Dear Editor,
Osteoarthritis (OA) remains the most common joint disorder afflicting patients in the United States. Knee osteoarthritis often presents as discomfort, stiffness, swelling, and functional impairment and is one of the most common causes of disability in the world [1]. Despite advances in medical management, there remains a large subset of patients with knee OA who achieve suboptimal analgesia and disability despite analgesics and intra-articular injections. Radiofrequency ablation (RFA) of the genicular nerves is an alternative that has been shown to offer significant pain reduction and functional improvement in this patient subset, with very few complications reported [2–5]. The most significant complication reported was a third-degree skin burn from RFA of the inferiomedial genicular nerve [6–8]. We would like to report the first significant infectious complication of genicular nerve RFA.
The patient was a 67-year-old female (body mass index 29.1 kg/m2) with a medical history significant for coronary artery disease with drug-eluding coronary stents (on clopidogrel), hypertension, non-insulin-dependent diabetes mellitus (well controlled, last hemoglobin A1c <6%), morbid obesity, and knee osteoarthritis with significant functional impairment. She had had inadequate relief from multiple interventions for her knee pain, including oral duloxetine and intra-articular injections with viscosupplementation performed years prior. Plain x-rays of the knees showed bilateral advanced osteoarthritis with bone-on-bone apposition. As the patient continued to achieve only supoptimal analgesia with impaired function, the decision was made to perform genicular nerve blocks [4]. These blocks were performed twice, each time yielding >80% temporary pain relief and improved function.
She then underwent superiomedial, superiolateral, and inferiomedial genicular nerve RFA using an accepted protocol using cooled RFA [5,9]. Patient had normal vital signs including being afebrile and no effusion when she presented for the ablation. An aseptic technique was used for patient preparation, including chlorhexidine prep (allowed to dry for three minutes) and all personnel wearing hats and masks and the interventionist wearing additional sterile gloves and gown. One percent lidocaine was used to anesthetize the target points, and a 17-gauge 75-mm introducer needle with a 4-mm active tip was placed over the target sites (Coolief, Halyard Health, Alpharetta, GA, USA). Anterior-posterior views and lateral fluoroscopic views were used to confirm appropriate placement of the electrode. Motor nerve stimulation was performed at 2 V at 2 Hz without a motor response in the muscles around the knee joint, and the nerves were blocked using 1 mL of a mixture of 1:1 bupivacaine 0.5% and lidocaine 2%. Each site was lesioned for two minutes and 30 seconds to a set temperature of 60°C with normal impedances throughout the ablations, and needles were flushed with a total of 6 mL of 0.5% bupivacaine and 20 mg of triamcinolone, divided at the three target sites per side, to prevent postablation neuritis. Each puncture site was covered with pressure bandages, and instructions were given to the patient to keep the sites clean and dry and to contact the pain clinic if any new symptoms occurred. The patient was discharged home, reporting immediate relief from the procedure. However, within the subsequent 24 hours, the patient endorsed “fullness” and increasing pain in the left knee. The patient was advised to return to the emergency room, where her left knee was noted to have increased warmth, decreased range of motion, and fullness. Her vital signs were normal, including being afebrile. A plasma leukocyte elevation of 18.4 k/mm3 (reference range 4.8–10.8 k/mm3) was demonstrated, as well as a large effusion noted on computed tomography imaging of the left knee (Figure 1). The knee joint was then aspirated under ultrasound, yielding 60 mL of purulent fluid (Figure 2) that contained 82,466 leukocytes with 96% polymorphs and a negative gram stain. Given the presentation and aspirate results, a presumed diagnosis of septic arthritis was established, and intravenous ceftriaxone and vancomycin were empirically started. A left knee arthroscopy with washout was undertaken shortly after admission. Aspirate and synovial fluid cultures obtained intraoperatively grew methicillin-sensitive Staphylococcus aureus. The patient’s antibiotics were subsequently deescalated to intravenous cefazolin given bacterial culture sensitivity findings. To ensure thoroughness, a transesophageal echocardiogram was obtained and was negative for any vegetations. A peripherally inserted central catheter was placed once blood cultures were negative, and parenteral antibiotics were continued postdischarge at a skilled nursing facility. After approximately three months, the patient was deemed to have cleared the infection and she had left total knee arthroplasty five months after the initial presentation of infection.
Large effusion demonstrated on computed tomography imaging of the left knee. Coronal view (A) used for location at the epicondyles, with an axial view (B) of a large anterior effusion (white arrow).
Aspirate of left-sided prepatellar effusion, which yielded 80,000 white blood cells, 96% leukocytes, and a culture of Staphylococcus aureus.
Based upon this case, we speculate that while performing genicular nerve RFA, entrance into the joint capsule is possible. The cephalad aspect of the joint capsule, particularly the suprapatellar pouch, can be inadvertently traversed during the placement of the superior introducer needles. The superior-lateral approach and superior-medial approach to accessing the knee joint have quite similar needle trajectories compared with genicular nerve RFA needle placement. In fact, these approaches have been touted to have the most direct access to the synovium [10]. The authors speculate that the knee joint was traversed during RFA, with through-and-through penetration of the capsule, with the electrodes remaining extra-articular. Any contamination may not only lead to a superficial infection, but also a septic joint. In addition, any subclinical infection may be worsened by needle distribution.
As genicular nerve RFA is performed with increasing frequency as a safe interventional alternative for treating knee pain, the authors bring light to the first reported case of septic arthritis as a postprocedure complication. There remain questions about the etiology, as strict adherence to aseptic technique was maintained throughout. Upon further questioning, the patient did endorse having recently seen a hematologist for leukocytosis of unknown etiology on routine lab work, which was resolved before the procedure. One would expect an infection solely related to contamination during the procedure to occur within 72 hours, not necessarily within 24 hours. Despite this, genicular nerve RFA remains the likely source of her infection. This case serves to reinforce the importance of maintaining vigilance for sterile precautions and emphasizes the discussion of serious implications of joint infection when obtaining informed consent. Additionally, this case raises the question moving forward of what strategies should be followed for mitigating infection risk. For providers, there is perhaps a place for individualized consideration of antibiotic prophylaxis to cover organisms that are common isolates in septic arthritis, particularly skin flora with prophylactic antibiotics such as parenteral cefazolin. However, more research is needed to see if this is needed in all patients or perhaps only a subset of patients who are at high risk (such as patients with diabetes mellitus). In addition, there may be a role for changing the technique of electrode placement—coming in at more of an angle (starting more lateral for the superiorlateral electrode and starting more medial for the superiormedial electrode)—to lessen the chance of traversing the joint space. The authors hope this case raises awareness to a potential adverse outcome and aids in improving provider practice and considerations when performing genicular nerve RFA.
