Diagnosis of brucellosis: Combining tests to improve performance

Introduction Brucellosis, a zoonotic infectious disease caused by bacteria of the genus Brucella, remains a significant global health concern in many parts of the world. Traditional diagnostic methods, including serological tests, suffer from limitations, including low sensibility and high false-positive rates, emphasizing the need for improved diagnostic strategies. In this study, we aimed to optimize diagnostic accuracy by reevaluating serological tests and exploring novel diagnostic algorithms. Methods A retrospective observational study was conducted using sera collected between June 2012 and June 2023 at the French National Reference Center for Brucella. Various serological tests, including Rose Bengal plate test (RBT), standard agglutination test (SAT), Brucellacapt, and ELISA for IgM and IgG, were performed. Different diagnostic algorithms were evaluated, combining RBT with SAT, Brucellacapt, and ELISA to enhance the performance of diagnostic tests. Results Among 3587 sera analyzed, 148 were confirmed cases of human brucellosis. Individual serological tests exhibited good sensitivity and specificity but lacked diagnostic accuracy. However, combining RBT with SAT or Brucellacapt significantly improved diagnostic performance, with reduced false positives. The most promising results were observed when an algorithm was built combining RBT, Brucellacapt, and ELISA for IgM and IgG (a score value of 0.5 with 90.5% for sensitivity, 99.7% for specificity, 92.4% for PPV, and 99.6% for NPV). Conclusions Serological tests remain crucial for brucellosis diagnosis, but their limitations necessitate innovative diagnostic approaches. Combining multiple serological tests in diagnostic algorithms shows promise in improving diagnostic accuracy. Efforts to refine diagnostic, strengthen surveillance, and raise awareness are essential for effective brucellosis control, particularly in resource-limited settings.


Funded studies
Enter a statement with the following details: • Initials of the authors who received each award • Grant numbers awarded to each author • The full name of each funder • URL of each funder website • Did the sponsors or funders play any role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript?Did you receive funding for this work?

Competing Interests
On behalf of all authors, disclose any competing interests that could be perceived to bias this work.This statement will be typeset if the manuscript is accepted for publication.
Review the instructions link below and PLOS NTDs' competing interests policy to determine what information must be disclosed at submission.The authors have no conflict of interest to declare.
suis (domestic pigs), B. abortus (cattle), and B. canis (dogs) [1,4].Humans can become infected by consuming contaminated food, mainly raw milk and dairy products, or by direct contact with infected animals or their bodily fluids, including through the aerosolization of contaminated samples in laboratories [4,6,7].
Brucellosis is a significant human health concern in many parts of the world, especially in Mediterranean countries, the Middle East, South and Central Asia, and Central and South America [6,8].While some countries, such as France, are officially free of the diseasdiseasefree, cases still occur in individuals returning from endemic regions.The clinical presentation of brucellosis varies widely, ranging from acute febrile illness with nonspecific symptoms to chronic debilitating conditions affecting multiple organ systems [9].This variability underscores the importance of accurate and timely diagnosis to facilitate appropriate treatment and prevent complications and relapse [10].Definitive diagnosis of Brucella infection relies on bacterial isolation, but culture sensitivity decreases as the disease progresses (i.e., during subacute and chronic phases) [8].
Traditional diagnostic methods for brucellosis include culture, serology, and molecular techniques [7,11].The identification of human cases of brucellosis heavily depends on microbiological analysis, given the variability and lack of specificity of the associated symptoms of this disease [5].While these approaches have been instrumental in identifying Brucella species and confirming infection, they often suffer from limitations such as low sensitivity, cross-reactivity with other pathogens, and the need for specialized laboratory facilities [7].Due to their simplicity, cost-effectiveness, and high negative predictive value (NPV), serological tests remain the primary diagnostic tools in endemic areas with limited resources and in countries where the prevalence is low (and considered as a neglected disease) [11].However, their results require interpretation that is often difficult and frequently inconclusive [12].[11].
Strategies to reduce false positives include using dilution-based cut-offs for positivity, combining multiple tests to improve diagnostic accuracy, and incorporating clinical algorithms that integrate serological modalities.Moreover, commercialization of the serological kit for SAT, validated for human diagnostics, has recently been discontinued, making it necessary to identify new serological diagnostic solutions.
Ultimately, improved diagnostic tools and strategies are essential for effective disease management, surveillance, and control, thereby reducing the burden of brucellosis on human and animal populations worldwide.
In this study, our objectives were to (i) reevaluate the utility of RBT dilution for diagnosing of human brucellosis and (ii) assess the performance of diagnostic algorithms combining RBT and more specific tests.

Patients and human sera
This retrospective observational study was conducted in accordance with the Declaration of Helsinki and French law [16,17].Sera sent to the French National Reference Center for Brucella

Serological tests
All samples from a given patient were processed simultaneously and kept frozen at -20°C.
For the titrated RBT, 30 L of non-diluted serum was mixed with an equal volume of a suspension in an acidic medium (pH 3.65 ±0.05) and buffered commercial Brucella abortus antigen stained with rose bengal (BioMérieux, Marcy l'Etoile, France) [15].The solution was previously equilibrated at room temperature and shaken to resuspend any bacterial sediment.
The mixture was rocked at room temperature for 4 min according to the manufacturer's recommendations.Any visible agglutination and/or a typical rim were considered positive.
Positive sera were titrated by serial twofold dilutions in physiological saline (1/2 to 1/256 dilutions), with 30 L of dilutions being tested in the same manner as pure serum.
The Wright standard tube agglutination (STA) was performed using a commercial B. abortus antigen (issued from killed B. abortus 119-3 strain) and saline as the diluent (50 μL per well) in tube format (Brucella-WRIGHT®, Bio-Rad, Marnes-la-Coquette, France) following the manufacturer's instructions.A large dilution panel (1/20 to 1/10240) of dilutions was tested to avoid negative results due to a prozone phenomenon.The SAT reactions were read after a 24h incubation at 37°C.The highest serum dilution showing >50% agglutination was considered the agglutinating titer [12].A titer of 160 is currently considered the positivity threshold in nonendemic areas [11].
The Brucellacapt® (Vircell, Granada, Spain) test employs an immunocapture agglutination technique to detect all classes of antibodies specific to Brucella species in serum samples.
Brucellacapt® tests were performed according to the manufacturer's instructions.Serum dilutions varying between 1/80 to 1/5120 were performed in strips of microtiter wells coated with anti-human immunoglobulins.Diluted patient serum and antigen were added to the wells, and the strips were incubated for 24-h at 37°C.Positive reactions show agglutination over the bottom of the well, while negative reactions are indicated by a pellet at the center of the bottom of the well.A cut-off titer of 1/160 is currently used [18][19][20].
ELISA for IgM and IgG antibodies were was performed on each serum sample using a commercial kit (Vircell).The testing procedure followed the manufacturer's instructions.The test is performed in 96-well microtiter plates that are precoated with Brucella antigen.Titers over 11 U for IgG and for IgM were considered positive.

Statistical analyses
Patients For each score, a ROC curve was generated, and optimal thresholds were selected based on optimized sensitivity and specificity.Diagnostic performances, including AUC, sensitivity, specificity, PPV, and NPV, were presented along with their confidence intervals (95%CI).
Given that predictive models were constructed, the data were partitioned into a training set (2/3) and a test set (1/3) while maintaining a balance in disease status.Performances were computed in the test set, but final models (leading to the provided formulas) were derived using the entire dataset.Performance metrics for the entire dataset were also provided and demonstrated similarity to those of the test set.

Sera analyzed
Among the 3674 sera received at the French National Reference Center for brucellosis during the study period, 13 were excluded due to missing at least two of the test results (due to insufficient serum quantity), and 74 were excluded due to an undetermined status of the serology.Among the remaining 3587 sera definitively included, 148 were considered as confirmed human brucellosis cases, and 3439 as were negative (Figure 1).

Performance of serological tests
The performance of the five serological tests used in this study is presented in Table 1.
Overall, these tests exhibited excellent specificities (94.4% to 99.3%) and good sensitivities (72.4% to 97.3%).However, the poor PPVs observed (42.6% to 84%) indicated that the different serological tests needed to be more suitable for drawing clear conclusions on the diagnosis.
Among the serological tests, the SAT is considered as the first screening test.2).
However, 22 positives were not detected using this limit, but the number of false positives was reduced from 194 to 24.

Performance of diagnostic algorithm combining RBT and SAT
Firstly, we combined the two main serological tests, RBT and SAT results, to determine the better diagnostic algorithm to diagnosest diagnostic algorithm for diagnosing brucellosis.We * RBT and SAT are the dilutions from which the test becomes positive.A negative result takes the value 2.
On the whole data set, the performances are 85 The AUC of the association of these two tests was 0.991 [0.984, 0.998].

Discussion
Brucellosis poses significant diagnostic challenges, complicating public health efforts due to the broad extent of individuals' contact with infected animals or their products.The nonspecific symptoms associated with brucellosis, such as fever and joint pain, often lead to misdiagnosis or underdiagnosis [11].Healthcare providers may overlook or misinterpret the disease, delaying appropriate treatment.Moreover, diagnostic inaccuracies are prevalent in regions with limited access to advanced diagnostic tools and healthcare infrastructure, resulting in underestimated disease burden and inadequate public health responses.Paradoxically, diagnostic difficulties can lead to the overestimation of brucellosis cases in regions reliant on less specific diagnostic methods [7,21].Therefore, there is an urgent need to improve diagnostic methods.
Serological tests, although lacking sensibility and sometimes yielding challenging results in individuals repeatedly exposed to Brucella organisms, remain a cornerstone in diagnosis, particularly in resource-limited countries [11,22].They are also important tools in countries where brucellosis has been eradicated and where the disease has been 'forgotten'.
Evaluating the serodiagnosis tests for brucellosis is complex due to the absence of a single gold standard test against which all other laboratory assays can be measured.Combining existing and cost-effective tests presents an interesting solution for early diagnosis and treatmentearly diagnosis and treatment solution.In this study, we examined the diagnostic potential of five tests for detecting exposure to Brucella: the RBT, SAT, Brucellacapt®, and ELISA tests for IgM and IgG.RBT, known for its widespread use as a screening test, often yields a weak PPV due to the often low prevalence of brucellosis and its high sensitivity, particularly as serum dilutions are not usually tested.To address this, our study aimed to optimize test performance while maintaining sensitivity and reducing false positives by establishing dilution-based cutoffs for positivity.We explored various scenarios, from individual test considerations to combinations aimed at enhancing performance.
Utilizing dilutions in RBT gave a significant improvement in diagnostic accuracy, highlighting the importance of this approach in refining testing protocols as previously reported [14,15].By diluting serum samples, nonspecific reactions can be reduced, enhancing test specificity, leading to more reliable results, and ultimately improving diagnostic accuracy.Furthermore, our investigation into combining tests underscores the potential for enhancing diagnostic capabilities in the detection of anti-Brucella antibodies.Integrating multiple testing methods allows us to use the strengths of each, potentially achieving greater overall accuracy in diagnosis.This integrated approach represents a promising avenue in clinical management, demonstrating the potential benefits of a comprehensive testing strategy.The effectiveness of combining RBT with SAT has already been established, particularly where these tests are the primary diagnostic options available [11].This underscores the importance of maximizing diagnostic capabilities with the resources at hand.However, due to the discontinuation of SAT commercialization in Europe, it was important essential to establish a new strategy to enhance diagnostic accuracy and improve patient care.Finally, combining RBT and Brucellacapt® with ELISA IgM and IgG shows a very promising performance, drastically limiting false positives and suggesting the potential for enhanced diagnostic accuracy with a multi-test approach.To our knowledge, this approach is original and has never been reported.However, further investigation is warranted to determine the reliability of this combined algorithm.Specifically, evaluating the algorithm's performance beyond just the AUC is crucial, as other metrics such as sensitivity, specificity, and PPV provide a more comprehensive assessment of diagnostic accuracy.
Despite their numerous drawbacks [23] such as low sensibility and the difficulty in interpreting results [12,21]), serological tests remain an indispensable diagnostic tool for human brucellosis.Traditionally, the principle was to use a screening test to evoke the diagnosis, followed by a confirmatory test [24].Here, we show that combining tests can improve performance and represents a promising avenue for enhancing diagnostic capabilities in brucellosis detection.By integrating multiple testing methods and combining the strengths of each, such as the sensitivity of one test and the specificity of another, we can achieve greater overall accuracy in diagnosis.This approach demonstrates the potential benefits of a comprehensive and integrated testing approach in managing brucellosis.Research efforts focused on improving the performance of serological tests and developing novel diagnostic approaches techniques are also underway.

(
Department of Microbiology, Nîmes University Hospital, Nîmes, France) between June 2012 and June 2023 were included in the analyses.The local ethics committee (Institutional Review Board of Nîmes University Hospital; registration number n°2024.03.05) approved their use in this research.Before inclusion, patients were informed of the study and their rights to oppose the use of their serology results.Patients were classified based on the expertise of the National Reference Center: i) confirmed human brucellosis if they exhibited epidemiological and clinical findings and the positivity of blood/tissue culture or had negative cultures but serological tests positive (SAT titer  1/160; RBT positive; IgG/IgM >11 UI; Brucellacapt® 1/160); ii) negative in other situations (no epidemiological and clinical findings, no cultures and negative serological tests).
with an "undetermined" diagnosis (< 2 positive serological tests and without any clear epidemiological or clinical data confirming brucellosis) and those with missing values for two of the five tests performed were excluded from the analyses.Diagnostic performances of the different tests and predictive models under evaluation were assessed using sensitivity, specificity, NPV, and PPV.The Area under the ROC curve (AUC) was also reported.All statistical analyses were conducted using R statistical software version 4.2.0.It should be noted that both quantitative and qualitative variables (the test results) are included in the model.RBT, SAT, and Brucellacapt® values represent the dilution levels at which the results become positive.This information was considered a quantitative variable in the model after transforming negative results (samples that never became positive regardless of dilution) to a value of 2. Pure serum was considered as 1, a ½ dilution as 0.5, and so on.IgM and IgG ELISA tests were qualitative variables taking only positive/negative values.The best thresholds were selected to maximize the PPV while maintaining good sensitivity (close to 85% or greater).Three algorithmic propositions were tried to enhance the diagnosis of brucellosis.The first proposition relies solely on the diluted RBT to improve specificity.The second proposition combines the diluted RBT and SAT or Brucellacapt® results.A logistic regression model was employed to integrate these variables and enhance diagnostic performance.The third proposition incorporates four serological tests (RBT, Brucellacapt® and IgM and IgG ELISA), also utilizing logistic regression.The formulas for the different models are provided in the article; the indicator function, denoted as 1{x=a}, yields a value of 1 if the condition within the braces is true (x=a here) and 0 otherwise.

Figure 1 .
Figure 1.Flowchart of the study.
However, as we have seen, it is no longer commercially available in Europe.Interestingly, in countries with low brucellosis prevalence, it is recommended to use a threshold of 1/160 to define test positivity.

Table 2 .
Performance of diluted Rose Bengal test (RBT)The performance of the diluted Rose Bengal test (RBT) was used to diagnose brucellosis.

Table 3 .
Comparison of the performance of different diagnostic algorithms using various combinations of serological test results to diagnose brucellosis.
*Evaluation was exclusively performed on 2105 sera due to the discontinuation of the kit's commercialization