A Case of Native Joint Septic Arthritis Caused by

Background

Native joint septic arthritis (NJSA) is an uncommon condition, with an incidence rate of 12/100 000 person-years [1]. Mortality is variable depending on co-morbid conditions and causative organism, but a case series studying 109 patients with NJSA found a mortality rate of 5.6% [2]. The most common organisms isolated in septic arthritis are Staphylococcus aureus, Streptococcus pyogenes, and methicillin-resistant Staphylococcus aureus, accounting for 53%, 15%, and 13% of cases, respectively [3]. On the contrary, septic arthritis due to anaerobic bacteria, particularly B. fragilis, is exceedingly rare. For example, a review of 308 patients with native joint septic arthritis found only 2 patients who had B. fragilis as the causative organism [4]. B. fragilis NJSA is more often attributed to hematogenous spread, with detection of B. fragilis in the blood 65% of the time [4].

B. fragilis is an obligate anaerobe and gram-negative rod that is typically isolated from the oral cavity, gastrointestinal tract, genitourinary system, and skin [4]. Clinically significant infections caused by B. fragilis have been found in many areas throughout the body, including intra-abdominal, intestinal, hepatobiliary, skin and soft tissue, blood, lung, urinary, gynecologic organs, brain, cerebrospinal, oral, cardiac valve, and bone [5]. Risk factors for B. fragilis infection include inflammatory arthritis, malignancy, sickle cell disease, and ischemic heart disease [4].

We report a case of B. fragilis NJSA, which we attribute to direct inoculation from a prior knee injection in the setting of negative blood cultures, in a patient with multiple risk factors, including cardiac disease, kidney disease, and diabetes.

Case Report

A 65-year-old man with a complex past medical history, including coronary artery disease with stent, heart failure, atrial fibrillation, sick sinus syndrome with a pacemaker, chronic kidney disease, diabetes mellitus, gout, asthma, hypertension, obstructive sleep apnea, anxiety, and prostate cancer status post prostatectomy, presented to the emergency room with a chief concern of 5 days of right knee pain associated with redness, swelling, and warmth. His symptoms had progressively worsened such that he was unable to bear weight on the affected leg. He denied any fevers or chills. Of note, 2 weeks prior to his presentation, he sought care for a gout flare of his right knee and was given colchicine and prednisone. He had undergone steroid injections in the right knee every 3 months, the last of which was 2 months prior to the current presentation. He denied intravenous drug use. His medication list included albuterol, fluticasone-vilanterol, linaclotide, losartan, amlodipine, famotidine, montelukast, furosemide, glipizide, clonazepam, pantoprazole, evolocumab, and febuxostat.

His vital signs included a temperature of 38.5°C, heart rate of 80 beats per minute, blood pressure of 137/73 mmHg, respiratory rate of 16 breaths/minute, oxygen saturation of 98% on room air, and weight of 92.2 kilograms. His physical exam was remarkable for a right knee that was erythematous, warm, and swollen, and was exquisitely tender to palpation. The patient was unable to tolerate passive range of motion. The remainder of the joints assessed were unremarkable. Auscultation of his heart revealed an irregularly irregular rhythm but was otherwise negative for murmurs.

Notable lab results included a white blood cell count (WBC) of 16.7 K/cumm (reference range 4.5–10 K/cumm) with 78% neutrophils and 1% bands, hemoglobin 12.7 g/dL (reference range 13.5–16.5 g/dL), blood urea nitrogen (BUN) 48 mg/dL (reference range 8–20 mg/dL), creatinine 1.86 mg/dL (reference range 0.64–1.27 mg/dL), c-reactive protein (CRP) 9.15 mg/dL (reference range 0–0.74 mg/dL), and erythrocyte sedimentation rate (ESR) 49 mm/hr (reference range 0–20 mm/hr). Light brown, purulent synovial fluid from the right knee was obtained prior to antibiotic administration and was notable for a nucleated cell count of 218 325 K/cumm (reference range 13–180 K/cumm) with 87% neutrophils (reference range 0–25%), and a red blood cell count of16 127 K/cumm (reference range 0–2000 K/cumm). Sterile specimens were sent to the microbiology lab for aerobic and anerobic screening, revealing no organisms on original gram stain. Cultures plated per hospital protocol on multiple culture media including Blood, Chocolate, MacConkey, and Brucella with H&K (BRUHK) agar showed no growth at 24 h. A radiograph of the right knee demonstrated a small-to-moderate-size joint effusion (Figure 1).

The patient was started on empiric treatment with vancomycin and ceftriaxone for septic arthritis. Orthopedic Surgery was consulted for surgical management. The patient was taken to the operating room on hospital day 1 for irrigation and debridement of his right knee, without any complications. By hospital day 2, the synovial fluid culture obtained from admission revealed heavy growth of B. fragilis on BRUHK media (Hardy Diagnostics A30), and subsequent intraoperative cultures grew B. fragilis as well. As such, per Infectious Disease recommendations, the patient’s antibiotic regimen was changed to intravenous metronidazole with an eventual plan for a 4-week treatment course on oral metronidazole. The source of infection was attributed to direct inoculation from his prior corticosteroid joint injection, given that he had multiple blood cultures that were negative and no other potential sites of infection were identified. Although the patient remained afebrile and his leukocytosis resolved post-operatively, he continued to have pain and swelling with an associated rise in CRP to 25.4 mg/dL on hospital day 4. Orthopedic Surgery thus planned for repeat irrigation and debridement, but on hospital day 5, the patient left the hospital against medical advice and unfortunately was lost to follow-up thereafter.

Discussion

No controlled trials or observational studies regarding the management of NJSA due to B. fragilis could be found in the literature. As with cases of NJSA due to other organisms, treatment of NJSA due to B. fragilis is thought to require drainage of the affected joint in combination with antibiotic therapy.

The optimal means of achieving source control in NJSA in general remain uncertain. Needle aspiration, arthroscopy, and arthrotomy have all been explored in the literature, but none of these modalities have demonstrated statistically significant superiority [6]. Arthrotomy and needle aspiration have both been employed as modes of source control in cases of NJSA due to B. fragilis [7–12]. However, this has been shown in some case reports to be insufficient, and 1 case report described the need for hemipelvectomy due to difficulty with achieving adequate source control [13].

In addition, antibiotic selection has varied widely in documented cases of NJSA due to B. fragilis with agents including chloramphenicol, clindamycin, doxycycline, erythromycin, meropenem, metronidazole, piperacillin/tazobactam, and tobramycin prescribed, both individually and in combination, after synovial fluid culture speciation [7–12]. Metronidazole is currently the most commonly used antibiotic in this setting [4]. Metronidazole demonstrates excellent in vitro activity against B. fragilis and has been shown to penetrate well into synovial fluid, thus making it an extremely viable option for treatment [14,15]. In addition, metronidazole resistance is rare, with >99% of gram-negative anaerobes demonstrating susceptibility to this agent. This demonstrates superiority to alternatives with B. fragilis showing 10–42%, 1–22%, and 0–5% resistance to clindamycin, meropenem, and piperacillin/tazobactam, respectively [14]. Metronidazole is generally well-tolerated, with common adverse effects including nausea, diarrhea, and abdominal pain. However, extended courses of metronidazole (longer than 4 weeks) may result in peripheral neuropathy, which is reversible upon discontinuation [16].

The duration of antibiotic therapy in cases of NJSA due to B. fragilis in the literature has ranged from 3 to 16 weeks, whereas 3 to 6 weeks of antibiotic therapy are traditionally recommended in the setting of NJSA in general [4]. Of note, a recent unblinded randomized trial of NJSA patients following surgical management compared a 2-week course of antibiotics with a 4-week course, finding the 2-week course to be non-inferior [17]. This approach has not yet been attempted in any documented case of NJSA due to B. fragilis.

NJSA due to B. fragilis typically responds well to the therapy outlined above. In the literature, we found documentation of only a single recorded death due to B. fragilis NJSA and a single case of relapsed disease, both as outlined below.

The documented mortality occurred in a 71-year-old man with seropositive rheumatoid arthritis and NJSA of his right knee and elbow. The patient had additional comorbidities including chronic obstructive pulmonary disease, coronary artery disease, and congestive heart failure, similar to our patient. Prior to his death, the patient had received surgical drainage of the affected joints and antibiotic therapy with ampicillin, chloramphenicol, clindamycin, and tobramycin. No autopsy was performed [8].

The case of relapsed disease occurred in a 44-year-old man with insulin-dependent diabetes mellitus and NJSA of his right knee. During his initial hospitalization, this patient received a right knee arthrotomy and a 5-week course of clindamycin following the isolation of B. fragilis in his synovial fluid. The patient was readmitted 3 weeks after discharge for management of purulent drainage from his right knee, and B. fragilis was again isolated from the patient’s synovial fluid. The patient received surgical drainage of the affected joint and antibiotic therapy with a 19-day course of clindamycin followed by a 2-week course of doxycycline. The patient recovered and retained full range of motion in his right knee thereafter [8].

Conclusions

NJSA due to B. fragilis is very rare. It is more common in the setting of B. fragilis bacteremia, occurring in about 65% of cases, or in the setting of contiguous spread from another concomitant infection, occurring in about 25% of patients [4]. In our case described above, the patient did not have B. fragilis isolated from the bloodstream, nor was a local concomitant infection present, so the infection was attributed to direct inoculation from the prior corticosteroid injection.

Management requires both antibiotic therapy and source control with surgical drainage. However, the preferred modality of surgical drainage remains controversial [6]. The duration of antibiotic therapy after surgical drainage is dependent upon whether there is a concomitant primary infectious process, such as endocarditis or osteomyelitis, as these would require longer treatment courses. If none of these are present, then 2 weeks of targeted antibiotic therapy is now considered adequate [17].