The efficacy and the safety of intra-articular injections of sodium hyaluronate were studied in patients with osteoarthritis of the knee in a randomized multicenter double-blind study. Two hundred and nine patients received five injections of either 25 mg hyaluronate/2.5 ml (verum, N = 102) or 0.25 mg hyaluronate/2.5 ml (control, N = 107) at weekly intervals. Seven patients in each group were excluded from the protocol-correct efficacy analysis. The Lequesne Index, the first main criterion, showed a significant superiority of the verum-treated patients after the third injection up to the final follow-up examination 9 weeks after the last injection (MANOVA, P < 0.025). The consumption of paracetamol was defined as a complementary main criterion that did not reveal significant differences between the treatment groups. Most of the individual secondary endpoints demonstrated a much better response to the active treatment without reaching the significance level in the intergroup comparisons for the single time-points. Side-effects were confined to local reactions of minor severity and short duration in four patients (six events) of the verum group and in five patients of the control group. Clinical chemistry and hematology remained essentially unchanged.