First Report of Klebsiella pneumoniae subsp. ozaenae Meningitis in Japan

Abstract

Although the frequency of community-acquired infections caused by Klebsiella pneumoniae subsp. ozaenae (K. ozaenae) is low, they are often detected in sputum specimens. In addition, lung abscesses, necrotizing pneumonia, and urinary tract infections caused by K. ozaenae have also been reported. We herein report the first detection of K. ozaenae as an etiological agent of bacterial meningitis in Japan. Cases of K. ozaenae meningitis complicated by diabetes mellitus and sinusitis have been reported elsewhere. When Klebsiella pneumoniae is detected in such cases, it is important to use other detection methods in addition to mass spectrometry for correct identification.

Introduction

Klebsiella pneumoniae subsp. ozaenae (K. ozaenae) was first reported in the 1800s. It was thought to be specific to nasal diseases and its isolation from normal clinical samples is relatively rare. Saito et al. reported in 1973 that K. ozaenae was frequently detected in the sputum of patients with respiratory diseases and it was involved in the exacerbation of chronic bronchitis and bronchiectasis in many cases (1). Recently, cases of necrotizing pneumonia (2), lung abscess (3), skull base abscess, and urinary tract infection caused by K. ozaenae have been reported.

We encountered a case of meningitis caused by K. ozaenae in a female Japanese patient who lived in Japan. The causative organism was identified as K. pneumoniae by matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry (MS) performed in our microbiology laboratory, but it was later identified as K. ozaenae by whole-genome sequencing. In clinical practice, the use of MS for the identification of bacterial species has been on the rise because of its rapid identification capability and cost-effectiveness, especially in medium- to large-sized laboratories; however, it has limitations in differentiating microorganisms with similar MS profiles, such as Streptococcus pneumoniae and members of the Streptococcus mitis group, Shigella species, and Escherichia coli (4,5). Therefore, instead of relying only on MS, careful identification using multiple methods is required to correctly identify and understand the pathogenesis of the disease and to select an appropriate treatment depending on the circumstances.

Case Report

The patient was a Japanese woman in her 70s. She was diagnosed with diabetes but was not treated. She had been aware of a headache for 1-2 months without seeing a doctor. She was working as usual in the morning, but her headache suddenly worsened and thereafter she began to vomit. Therefore, she visited a neurosurgery clinic. Magnetic resonance imaging of the head showed left frontal sinusitis (Fig. 1) but no apparent abnormality in the cranium. During the clinical examination, the patient showed a disturbance of consciousness, such as difficulty talking and conjugate deviation to the left, so she was transferred to the Department of Neurology at our hospital for further evaluation and treatment. On arrival, her temperature was 37.7°C, blood pressure was 160/80 mmHg, and heart rate was 75 beats/min. Her orientation was normal, but her consciousness was unclear, and she showed cervical rigidity and Kernig's sign. Other physical examination results were unremarkable. Blood test results showed an elevated white blood cell count of 12.3×10³ cells/μL, blood glucose level of 254 mg/dL, and hemoglobin A1c level of 11.6% (National Glycohemoglobin Standardization Program). Cerebral spinal fluid (CSF) examination revealed a cloudy white color, a cell count of 4,093 cells/μL (polymorphonuclear cells, 86%), a protein level of 373 mg/dL, and a glucose level of 110 mg/dL. The urine test results were negative for S. pneumoniae antigen, and the nasal influenza antigen test result was negative. Bacterial meningitis was suspected, and empirical treatment with meropenem (2 g every 8 hours) and vancomycin (1.25 g every 24 h) was initiated intravenously. On the fourth day of incubation, the organism that grew from the spinal fluid sample was identified as K. pneumoniae with a sufficiently high score value (2.371) using the MALDI Biotyper with library version 5 (Bruker Daltonics, Billerica, USA). No organisms grew from the two sets of blood cultures that were collected simultaneously. The antimicrobial agent was changed to ceftriaxone (2 g every 12 h) with reference on the drug susceptibility results (Table), and the treatment was terminated 4 weeks after an improvement in the CSF findings was confirmed. No sequelae or relapse were observed. After diabetes mellitus was treated and the patient's condition stabilized, otolaryngologists performed surgery for left frontal sinusitis. At the time of surgery, no findings were suggestive of atrophic rhinitis (Fig. 2).

Figure 1.

Computed tomography in coronal sections of the paranasal sinuses. a: Before the surgery. There are no findings of bone destruction in the left frontal sinus. b: After the surgery.

Table.

Drug Susceptibility Results.

	MIC (μg/mL)	Susceptibility
Ampicillin	>16	R
Sulbactam/ampicilin	≤8	S
Cefazoline	≤4	S
Cefepime	≤2	S
Imipenem	≤1	S
Meropenem	≤1	S
Aztreonam	≤4	S
Minocycline	≤2	S
Levofloxacin	≤0.5	S
Sulfamethoxazole/trimethoprim	≤40	S
Fosfomycin	16	I

S: susceptible, I: intermediate, R: resistant

Figure 2.

The sinus mucosa during surgery. No findings suggestive of atrophic rhinitis are seen.

Since community-acquired meningitis caused by K. pneumoniae is rare in Japan, we performed whole genome sequencing of the strain isolated from the CSF using MiSeq (Illumina, San Diego, USA). A QIAseq FX DNA Library Kit (Qiagen, Venlo, the Netherlands) was used for library preparation, and the resulting sequencing reads were assembled using SPAdes (version 3.13) to obtain draft genomic sequences. The genomic sequences of strain (FUJ00291) were deposited in the National Center for Biotechnology Information database under accession No. JAHQCT000000000 (6). The draft genome data were subjected to genetic analysis using the Pathogenwatch web-based analysis tool (7). The strain was identified as capsular genotype K4 and sequence type (ST) 91. The strain carried virulence genes for aerobactin (iucAB) and yersiniabactin (ybtEPQTX) and a narrow-spectrum β-lactamase gene (bla_SHV-11). Because the capsular genotypes K4 and ST91 were previously reported to be strongly associated with K. ozaenae, the draft genome of this strain and the draft genome of ATCC 11296, a type strain of K. ozaenae (GenBank: CDJH00000000.1) (8), were compared. The average nucleotide identity is 99.76%, indicating a high degree of similarity (9). This strain was genetically identified as K. ozaenae K4-ST91.

Discussion

We encountered a case of bacterial meningitis caused by K. ozaenae associated with diabetes mellitus and sinusitis. The bacterium that grew from the spinal fluid was first identified as K. pneumoniae by MALDI-TOF MS but was later identified as K. ozaenae by whole genome sequencing. No broad-spectrum β-lactamase genes were found in this organism. The patient did not have a high fever at the time of admission, and the onset of headache was subacute, which differs from the typical course of bacterial meningitis caused by more virulent pathogens, such as S. pneumoniae. Although no antimicrobial agents were administered prior to the patient's visit to our hospital, blood samples taken at the time of admission showed no bacteremia, suggesting that the patient had developed meningitis by a direct invasion from complicated sinusitis.

The incidence of adult bacterial meningitis caused by K. pneumoniae in Japan is as low as 5.6% (10). The Infectious Diseases Society of America guidelines recommend vancomycin plus ampicillin plus a 3rd generation cephalosporin for empirical therapy of community-acquired meningitis in patients over 50 years of age, with the assumption that S. pneumoniae, Neisseria meningitidis, Listeria monocytogenes, and aerobic gram-negative bacilli are the major causative agents (11). In Japan, treatment with vancomycin plus meropenem is recommended as an alternative regimen for empirical therapy when the involvement of extended-spectrum β-lactamase-producing bacteria is suspected (12). In the present case, we suspected bacterial meningitis based on the predominance of polymorphonuclear cells in the spinal fluid examination, started empiric treatment with vancomycin and meropenem, and then switched to ceftriaxone after obtaining the identification and drug susceptibility results for the pathogen. The recommended duration of treatment for meningitis caused by gram-negative rods is 21 days, but in this case, we prolonged it because this strain demonstrated hypermucoviscosity (the string test was positive).

Klebsiella ozaenae is a subspecies of K. pneumoniae that was considered to colonize only the nasopharynx or to be a putative cause of ozaena (13,14). In 1978, Goldstein et al. investigated 64 clinical isolates of K. ozaenae, and reported that the organism was detected in sputum, urinary tract, blood, and middle ear, indicating that the spectrum of diseases caused by K. ozaenae may be more extensive than previously appreciated (15). Between the 1970s and the 1990s, three cases of meningitis due to K. ozaenae were reported in Taiwan, two in the United States, and one in Indonesia (16-20). All three cases of K. ozaenae meningitis reported in Taiwan were over 50 years old and had primary diseases of the nasopharyngeal pathway (19). Using conventional methods that are routinely used in microbiological laboratories, new species such as K. ozaenae and Klebsiella rhinoscleromatis, are largely indistinguishable from K. pneumoniae and Klebsiella oxytoca. In 2004, Hansen et al. suggested a test panel consisting of 18 biochemical tests to identify all Klebsiella spp. and subspecies (21). Although K. pneumoniae and K. ozaenae can be differentiated using the Voges-Proskauer test and malonic acid digestion, differentiation using MALDI-TOF MS is currently difficult at the present time. The clinical manifestations of different Klebsiella species overlap, and it may not be necessary to correctly identify all of them. However, K. ozaenae is frequently reported in elderly patients with diabetes and nasal diseases, suggesting that these patient backgrounds may be related to this species. Differences in the antimicrobial susceptibility profiles of K. pneumoniae and K. ozaenae have not been fully documented. In cases of serious infections such as central nervous system infections, as in this case, it is better not to give up easily when trying to correctly identify the causative organism in order to collect information on the epidemiology. We reaffirmed that it is important not to rely solely on MS and that the use of classical tests, such as biochemical characterization, is still needed to accurately identify severe infections and rare cases. Even after enrichment of the Klebsiella spp. database in Bruker Daltonics library version 6, the MS data for K. ozaenae remained unchanged with only three strains (22). Further research is needed to determine whether the enrichment of MS data will enable the differentiation of these species and strains.

Although community-acquired meningitis caused by K. pneumoniae is rare in most countries, it is frequently encountered in countries where hypervirulent K. pneumoniae is endemic (23). A single-center study in Taiwan documented that K. pneumoniae was the leading pathogen causing acute bacterial meningitis in patients without a history of neurosurgery (24). Although there is insufficient data on the epidemiology of the causative organisms of community-onset meningitis in Japan, a recent nationwide multicenter study reported that 18.6% of K. pneumoniae detected in blood cultures were hypervirulent strains (25). Considering that K. pneumoniae and K. ozaenae are indistinguishable by MALDI-TOF MS, when K. pneumoniae is detected as the causative organism of community-onset meningitis, it is important to consider the possibility of K. ozaenae being the causative organism and the involvement of hypervirulent strains.

One limitation associated with this case report is that no culture from the sinus cavities was collected at the time of admission, and K. ozaenae did not develop from the culture specimens collected during sinus surgery; therefore, we could not confirm that K. ozaenae was directly transmitted from the sinus cavities and thus caused meningitis.

In conclusion, from the perspective of treatment and epidemiology, the correct identification of K. ozaenae while considering the possibility of hypervirulent K. pneumoniae is important. MALDI-TOF MS is rapidly gaining popularity in clinical microbiology laboratories as it can provide rapid results; however, it is important to understand its limitations and to use other methods when the accuracy of identification is questionable based on clinical circumstances.

Written informed consent was obtained from the patient for the publication of this case report and any accompanying images. A copy of the written consent form is available for review by the editor of this journal.

The authors state that they have no Conflict of Interest (COI).