The objective of this study was to test if spa therapy can play a role in the management of severe knee osteoarthritis (OA). Twenty patients with radiologically and clinically severe knee OA were randomly assigned into spa and drug therapy groups. Spa group (n = 10) traveled to a spa town and stayed at a hotel for a 10-day spa therapy course. They followed a balneotherapy regimen including thermal pool baths at 37 degrees C for 20 min two times daily. Drug therapy group (n = 10) stayed at home and followed their individually prescribed drug therapy (NSAIDs and paracetamol). Patients were assessed at baseline (week 0), after spa therapy at 2 weeks (week 2) and during follow-up period at 12 (week 12) and 24 (week 24) weeks by a blinded investigator. Patients assessed with Lequesne algofunctional index (LAFI), pain (visual analogue scale, VAS), patient's and investigator's global evaluation (VAS), ten-stairs stepping up and down time, 15 m walking time and three times squatting up and down time. Significant improvement in pain and LAFI scores were found at week 2, week 12 and week 24 in the spa therapy group compared to baseline. Comparing the two group differences, spa therapy was superior to drug therapy in pain reduction and in physician's global assessment at all time points. This superiority was also found in LAFI scores and patients' global assessments at week 12 and week 24. A 10-day course of spa therapy may be beneficial in short- and medium-term up to 24 weeks by reducing pain and improving functional status and overall well-being in patients with severe knee OA and may be considered as an effective therapeutic tool for such patients in countries like Turkey where it is widely available and (at least partly) reimbursed.