Genicular Artery Embolization, Radiofrequency Ablation, and Corticosteroid Therapy for Knee Osteoarthritis: A Cost-Effectiveness Analysis Using Randomized Clinical Trial Data.

Background: Many patients with symptomatic knee osteoarthritis (KOA) are refractory to traditional nonsurgical treatments such as intraarticular corticosteroid (CS) injection but are not yet eligible for or decline surgery. Genicular artery embolization (GAE) and radiofrequency ablation (RFA) are emerging adjunctive or alternative minimally invasive treatments. Objective: To perform a cost-effectiveness analysis (CEA) comparing CS, GAE, and RFA, for treatment of symptomatic KOA using a Markov model based on a de novo network meta-analysis (NMA) of randomized control trials. Methods: CEA was conducted to compare GAE and RFA to CS using a Markov cohort state-transition model from a U.S. Medicare payer's perspective over a 4-year time horizon. The model incorporated each treatment's success and attrition rates, costs, and utility benefit. Utility benefit values were derived at short-term (0.5-3 months) and long-term (6-12 months) posttreatment follow-up from NMA of published RCTs using an outcome of improved knee pain and/or function. Analyses were conducted at a willingness-to-pay threshold of $100,000 per quality-adjusted life year (QALY). Sensitivity analyses were performed, including when simulating various cost setting scenarios (i.e., office vs hospital outpatient treatment). Results: RFA demonstrated larger treatment effect than GAE, more pronounced at short-term [standardized mean difference (SMD), -1.6688, 95% CI [-2.7806; -0.5571], p=.003] than long-term (SMD -0.3822, 95% CI [-1.9743; 1.2100], p=.64) follow-up. Across cost setting scenarios, incremental cost-effectiveness ratios relative to CS were $561-1563/QALY for GAE versus $76-429/QALY for RFA (not counting scenarios in which RFA was dominated by CS). GAE demonstrated higher cost-effectiveness probability compared to RFA (41.6-54.8% vs. 18.4-29.2%, respectively). GAE was more cost-effective than RFA when the GAE clinical success rate and post-GAE utility value exceeded 32.1-51.0% and 0.562-0.617, respectively, and when the GAE quarterly attrition rate was less than 8.8-17.4%. RFA was more cost-effective when baseline pre-treatment utility values exceeded 0.695-0.713. Neither GAE costs nor RFA costs were sensitive parameters. Conclusion: Across scenarios, GAE was consistently the most likely cost-effective treatment option compared to RFA and CS, although clinical success rates, attrition rates, and utility values impact its cost-effectiveness. Clinical Impact: GAE is likely to be more cost-effective than RFA or CS for treatment of symptomatic KOA.