Limitations of the Standard Agglutination Test for Detecting Patients with Brucella melitensis Bacteremia

Introduction

B rucellosis remains a highly endemic zoonosis among domesticated animals in many regions, including Middle East countries (Pappas et al. 2007). Brucella melitensis infections are common in southern Israel among herds of sheep and goats and cause substantial human morbidity in the seminomadic Bedouin population of the area that consumes unpasteurized dairy products (Banai 2010, Slater et al. 1990–1991).

The entire population of southern Israel is covered by a comprehensive national health insurance system and served by a vast net of community health clinics and a single hospital (Soroka University Medical Center, SUMC). Diagnostic blood cultures are only provided at the SUMC for inpatients and for individuals referred to the Emergency Department, and no blood cultures are performed in the outpatient setting. Serologic tests for brucellosis are performed at the Clinical Microbiology Laboratory of the SUMC (CML-SUMC) as well as in a few smaller laboratories in the region. It is estimated that the CML-SUMC provides diagnostic serologic testing for brucellosis to 90% of the relevant Bedouin population.

Materials and Methods

Results and Discussion

During the 8-year study period 2002–2009, a total of 850 culture and/or serologically confirmed brucellosis cases were identified at the CML-SUMC, of which 264 (31.1%) met the criteria for inclusion in the data analysis. The remaining 586 patients comprised individuals with culture-confirmed and/or serologically proven disease but from whom either cultures or serologic tests were not obtained, were drawn >1 week apart, or culture results were negative and, therefore, did not enable to assess the SAT sensitivity.

All bacterial isolates were identified as B. melitensis. The distribution of cases by SAT titer as well as the number of cases missed and the sensitivity rate at each cutoff are shown in Table 1. The cutoff of ≥1:160 missed 8.3% of culture-positive patients, whereas a cutoff of ≥1:320 would have missed 17.4% of patients.

Human brucellosis is not usually a life-threatening disease, and therefore, affected individuals are commonly diagnosed and managed as outpatients. Because in southern Israel neither blood cultures nor nucleic acid amplification are available in the outpatient setting, the sensitivity of serologic assays is crucial for a reliable diagnosis of the disease. The present study shows that the performance of the SAT assay in the Bedouin population of the area is suboptimal and many patients with culture-proven brucellosis could have been missed by serologic tests. Needless to say, increasing the diagnostic SAT cutoff titer to 1:320, as recommended to decrease false-positive rates in endemic areas (Pappas et al. 2005), would have overlooked an even greater fraction of cases.

Obviously, the sensitivity of a diagnostic test strongly depends on the comparator chosen as a gold standard. Evaluation of serologic tests for brucellosis has been complicated by the fact that combinations of epidemiologic background, symptoms compatible with the disease, isolation of the organism, and seroconversion have been used as a gold standard for assessing sensitivity (Mantecon et al. 2006). Because the clinical presentation of human brucellosis is unspecific, antibodies against smooth lipopolysaccharide antigens show cross reactivity with other Gram-negative organisms, and high background titers may occur in endemic areas in the absence of active disease, only isolation of Brucella organisms can be considered an irrefutable proof of the disease (Yagupsky 1999). In the present study, positive isolation of brucellae from blood was chosen as the gold standard to avoid the uncertainties of a less-rigorous disease-defining criterion. Obviously, not all patients with brucellosis have a positive Brucella culture by the time of presentation and the chances of detecting the organism substantially diminish in individuals with prolonged illness or receiving antibiotic therapy (Yagupsky 1999).

In recent years, nucleic acid amplification assays have emerged as a sensitive and specific alternative approach for confirming the disease (Vrioni et al. 2008). Although promising, these tests have not been standardized, and the significance of a persistent positive PCR result in appropriately treated and apparently cured patients is unclear (Vrioni et al. 2008). Equally important, PCR requires expensive technology, adequate laboratory facilities, and technical expertise, which are frequently unavailable in developing countries where brucellosis is endemic. Thus, serologic tests continue to play an important role in establishing the diagnosis. Evaluation of serologic assays for brucellosis in general and that of the SAT, which remains the most popular method, has identified serious drawbacks such as inability to detect Brucella canis infections, poor specificity, presence of blocking antibodies, or the prozone effect (Pappas et al. 2005; Araj 2010).

An important drawback of the present study is the fact that relevant clinical information, such as illness duration, was not available, making it difficult to interpret false-negative SAT results. However, the results indicate that the recommended 1:320 cutoff may overlook nearly 1 in 5 patients in the bacteremic phase of brucellosis, and therefore, the opportunity to establish the diagnosis and administer early antimicrobial therapy could be missed. This may lead to chronic infection characterized by focal complications, persisting health complaints, and permanent sequelae (Pappas et al. 2005). Physicians working in regions endemic for brucellosis should be aware of the limitations of SAT assay for detecting patients with B. melitensis bacteremia.