Co-infections or secondary infections with bacteria is uncommon among novel coronavirus disease 2019 (COVID-19) patients, seen in 1% of all patients according to a UK national surveillance study [1]. Nevertheless, bacterial infections are more frequent in critically ill COVID-19 patients, many of which occur as hospital-acquired infection (HAI) [2]. Pneumonia and bloodstream infections are common types of HAIs, and Klebsiella pneumoniae is one of the most common causative organisms along with Staphylococcus aureus and Pseudomonas aeruginosa [1,2,3,4]. Among the phylogroups of K. pneumoniae complex, K. pneumoniae is the most concerning, with hypervirulent strains described in Asian countries in association with severe, pyogenic infections, as exemplified by a fatal case of COVID-19 complicated by respiratory tract infection and bacteremia caused by hypervirulent K. pneumoniae sequence type (ST) 86 [5]. While Klebsiella variicola has been recently recognized as an important human pathogen, the clinical characteristics and impact of co-infection or secondary infection among COVID-19 patients remain unclear [6].

Here, we present a case of fatal secondary infection caused by K. variicola that complicated the clinical course of severe COVID-19.

Case presentations

In May 2021, a 71-year-old man presented with fever, altered mental status and generalized weakness. He had had fever over 38.0 °C for 4 days prior to admission and had a positive SARS-CoV-2 PCR test at a primary care clinic. He had no known past medical history but was newly diagnosed with type II diabetes mellitus (hemoglobin A1c of 10.3%) upon admission. His vital signs included body temperature of 40.0 °C, blood pressure of 155/86 mmHg, pulse rate of 100 beats/min, respiratory rate of 35 breaths/min, and oxygen saturation of 85% with ambient air. Computed tomography scan of the chest showed bilateral ground-glass opacities and consolidations, consistent with COVID-19 pneumonia. Initial laboratory data showed white blood cells count of 4000/μL, C-reactive protein of 14.14 mg/dL (reference, 0.00–0.14).

He was admitted to intensive care unit and received dexamethasone (6.6 mg/day), remdesivir, and intravenous unfractionated heparin. He subsequently required high flow nasal cannula. No empiric antibiotics were given as no bacterial infection was suspected on admission. Initial bacterial cultures of blood, sputum and urine were all negative.

On hospital day 3, his respiratory status deteriorated and he was intubated (Figs. 1, 2). On hospital day 6, blood and sputum cultures were repeated due to fever but the blood cultures revealed no growth. His fever persisted and sputum cultures were repeated on hospital day 8. Gram stain of the sputum samples collected on hospital days 6 and 8 revealed Gram-positive cocci (GPC) in clusters. GPC was becoming more predominant and seen within phagocytes in the sample on hospital day 8. The sputum cultures on hospital days 6 and 8 subsequently grew methicillin-susceptible Staphylococcus aureus (MSSA) and Klebsiella pneumoniae complex. Superimposed MSSA pneumonia was suspected and cefazolin was started on hospital day 10.

Fig. 1
figure 1

Clinical parameters of the COVID-19 case with K. variicola. SOFA sequential organ failure assessment, Max BT maximum body temperature in a day, P/F ratio (partial pressure of arterial oxygen, PaO2/fraction of inspired oxygen, FiO2, %) ratio, CT computed tomography, CEZ Cefazolin, PIPC/TAZ Piperacillin/Tazobactam, VCM Vancomycin, MEPM Meropenem

Fig. 2
figure 2

The serial chest radiographs of the patient

On hospital day 11, cefazolin was switched to piperacillin–tazobactam. On hospital day 12, the blood culture taken on hospital day 11 grew Gram-negative rods. All the intravenous catheters were replaced. On hospital day 13, he rapidly developed hypotension, worsening oxygenation and lactic acidosis (Fig. 1). The serial chest radiographs remained unchanged between hospital days 10 and 13 (Fig. 2). Despite prompt interventions including escalation of antibiotics to vancomycin and meropenem, he died on the same day.

Subsequently, blood and sputum cultures collected on hospital days 11, 12 and 13 all grew K. pneumoniae complex identified using MicroScan (Beckman Coulter, CA, USA). All isolates were susceptible to cefazolin, piperacillin–tazobactam, meropenem, sulfamethoxazole-trimethoprim, levofloxacin and fosfomycin except the isolates from sputum cultures collected on hospital days 8 and 13 were intermediate to fosfomycin.

The isolates from the sputum cultures collected on hospital days 6, 12 and 13, the blood cultures collected on hospital days 12 and 13, and the culture of the central venous catheter removed on hospital day 12 were subjected to molecular characterization. All isolates were identified as K. variicola by a multiplex PCR protocol, which distinguishes the phylogroups of the K. pneumoniae complex based on polymorphisms of the SHV-type β-lactamase gene [7]. Another multiplex PCR, which examines the virulence gene profile, demonstrated that all strains carried kfu, entB, and mrkB. These results suggested that the isolates recovered from different body sites on different dates were clonal K. variicola isolates. A representative isolate (FUJ01370), which was an isolate from the blood culture on hospital day 12 was further analyzed with whole-genome sequencing by using NextSeq 2000 (Illumina) as described previously [6], and genetic features were characterized using the Pathogenwatch website ( The capsular genotype and O locus type were determined as KL11 and O3, respectively. The isolate carried a gene for LEN-type β-lactamase (blaLEN-16), which is a characteristic genetic feature of K. variicola [8], and oqxA, oqxB, fosA as antimicrobial resistance genes. Replicons for IncFIB (pNDM-Mar) plasmid, FIA (pBK30683) plasmid, and IncHI1B (pNDM-MAR) plasmid were also identified. The isolate had a novel multilocus sequence typing allelic profile (gapA-infB-mdh-pgi-phoE-rpoB-tonB: 16-24-21-27-52-17-152), to which sequence type 5794 was assigned (GenBank assembly accession: GCA_019042755.1).


We report a case of fatal respiratory and bloodstream infection of K. variicola complicating severe COVID-19. The causative organism was initially reported as K. pneumoniae, but genetic analysis revealed that it was indeed K. variicola. K. variicola have been reported among patients with immunocompromising conditions such as malignancy and diabetes [6], and a recent study in Czech Republic reported that the percentage of K. variicola as an etiological agent causing nosocomial infections increased significantly (from 1 to 6%, p = 0.0004) in COVID-19 patients compared to the pre-COVID-19 period [9]. But the impact of co-infection or secondary infection among COVID-19 patients remain uncertain.

Klebsiella variicola is a member of the K. pneumoniae complex, originally reported as a bacterium in plants, but there have been an increasing number of reports on human infections in recent years. Conventional biochemical methods and automated instruments are unable to distinguish K. variicola from K. pneumoniae, and genetic analysis is required [8]. A recent study in Japan conducted a whole genome sequencing analysis of 140 strains collected as causative organisms of bloodstream infections and initially identified as K. pneumoniae reported that 24 (17.1%) were genetically identified as K. variicola [6]. As with hypervirulent K. pneumoniae, the epidemiology may vary greatly by geographic region, and secondary infections in COVID-19 patients may also be under-reported.

The pathogenic potential of K. variicola remains uncertain. In a single-center study in Sweden, K. variicola was a significant risk factor for 30-day mortality [10]. However, whether this resulted from virulent clone(s) representing the majority of K. variicola cases at the institution is unknown. Some strains of K. variicola show hypermucoviscosity of colonies on agar plates or carriage of cardinal virulence genes (rmpA, rmpA2, and genes for salmochelin and aerobactin), which are biomarkers of hypervirulent K. pneumoniae. However, the relationship between these features and virulence in human is unclear [6, 11].

Although K. variicola infection was fatal in our patient, the strains detected were neither highly viscous nor associated with cardinal virulence genes. It is reported that some strains have no known microbiological and genetic characteristics of virulence but are hypervirulent clinically [12, 13], raising the possibility that key virulence determinants are yet to be identified in K. variicola. Interestingly, mapping of the contigs of the sequenced strain to the virulence plasmids of K. variicola isolates (pKV8917, p15WZ-82_Vir), which were reported to be hypervirulent, revealed a common genetic region spanning approximately 150 kb (data not shown). However, this region does not contain any genes that have been associated with hypervirulence of K. pneumoniae, and the significance of this plasmid region is unknown [12, 14]].


In conclusion, we report a fatal case of K. variicola infection complicating severe COVID-19. Co-infection or secondary infection of K. variicola with COVID-19 is likely under-recognized and can be fulminant as in this case. Further research is required to better describe the epidemiological, clinical, biological and genetic characteristics of K. variicola.