

FBAT Kit	Cohen’s Kappa (95% CI)	p-Value
Atlas	0.30 (0.18–0.42)	0.0003 *
Labcare	0.07 (−0.05–0.19)	0.25
Expert	0.14 (0.02–0.26)	0.08
Shubham	0.04 (−0.008–0.16)	0.39
Fortress	−0.15 (−0.27–0.03)	0.91

p-value * indicates statistically significant agreement with the ELISA reference standard (p < 0.05).

). In contrast, the other test kits showed weak or non-significant agreement with ELISA. The Fortress kit yielded a negative Kappa value (−0.15), suggesting substantial disagreement with the reference standard.

4. Discussion and Conclusions

This study provides a comprehensive evaluation of five commercial Brucella point-of-care FBAT diagnostic kits. The findings highlight significant variability in diagnostic performance, reinforcing the necessity of local validation before deployment. Some of the FBAT kits yielded test results that were moderately well correlated with the ELISA test results, but overall, the sensitivity and specificity we demonstrated were not considered satisfactory for any kit, as neither approached a combined total of 1.5 [25]. This relative lack of FBAT validity has been reported by other investigators, both in Kenya [15] and in other countries [26,27,28]. Such poor test performance could result in false positive test results and consequent unnecessary treatment, an issue that may cause long-term health consequences, especially in the face of increased antimicrobial resistance [8,16]. Alternatively, false negative test results would likely result in a lack of appropriate antimicrobial treatment and chronic infection.

An important feature of this study was the inclusion of training for laboratory technicians on in-house validation techniques. This approach not only builds local capacity but also contributes to decentralized diagnostic quality assurance. Peer-reviewed literature strongly supports such strategies, emphasizing that empowering local laboratories to assess the tools they use improves both accuracy and accountability [29]. Our findings support broader efforts to implement decentralized, quality-assured diagnostic strategies in Kenya.

While a validated, commercial ELISA served as the reference standard in this study, its limitations must be acknowledged, particularly in distinguishing active vs. past infections. Additionally, the sensitivity and specificity of the ELISA we used were not specified by its manufacturer, which, to some extent, may compromise our comparisons with FBAT kits. Both ELISA and FBAT are antibody-based and may yield false positives in chronically exposed populations or false negatives in early infections [10,18]. In such settings, background antibody levels may remain elevated, reducing test specificity and limiting diagnostic and epidemiologic inference [7]. The Rose Bengal agglutination test is an alternative serologic test that has been repeatedly shown to have high sensitivity and specificity for diagnosing Brucella infection in humans and animals [30]. The Kenyan government just recently indicated that Rose Bengal testing is an acceptable diagnostic test for human brucellosis, and our view is that this procedure should replace FBAT.

The gold standard for diagnosis of brucellosis is culture [27,30], but isolation of this pathogen should be performed under BSL3 containment, takes many days, and is simply not feasible in most environments where brucellosis is endemic. Given these limitations, future diagnostic strategies should consider further development and validation of molecular assays, such as PCR or Loop-Mediated Isothermal Amplification (LAMP), which have promise for detecting Brucella DNA rapidly and with high sensitivity, making them particularly valuable in endemic, resource-limited settings [31]. Karthik and colleagues [32] demonstrated that a LAMP assay targeting B. abortus yielded 100% sensitivity and specificity when compared with the Rose Bengal and tube agglutination tests, and could be completed within an hour with minimal equipment. Subsequent studies have confirmed its utility as a reliable alternative to PCR, especially where molecular capacity is limited [32]. Ultimately, this study supports the prioritization of in-house validation of kits, coupled with investment in local diagnostic training and scalable molecular alternatives. These steps are essential to ensuring patient safety, accurate surveillance, and sustained control of brucellosis in endemic zones.

In conclusion, the findings of this study emphasize the urgent need for improved POC diagnostic tools for brucellosis in endemic regions. We recommend three key strategies:

• Discontinue use of commercial FBAT kits, as has been recommended by others [14,15,16].
• Adopt and scale up the use of validated alternatives such as the Rose Bengal test, where culture and molecular diagnostics are unavailable.
• Invest in molecular diagnostic platforms, including LAMP, alongside structured laboratory training for in-country validation.

These actions are essential to ensure accurate case detection, guide appropriate treatment, support surveillance, and ultimately reduce the burden of brucellosis in Kenya and other endemic regions.