OBJECTIVES: To develop a cost-effectiveness and budget impact analysis of intra-articular sodium hyaluronate 25mg/2.5ml (SH) versus total knee arthroplasty (TKA) in the treatment of knee osteoarthritis grades II and III of Kellgren-Lawrence classification, in patients who failed conservative treatment, under the Brazilian public healthcare system (BPHS) perspective. METHODS: A Markov model was developed to project costs and outcomes associated with knee osteoarthritis progression in 3 years. In the model all patients were eligible to surgery and could either receive SH or undergo TKA immediately. SH could delay or cancel surgery. The cycle duration was 3 months and the corresponding transition probabilities were obtained from the literature. The outcome was expressed as the number of avoided TKAs. The analysis considered only direct medical costs, including drugs, physician visits, procedures and complications treatment. Costs and outcomes were discounted at 5% per year. A budget impact analysis was developed considering the estimated number of TKAs in the BPHS in 2009. Main parameters were evaluated in a sensitivity analysis. RESULTS: In a 3-year time horizon, total costs per patient were US$3445 (2005-PPP-index USD1.0=R$1.4) for the SH group and R$4381 for the TKA group. The reduction in the absolute number of surgeries was 72.8% in the SH group, compared with TKA. Likewise, there was a cost reduction of US$936, in favor of SH. Considering 3991 eligible patients in the public system, the total annual costs for SH and TKA groups were US$5,136,234 and US$17,484,003, respectively. As a result, the budget impact was US$12,347,769, in favor of SH. The sensitivity analysis about the time horizon showed a cost-saving result until 4 years. CONCLUSIONS: The use of SH after failure of conservative treatment in patients with grades II or III knee osteoarthritis showed to be effective, leading to a reduction in costs and number of surgeries.