Post-surgical meningitis caused by Klebsiella variicola.

Klebsiella variicola, a member of K. pneumoniae phylogroup, can cause severe infectious diseases. We report a case of K. variicola meningitis after neurosurgery. The bacterium was isolated from blood and cerebrospinal fluid, and bacterial species identification was carried out by using both matrix-assisted laser-desorption/ionization time-of-flight mass spectroscopy (MALDI-TOF MS) and whole genome sequencing. Initially, the organism was misidentified as K. pneumoniae by VITEK®2; automated system in the clinical laboratory examination. The patient recovered with the combination of surgical drainage and antimicrobial treatment. To our knowledge, this is the first case report of post-surgical meningitis caused by K. variicola. As experienced in this case, the automated bacterial identification system popularly being used in the clinical laboratory might not be effective enough for bacterial species identification. The use of MALDI-TOF MS for microbial identification may be helpful to physicians for appropriate management of K. pneumoniae phylogroup infection.

Introduction

Klebsiella variicola is genetically closely related to K. pneumoniae and K. quasipneumoniae; it is difficult to distinguish the three species using routine biochemical tests [1]. K. variicola is an environmental bacterium that colonizes soil and plants [2]. Recently, K. variicola has also been reported to cause human infectious diseases; these infections have a higher mortality rate than those caused by K. pneumoniae [3]. Because distinguishing K. variicola from K. pneumoniae is difficult in general clinical laboratories, the clinical importance of K. variicola may be underestimated [4]. Therefore, case reports describing the precise identification of K. variicola by genome analysis are needed. We report a case of post-neurosurgical meningitis caused by K. variicola, in which the causative agent was identified by both mass spectrometry and genome sequencing. To our knowledge, this is the first case report of meningitis caused by K. variicola. Furthermore, we reviewed the diagnostic methods used for identifying K. variicola infections.

Case report

A 31-year-old woman was transferred to our hospital with sudden onset of severe headache and loss of consciousness. Her medical history was significant for systemic lupus erythematosus on prednisolone 10 mg/day (0.25 mg/kg/day). Based on head computed tomography, she was diagnosed with a subarachnoid hemorrhage. An emergent craniotomy was performed for clipping. Epidural, ventricular, and cisternal drains were inserted during surgery. Thirty-three days after the first surgery, she underwent cranioplasty with; an artificial dura mater inserted, and her own cranial bone, which had been removed for brain decompression during the first surgery, was replaced and fixed with titanium plates. Cefazolin, administered as perioperative prophylaxis, was discontinued on postoperative day one.

Twelve days after the second surgery, she became febrile, and she experienced seizures. Blood and cerebrospinal fluid (CSF) samples were obtained for analysis and culture. On examination, the patient was febrile at 39.3 °C with a pulse rate of 157 beats/min and a Glasgow coma scale of E4VTM2. Examinations of her chest, abdomen, legs, and skin were unremarkable. The following results were observed on CSF analysis: protein level, 814 mg/dL; glucose, undetectable; and total nucleated cell count, 16,000/mm3 (99.6% polynucleated cells). Based on Gram staining of the CSF sample (Fig. 1), we diagnosed bacterial meningitis. Thus, we administered intravenous vancomycin 1 g every 12 h and intravenous meropenem 2 g every eight hours as empirical therapy according to the current Infectious Diseases Society of America guidelines [5]. On the following day, she underwent another surgery for open cranial drainage. The artificial dura mater and the replaced cranial bone were removed.

Fig. 1

Gram staining of the cerebrospinal fluid sample showing short gram-negative rods captured by a neutrophil.

Both blood and CSF cultures were positive. The microorganism isolated from these samples was identified as K. variicola using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) (MALDI biotyper, Bruker Inc., USA). We also determined the whole genome sequence of the strain detected from the blood sample (JMUB4806), and the sequence was compared with 37 genome sequences, including those of 13 known K. variicola strains (Fig. 2). Phylogenetic analysis indicated that JMUB4806 was classified as K. variicola. Contrastingly, an automated instrument for infectious disease testing (VITEK®2, bioMérieux S.A., France) identified the causative organism as K. pneumoniae.

Fig. 2

Phylogenetic relationship of K. variicola JMUB4039 with other Klebsiella species.

The maximum parsimony tree was constructed using the majority of single nucleotide polymorphisms present in at least 75% of genomes analyzed by kSNP3.0 program (ref. S.N. Gardner et al., 2015 Bioinformatics Vol. 31 2877–2878). The 39 genome sequences analyzed were obtained from seven Klebsiella species.

Based on the antimicrobial susceptibility profile of the detected K. variicola strain, therapy was adjusted to intravenous ceftriaxone 2 g every 12 h. Head magnetic resonance imaging (MRI) three days after the third surgery showed severe inflammation in the bilateral frontal lobes and fluid collection in the third ventricle (Fig. 3). The therapy was extended more than four weeks after the third surgery.

Fig. 3

Head magnetic resonance imaging three days after the third surgery showing severe inflammation in the bilateral frontal lobes and fluid collection in the third ventricle (white arrow).

a. T1, T1-weighted image; b. T2, T2-weighted image; c. FLAIR, fluid-attenuated inversion recovery; d. DWI, diffusion-weighted image.

Because of the development of drug-induced rash, antimicrobial therapy was transitioned from ceftriaxone to intravenous levofloxacin 500 mg every 24 h and then levofloxacin to intravenous aztreonam 2 g every 8 h. Thirty days after the third surgery, head MRI (Fig. 4) was performed; although hydrocephalus had worsened, the inflammation in the bilateral frontal lobes had diminished, and fluid collection in the third ventricle had disappeared.

Fig. 4

Head magnetic resonance imaging 30 days after the third surgery showing diminished inflammation in the bilateral frontal lobes and disappearance of fluid collection in the third ventricle.

a. T1, T1-weighted image; b. T2, T2-weighted image; c. FLAIR, fluid-attenuated inversion recovery; d. DWI, diffusion-weighted image.

Microbiological investigations

The genomic DNA of JMUB4039 was extracted using the DNeasy Blood and Tissue Kit (Qiagen, Hilden, Germany). The whole genome was sequenced using the Nextera XT DNA Library Preparation Kit and Miseq Reagent Kit v3 (Illumina Inc, USA). The raw reads were assembled by the CLC genomics work bench (CLC bio). The phylogenetic tree was created using kSNP3.0 and FigTree v1.4.3 using JMUB4039 and 38 Klebsiella strains whose genome sequences are available on GenBank.

Discussion

We encountered a case of post-surgical meningitis caused by K. variicola and treated it successfully. To our knowledge, this is the first case report of meningitis caused by K. variicola. In 2004, K. variicola was distinguished from K. pneumoniae based on the following factors: results of total DNA-DNA hybridization; monophyly observed in phylogenetic analysis that was derived from the sequences of rpoB, gyrA, mdh, infB, phoE, and nifH genes; and distinct phenotypic traits [1]. The draft genome sequence of K. variicola was proposed in 2014 [6].

Maatallah et al reported that bacteremic infections caused by K. variicola had higher 30-day mortality rates (29.4%) than those of infections caused by K. pneumoniae (13.5%) [3]. These results suggested that K. variicola has higher virulence than K. pneumoniae. Thus, identification of K. variicola in clinical settings is important for providing appropriate management and improving the outcomes of K. variicola infections. However, automated instruments used for infectious disease testing might identify K. variicola as K. pneumoniae. In fact, in our case, organisms detected from blood and CSF cultures were identified as K. pneumoniae by VITEK®2, but were identified as K. variicola by MALDI-TOF MS and whole genome sequencing. Adonitol fermentation may be useful for distinguishing K. variicola (adonitol-negative) from K. pneumoniae and K. quasipneumoniae (adonitol-positive) [7]. However, the negative predictive value of adonitol fermentation for detecting K. variicola is reported to be only 69.8% [7]. Moreover, adonitol fermentation is not routinely performed in clinical laboratories.

In a recently published study, both the sensitivity and specificity of MALDI-TOF MS for detecting K. variicola were determined to be 100% [8]. Thus, MALDI-TOF MS may be a suitable method for detecting K. variicola. Furthermore, the spread of MALDI-TOF MS among clinical laboratories is expected to facilitate identification of K. variicola. In conclusion, to distinguish K. variicola from K. pneumoniae, clinicians should reanalyze organisms that were identified by automated instruments as K. pneumoniae. This could enable physicians to provide appropriate management and improve the outcomes of patients with K. variicola infections.

Funding

No financial support was provided for work relevant to this article.

Author’s contribution

All authors meet the ICMJE authorship criteria.

Dr. Dai Akine, Dr. Teppei Sasahara and Dr. Yuji Morisawa contributed to acquisition, analysis and interpretation of both clinical and microbiological data of the case. Dr. Yohei Ishishita and Dr. Takashi Yamaguchi contributed to analysis and interpretation of clinical data. Dr. Shinya Watanabe and Dr. Longzhu Cui contributed to analysis and interpretation of microbiological data.

All authors revised the work and approved of the final version of the work.

Declaration of Competing Interest

None.