Objectives To evaluate the medico-economic benefits of Hylan G-F 20 over 9 months in the treatment of patients suffering from knee osteoarthritis (OA).

Methods Patients with radiographically proven knee OA, pain on movement ³ 40 mm (VAS) and treated by at least two 10-day NSAIDs courses or by slow-acting anti-OA drugs within the previous 3 or 2 months respectively were enrolled by 81 centres (21 hospitals and 60 private rheumatologists). After stratification on known OA risk factors (patients? age, BMI,...), patients were centrally randomised in Synvisc (S) or in usual treatment (UT) groups. S-patients were treated by 3 intraarticular injections of Hylan GF 20 one week apart and encouraged to stop or spare any other OA treatments. If needed S-patients could be treated by Synvisc in both knees. Primary outcome was difference in specific knee-OA cost (visits, exams, hospitalisations, treatments and adverse event treatment/prevention related to OA treatments) divided by difference in AUC mean change of Lequesne index between groups. Data were collected at baseline, day 28, months 3, 6 and 9.

Results Out of 518 patients randomised, 506 (253 per group) were treated and evaluable. Both groups were comparable at baseline. Forty-five S-patients (18%) were treated in both knees. Cost per patient was analysed over 9 months in social security perspective (year 1998 value). All direct and indirect costs were decreased in S-group compared to UT-group (medications ? 40%, non-drug treatments ? 19%, hospitalisation ? 5%, exams ?2%, seak-leave ? 15%) with the only exception of physician visits cost (+ 33%) due to Synvisc injection procedure.

Specific OA total costs (including Synvisc) were 829.12 Euros in S-group and 821.34 Euros in UT-group. Synvisc cost in social security perspective (year 2000 value) is 99.09 Euros per treatment.

AUC mean change in Lequesne Index was 7.9 points in S-group vs 9.9 in UT-group (p < 0.0001). All together the cost/efficacy ratio (including Synvisc cost) was 7.78 Euros for a 2-point improvement of Lequesne index (difference clinically significant) between groups.

Conclusion Viscosupplementation by Synvisc improves algofunctional status of patients and has no significant impact on specific knee OA cost.

Acknowledgement Synvisc Experts Group: Prs. Bardin T., Dougados M., Kahan A. (Paris), Mazières B (Toulouse), Trèves R. (Limoges), Vignon E. (Lyon), Drs Auvinet B. (Laval), Naim C. (Marseille), Rossignol O. (Fontenay).