2025 年 12 巻 p. 435-438
The clinical course of pyogenic spondylodiscitis varies based on its etiology. To the best of our knowledge, there have been only 7 cases of spondylodiscitis caused by Klebsiella oxytoca. This bacterium typically affects patients with underlying comorbidities or develops as a secondary infection. This case illustrates a rare instance of lumbar spondylodiscitis caused by Klebsiella oxytoca. A direct biopsy via endoscopic surgery may be beneficial when multiple bacterial culture investigations yielded negative results. It is important to recognize that identifying the causative pathogen is crucial when initiating antibiotic therapy for spondylodiscitis.
Pyogenic spondylodiscitis is sometimes encountered, and its clinical course varies based on its etiology. It is typically caused by Staphylococcus aureus, followed by Escherichia coli.1) To our knowledge, there have been only 7 cases of spondylodiscitis in which the causative bacterium was Klebsiella oxytoca.1-6)Klebsiella oxytoca is a commensal bacterium; therefore, most bacteremia associated with this pathogen occurs in patients with underlying comorbidities or develops secondary to hepatobiliary infections.4) Herein, we present a rare case of lumbar spondylodiscitis caused by Klebsiella oxytoca, which was identified in samples obtained during endoscopic surgery.
A 77-year-old man presented with fever of unknown origin and back pain 1 month ago. His family physician examined a blood sample. His initial C-reactive protein (CRP) was 18.72 mg/dL, and white cell count (WBC) was 11,600/μL. At first, oral antibiotics with a new quinolone were administered for 2 weeks; however, the inflammatory markers in the blood tests further increased. Thus, it was changed to Cefcapene pivoxil and administered for 10 days. The fever gradually decreased, but his back pain worsened. His medical history included type 2 diabetes mellitus (DM), angina pectoris, benign prostatic hyperplasia, cervical spondylosis (C7/Th1 post-laminectomy), dyslipidemia, and appendicitis. He visited my outpatient clinic without a referral. He suffered from intolerable back pain but showed no neurological deficits. He was urgently admitted to our hospital.
Labo data and radiographical examinationOn admission, laboratory findings demonstrated increased CRP (4.86 mg/dL), WBC (10,800/μL), and erythrocyte sedimentation rate (71/mm at 1 hr). His hemoglobin A1c was 6.3% indicating good control. His renal function and coagulation tests (prothrombin time and activated partial thromboplastin time) were within normal ranges. Hepatitis B surface antigen and anti-hepatitis C virus antibody were negative. Blood and urine culture examinations were negative. Cardiac echocardiography showed no vegetations, suggesting infectious endocarditis.
A whole-body computed tomography (CT) scan revealed no infectious lesions. The L2 lower and L3 upper endplates were damaged, but the vertebral bodies were not yet severely destroyed (Fig. 1A). Magnetic resonance imaging (MRI) indicated that spondylodiscitis was identified at the L2/3 level. The cartilage endplates and intervertebral disc were grossly disrupted, displaying extensive high signals on T2-weighted images (Fig. 1B). He was diagnosed with spondylodiscitis.

Initial clinical images.
(A) Initial CT demonstrated bony erosion at the L2 lower and L3 upper endplates. Initial magnetic resonance imaging on emergent admission.
(B) T2-weighted image indicated extensive high signals that revealed spondylodiscitis at the L2/3 level.
CT: computed tomography
He wore a soft corset and remained on bed rest as much as possible. Considering the incidence of causative bacteria reported in previous literature, we selected ampicillin (ABPC), followed by clindamycin (CLDM) and sulbactam/ampicillin (SBT/ABPC) targeting Staphylococcus aureus and Escherichia coli. However, these drugs were not effective. Three weeks after admission, CRP was increased to 11.05 mg/dL, while WBC decreased to 5,900/μL. Therefore, we performed endoscopic surgery to obtain samples and perform drainage. The endoscopic surgery was performed via the left L2/3 transforaminal approach using the inside-outside technique after discography with indigocarmine. Intraoperative endoscopic images showed fragmented cartilage endplates and an abscess in the damaged intervertebral disc space. We collected endplates and lumbar disc materials for culture samples. Then, we irrigated the operative field with approximately 6 liters of saline solution and coagulated the fragmented tissues.
Postoperative outcomesThe next day, laboratory data showed a decreased CRP level (5.17 mg/dL). The intraoperative specimen culture was positive for Klebsiella oxytoca. As a result, the preoperatively administered ABPC, CLDM, and SBT/ABPC were resistant to this bacterium. Based on the susceptibility, we selected antibiotics consisting of 2,000 mg per day of cefotaxime (CTX) intravenously in combination with oral minocycline (MINO) at 200 mg per day for one month, during which his laboratory data returned to the normal range. We discontinued CTX, but continued MINO for another month until radiographical images showed improvement. His back pain gradually improved in parallel with laboratory and radiographic findings. We instructed him to wear a soft corset for 6 months. Finally, the collapsed vertebral bodies did not fuse on the postoperative 1-year CT image (Fig. 2A). There was remarkable deformity without instability on the postoperative 1-year dynamic X-rays (Fig. 2B-D). However, additional fusion surgery was not needed due to improvement in back pain and the absence of new neurological deficits.

Postoperative 1-year images.
(A) Bony fusion was not achieved on the postoperative 1-year CT image.
(B, C, D) Obvious instability was not present on the postoperative 1-year dynamic X-rays.
CT: computed tomography
Diagnostic delay in spondylodiscitis was an independent risk factor for an unfavorable outcome.7) To make an early diagnosis, prompt referral to experts and timely MRI investigation are important.7) Blood culture positivity in pyogenic spondylodiscitis was relatively low, 53.8%.8) Percutaneous CT-guided biopsy can increase bacterial identification rates to 47%-76%.9)
The causative bacteria in spondylodiscitis were Staphylococcus aureus (48.2%), followed by Escherichia coli (11.8%), coagulase-negative staphylococci (6.7%), and Pseudomonas aeruginosa (5.1%).1,7) Therefore, the detection of Klebsiella oxytoca as a causative bacterium of pyogenic spondylodiscitis is extremely rare. Klebsiella oxytoca can cause a wide range of diseases, including colitis, infective endocarditis, urinary tract infections, and respiratory infections.4) This pathogen is a commensal bacterium. As a result, most bacteremia associated with this pathogen primarily occurs in patients with underlying comorbidities such as DM or develops as a secondary infection.4) Other potential factors include prior antibiotics, urinary catheterization, corticosteroid treatment, recent surgery, and respiratory assistance.2) In the presented case, the patient had prior oral antibiotics for an infection of unknown origin and well-controlled DM. These factors may have contributed to immunological dysfunction and secondary infection in spondylodiscitis.
PubMed reported 7 cases of spondylodiscitis caused by Klebsiella oxytoca (Table 1).1-6) All patients whose details were described (cases number 1-4) were male and had previous infectious diseases. In our case, the invasive pathway could not be determined. Only case 2 involved a compromised host. The bacterium was not detected in systematic cultures in our case, but it was successfully isolated from the surgical specimens in cases 1, 2, and our case. All patients improved after proper management.
Summary of the Published Cases of Spondylodiscitis Caused by Klebsiella oxytoca
| No | Reference | Age | Sex | Comorbidities | Smoking | Alcohol | Steroid | Previous infection | Level | Surgery | Specimens | Antibiotics | Outcome |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| AP: angina pectoris, BPH: benign prostatic hyperplasia, CAZ: ceftazidime, CEZ: cefazolin, CPFX: ciprofloxacin, CS: cervical spondylosis, CTRX: ceftriaxone, CTX: cefotaxime, DL: dyslipidemia, DM: diabetes mellitus, HT: hypertension, LVFX: levofloxacin, M: male, MINO: minocycline, NA: not assigned, PLF: posterolateral fusion, UTI: urinary tract infection | |||||||||||||
| 1 | Brooks, 2022 [1] | 79 | M | BPH, HT, Obesity | NA | NA | None | UTI | Th5/6 | Needlebiopsy→PLF | Blood, urine, and lumbar disc | CEZ→CTRX→CPFX | Improvement |
| 2 | Sabio, 2002 [2] | 51 | M | Heroin addiction | NA | NA | None | Skin ulcers | Th12/L1 | Needle biopsy | Lumbar disc | CTRX | Improvement |
| 3 | Agrawal, 2009 [3] | 54 | M | NA | + | + | None | Sepsis after prostate biposy | C3/4 | None | Blood | CTRX→CPFX | Improvement |
| 4 | Iwadare, 2023 [4] | 76 | M | HT, DM | + | + | None | Choledocholithiasis | L4/5 | Retroperitoneal abscess drainage | Blood, urine, and retroperitoneal abscess | LVFX | Improvement |
| 5 | Burkhardt, 2019 [5] | NA | NA | NA | NA | NA | NA | NA | Cervical | + | NA | NA | NA |
| 6 | Park, 2016 [6] | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA | NA |
| 7 | |||||||||||||
| 8 | Our presented case | 77 | M | AP, BPH, CS, DM, DL | + | − | None | Unknown origin | L2/3 | Endoscopic biopsy and saline irrigation | Lumbar disc | CTX→CAZ→MINO | Improvement |
The clinical guidelines for the diagnosis and treatment of vertebral osteomyelitis by the Infectious Diseases Society of America do not recommend surgical debridement and/or stabilization in patients whose clinical symptoms, physical examinations, and inflammatory markers have improved despite progressing bony invasion at 4-6 weeks.10) However, the guidelines suggest the following surgical indications: 1) Progressive neurologic deficits, instability, or deformity caused by the collapsed vertebral body and intervertebral disc space; 2) uncontrollable infection despite proper medical treatments, including antimicrobial therapy. In our presented case, conservative treatments failed, as the previous doctor had already administered medication for 4 weeks without an effective course. We then changed to another antibiotic therapy for 3 weeks, combined with bed rest and a corset. Prolonged bed rest sometimes leads to disuse and its associated complications.11) Therefore, we selected a less invasive endoscopic surgery. Recently, several reports have described the usefulness of endoscopic procedures as an alternative to conventional surgical treatments in the subacute stage.9,12) Endoscopic surgery can collect adequate specimens, yielding higher diagnostic efficacy than blood cultures or CT-guided needle biopsies.9) Furthermore, the latest systematic review revealed that the endoscopic surgery group (84.38%) had a higher pathogen identification rate than the CT-guided biopsy group (36.59%).13)
Endoscopic surgery proved effective for lumbar spondylodiscitis that was refractory to conservative therapy, utilizing its minimally invasive characteristics. Identifying the causative pathogen may improve the success of postoperative antibiotic treatment and help avoid invasive procedures such as fusion surgery.
This case illustrated lumbar spondylodiscitis rarely caused by Klebsiella oxytoca. A direct biopsy by endoscopic surgery may be beneficial when several bacterial culture investigations yield negative results. We need to recognize that identification of the causative pathogen is very important when introducing antibiotic therapy for spondylodiscitis.
The authors have nothing to disclose.
Consent from the patient was obtained.
All authors have no conflict of interest.