Meningitis caused by F. tularensis is rare, but cases have been reported [7, 9, 12, 16, 24]. In our case F. tularensis was identified as the causative microbe by an in house PCR method and 16s rDNA sequencing. An assumption of subspecies holarctica as a causative agent was made based on the known geographical distribution of F. tularensis subspecies in Europe and Finland and the lack of recent traveling in patient history [18, 23]. However, since 16s rDNA is unable to distinguish between the F. tularensis subspecies [10] the possibility of other subspecies as a causative agent cannot fully be ruled out.

The initially negative screening result with meningitis/encephalitis PCR panel combined with CSF Gram staining result (Gram-negative rod), colony morphology on agar plates, local increase of positive tularemia cases led us to consider F. tularensis as a cause for the infection. In addition, the time of the year and geographic location in an area of relatively high prevalence of F. tularensis also supported this possibility. As a matter of fact, during the time patient was at the UH, 10 antibody positive samples out of 32 studied (31%) were observed in the laboratory. It is interesting that all blood cultures remained negative in our patient case. It remains a possibility that longer culture (6 or 7 days) instead our routine 5.5-day incubation might have enabled the growth of F tularensis. However, this would not have affected the clinical decision making in our patient case, but it does underline the importance of PCR-based identification of slow growing bacteria, such as Brucella in clinical laboratories. As a matter of fact, our laboratory does offer clinicians the possibility to ask for extended culture if needed.

The delay in identifying the bacteria, was partly caused by the inability of MALDI-Tof Vitek MS to identify the bacterial culture in routine clinical laboratory. Retrospectively, the in-house Francisella PCR could have been utilized on CNS upon Gram staining finding and we have now added to our SOP Francisella PCR if initial CNS PCR remains negative and Gram-negative rod is found in the CNS Gram staining. Fortunately, the patient recovered completely, and traditional antibiotic susceptibility testing played an important role in discovering the appropriate treatment option; the initially unidentified bacteria was found to be sensitive to ciprofloxacin which is in line with most of the F. tularensis meningitis cases described in the literature [7, 9, 12, 16, 24]. Several groups have also studied the in vitro antibiotic sensitivity of F. tularensis [4, 5, 17]. These comprehensive studies have shown that fluoroquinolones, aminoglycosides, and tetracyclines are all effective for F. tularensis as in our case.

F. tularensis is difficult to isolate from clinical samples [18]. It could be argued, that F tularensis should be added also to the Vitek Maldi-tof library as it may cause severe infection -and also pose a risk of laboratory acquired infections to the staff if careful good laboratory practices are not followed. In our laboratory, all positive blood cultures bottles and tests on cultured bacterial plates are performed in laboratory hoods [19]. Interestingly, Bruker Maldi-tof does identify F. tularensis, though only in species level [20]. Recently, de Vries et al. [8] made Maldi-tof library validation to identify F. tularensis.

After patient recovery, the case was analyzed in a joint meeting of hospital infectious diseases medical staff and laboratory. F tularensis is not a common causative bacterium for meningitis and as expected incidence of meningitis caused by F. tularensis is low [7, 9, 12, 16, 24], the treatment was not initially directed towards F. tularensis. The resistance of our unidentified bacteria against antibiotics towards Gram-negative rods in our laboratory left relatively little choices for antimicrobial medication, namely ciprofloxacin and tobramycin and for this ciprofloxacin was successfully administered.

Overall, we noted here a possible pitfall in diagnosing F. tularensis as a causative agent for meningitis as the commercial version of Vitek Maldi-tof does not recognize this pathogen. A good knowledge of bacteriology, surveillance of local epidemiological situation and close collaboration between clinicians and the laboratory was needed for successful bacterial identification, appropriate antibiotic treatment and above all, complete patient recovery. Our case illustrated also a need to apply more specific nucleic acid based detection methods for rarer infectious agents such as F tularensis. We are currently planning a modification to our laboratory SOP to facilitate and increase the use of F tularensis PCR in case epidemiological and clinical situation suggests F tularensis might be the causative agent behind the symptoms of a patient.