Ribosomal RNA Sequence Analysis of Brucella Infection Misidentified as Ochrobactrum anthropi Infection

Abstract

A Brucella isolate was identified from purulent material collected during a hip surgery. Two previous blood cultures from the same patient yielded Ochrobactrum anthropi. After rRNA sequencing, all the isolates were identified as Brucella species and subsequently serotyped as Brucella suis. Misidentification of Brucella species remains a problem with bacterial identification systems.

CASE REPORT

A 62-year-old female presented to the emergency department with left hip pain and was unable to bear weight on that hip. The day before presentation she had fallen while standing up from a sitting position. On presentation the patient was febrile (38.7°C) with a low white blood cell count of 4,300 cells/μl. One set of blood cultures was obtained, but no antibiotics were given.

The patient was admitted to the hospital for surgery to stabilize the impacted fracture with a pin. She continued to have a high temperature, and a second set of blood cultures was collected. The first blood culture became positive after 3 days of incubation, with a short, Gram-negative bacillus in the aerobic bottle only. This bacterium was identified as Ochrobactrum anthropi by the RapID NF Plus system (Remel, Lenexa, KS). The second blood culture remained negative after 7 days of incubation. Antibiotics were administered to the patient, and she was released from the hospital after a successful surgery that stabilized the femoral neck fracture. The patient was given 2 weeks of antibiotics as an outpatient.

One month later, the patient was readmitted for severe abdominal pain. Despite the absence of subjective fever, chills, or sweats, her temperature was 38.6°C. Her white blood cell count was 3,100 cells/ml with 66% neutrophils. Two sets of blood cultures were obtained, and antibiotics were started. The fever resolved after 2 days. Four days after blood cultures were collected, a short, Gram-negative bacillus was detected in a single aerobic bottle and was also identified as Ochrobactrum anthropi by the RapID NF Plus system. Three other sets of blood cultures remained negative.

Seven months later, the patient was again admitted to the hospital for a permanent repair of the femoral neck fracture. During surgery, purulent material was collected from inflammatory tissue around the screws and within the hip socket. Fluid analysis revealed chronic inflammation with 44,000 white blood cells/μl. The Gram stain showed few neutrophils with no organisms. The bacterial culture was positive, with growth of small colonies on the sheep blood agar plate after 3 days of incubation. The Gram stain of the colonies was initially reported as corynebacterium-like bacteria with Gram-variable properties.

The organism was not identified by the Vitek2 Gram-positive identification (GPI) card, and after review it was evaluated with the Vitek2 Gram-negative identification (GNI) card. The Vitek2 GNI card identified the isolate as Brucella melitensis with a 97% probability. After this identification, the two earlier blood culture isolates were reevaluated. The isolates that were identified as Ochrobactrum anthropi as well as the Brucella species were typed using rRNA sequencing. The rRNA results indicated that all the bacterial isolates were Brucella species, with 100% agreement to rRNA sequences of known brucellae and no homology to Ochrobactrum anthropi sequences. The isolate was subsequently serotyped by the CDC as Brucella suis.

The patient was started on appropriate antibiotics and was notified of this change in the identification of the bacterium. She denied any exposure to Brucella species, such as ingestion of unpasteurized milk or milk products, travel outside the United States, or exposure to domestic farm animals. However, the patient was an experienced medical technologist who had worked in a clinical microbiology laboratory at a large hospital for more than 20 years.

Brucella species are small, Gram-negative bacilli that grow in humans as intracellular pathogens. The bacterium was first isolated by Sir David Bruce on the island of Malta in 1893 (15). Since that time, Brucella species have been recognized worldwide as being pathogenic in a wide variety of animals, including humans.

There are several reasons why Brucella species are often misidentified in clinical laboratories (1, 4, 30, 31, 33). First, Brucella infections are rare, which leads to a lack of experience with this organism in clinical laboratories. Second, different bacterial identification systems are unable to correctly identify Brucella species, which misleads technical staff (3, 4, 11, 30, 31, 41, 42). Also, Brucella infections are often chronic and present with vague clinical signs. Thus, cultures may be sent to different laboratories at different times.

Brucella infections typically present with generalized symptoms and intermittent bacteremia (29, 42). Exposure to infected domesticated animals or ingestion of contaminated food, such as cheese or milk, have been associated with Brucella infections (7, 13). However, Brucella infections have often been acquired by individuals working in laboratories where the organism is cultured. These exposures occur in a variety of different settings, such as clinical diagnostic laboratories and laboratories preparing vaccine strains or conducting experimental animal work (2, 13, 37).

The average incubation period for brucellosis is 2 to 10 weeks, but it can be longer than 6 months (5). Symptoms include intermittent fever, chills, malaise, sweating, joint pain, headache, anorexia, and fatigue (3, 13, 18). Chronic untreated brucellosis can lead to abscesses in multiple organs, such as the liver, spleen, heart, brain, or bones. The sensitivity of culturing Brucella from blood cultures of infected patients ranges from 15% to 70% (34, 40, 41). Thus, a negative blood culture does not rule out Brucella. The current case was consistent with this rate; only 2 of 6 (33%) aerobic bottles from blood culture sets were positive. Only after the isolation of brucellae from a surgical site was the bacterium correctly identified.

The diagnosis of Brucella infections is based on the isolation of the bacteria by culturing, detection of Brucella-specific serology, or detection with molecular methods. Brucellae grow on standard laboratory media, such as sheep blood agar and chocolate agar, and in most blood culture systems. The bacteria are strictly aerobic, nonhemolytic, oxidase positive, and produce a large amount of urease. In this case, brucellae were detected in blood cultures after 3 to 4 days of incubation. The identification of the bacterium as Ochrobactrum anthropi, which is typically a soil pathogen, and the lack of any additional positive cultures made these organisms appear to be possible contaminants.

The bacterial colonies appear as small white or gray colonies and grow slowly on sheep blood or chocolate agar after 24 h (Fig. 1). The slow growth of this bacterium may require an incubation of at least 5 days. The Gram stain of the organism shows variously sized coccobacilli that often resist destaining and appear to be Gram variable (Fig. 2).

Fig. 1.

Growth of Brucella species on sheep blood agar after 1 day of incubation in an aerobic environment.

Fig. 2.

Gram stain of colonies from the blood agar plate. Note the variable sizes of the bacteria and inconsistent staining properties of the cells.

The misidentification of B. melitensis as O. anthropi by commercial bacterial systems has been previously reported (10). Brucellae have also been mistakenly identified as Oligella ureolytica or Psychrobacter phenylpyruvicus (formerly named Moraxella phenylpyruvica) by commercial bacterial identifications systems (4, 31, 38). Because this appears to be a recurring issue in diagnostic laboratories, a systematic approach should be used to alert the technical staff. After this event, the laboratory informatics system was used to alert laboratory personnel that an isolate may be a Brucella species. When an identification of Ochrobactrum anthropi, Oligella ureolytica, or Psychrobacter phenylpyruvicus is sent to the laboratory information system from any bacterial identification system, a note populates the computer and states “rule out Brucella species, work within a biosafety cabinet.” Since these organisms are rarely seen, it is anticipated that this strategy will help identify and contain Brucella isolates quickly.

In this case, the three positive cultures were isolated months apart. When the surgical specimen yielded a Brucella species, the identification of the two previous blood culture isolates was reevaluated. Confirmation of the identification was needed for the appropriate treatment of the patient and for the evaluation of exposed laboratory personnel. The isolates were referred for rRNA sequencing after heat inactivation to prevent exposure to laboratory workers.

The isolates were analyzed by bidirectional sequencing of the entire 16S rRNA gene. The sequences were compared to related sequences in the SmartGene 16S eubacterial database (SmartGene, Inc., Raleigh, NC). This database is compiled and constantly updated from sequence data in the public domain. The integrity of the data is ensured by using specific profiles to filter out unreliable sequences. The extraction algorithm addresses the variability of the 16S rRNA gene across the bacterial kingdom. The database is frequently updated and takes into account recently described organisms. All sequences included in the database are linked to their original source without alteration. The genus-level identification of this isolate assigned a 100.0% match to 25 reference Brucella sequences. There are currently 170 Brucella rRNA sequences available in the SmartGene system. Assignment to a particular species of Brucella was not possible due to the high rate of homology among Brucella species sequences within the 16S rRNA gene. There was little homology (<50%) to the 133 Ochrobactrum species 16S rRNA sequences.

The family Brucellaceae contains the genera Brucella, Mycoplana, and Ochrobactrum. The most pathogenic bacteria in this family are the Brucella species. At the molecular level, all species of the genus Brucella are closely related. The second edition of Bergey's Manual of Systematic Bacteriology reports a high degree of DNA relatedness between all members of the Brucella genus, consistent with a single species (15). All Brucella strains share a highly similar genome structure with differences occurring in the arrangement of the genome. There are five Brucella species that are pathogenic with structural differences that correlate to host specificity; B. melitensis (found in sheep/goats), B. suis (found in swine), Brucella abortus (found in cattle), Brucella canis (found in dogs), and Brucella ovis (found in sheep) (15). All have been identified as pathogenic in humans. The distinction between these species is based on host specificity and antisera reactivity. They cannot be distinguished by genetic analysis of the 16S RNA gene. In spite of this finding, there are a large number of studies that have used molecular techniques to identify brucellae to the species level (6, 9, 14, 16, 19, 20, 22–26, 28, 32, 35). These studies have used different bacterial gene targets, indicating that the molecular differentiation of Brucella species requires the evaluation of several genes, especially the multilocus-variable tandem repeats (6, 14, 20, 23).

Approximately one-fourth of all laboratory-acquired bacterial infections are caused by Brucella species (12, 21, 27). The infectious dose of Brucella is 10 to 100 organisms by aerosol route (17, 37). Several cases have occurred after individuals were working with the organism on bench tops without wearing protective hoods, masks, or shields (17, 36, 37). However, cases have been reported that cannot accurately determine the exact mechanism of transmission. Due to the infectivity of Brucella species, biosafety level 3 practices are recommended for all work with cultures. However, the laboratory must first suspect a Brucella species in order to comply with these measures. This is difficult since species of Brucella phenotypically look similar to several other small Gram-negative bacteria. In addition, Brucella species are not often encountered in the United States. The CDC reported an average of 114 Brucella infection cases per year in the United States between 2005 and 2009 (www.cdc.gov/mmwr/preview/mmwrhtml/mm5916md.htm). Since these organisms are rarely isolated and resemble many other bacteria on agar plates, they are easily missed, even by experienced technical staff.

Any exposure or potential exposure to species of Brucella requires that the individual be monitored for Brucella-specific antibody. Exposed workers should have a baseline serum collected immediately after known exposure, and additional Brucella-specific antibody testing should be performed at 2, 4, 6, and 24 weeks after exposure. If symptoms or seroconversion develops, the individual should receive prophylactic antibiotic therapy for at least 6 weeks. All individuals with potential exposure must be monitored for symptoms for 4 weeks and actively monitored for increased body temperature (8, 39).

However, since Brucella species are often misidentified in the laboratory, exposures may not be recognized. Because laboratory misidentification is possible, some individuals are exposed before they are aware of the need for precautions or monitoring. If isolates are misidentified, individuals may not be aware that the exposure ever occurred.

This case demonstrates the difficulties of initially identifying Brucella species from specimens collected over the course of 9 months. Brucellosis is not often immediately life-threatening, but over time it can be severely debilitating. Since the symptoms of brucellosis are nonspecific and it is a rare disease in the United States, health care providers may not consider brucellosis in a differential diagnosis. However, brucellosis remains an occupational hazard for any laboratory personnel working in a clinical microbiology laboratory (27). From this case, it is clear that clinical microbiology staff may be exposed to brucellae without a follow-up if an isolate is not correctly identified. In this case, the patient's only obvious risk was her experience as a medical technologist working in a microbiology laboratory for more than 20 years. There was no specific event involving work with a Brucella isolate. However, it remains possible that this infection originated from working with an unknown Brucella culture. A practical approach to assessing exposure to brucellae in clinical laboratory personnel would be to perform regular Brucella-specific antibody surveillance. This practice could prevent long-term consequences if even a few laboratory personnel were identified and treated appropriately before chronic long-term damage occurred due to an occult Brucella infection.