Putative disease-modifying drugs are usually clinically used in osteoarthritis with two main aims: not only stopping or reducing the cartilage degenerative process after a long-term treatment, but also controlling the symptoms of the disease within a few days or weeks, thus avoiding or diminishing the use of symptomatic medications. Due to the difficulties of implementing the first aim, the latter aim was more often investigated, even if most often with inadequate study design and insufficient numbers of patients. We have recently carried out three double-blind, controlled, parallel groups, randomized, 4-6 week trials of glucosamine sulphate versus placebo or the NSAID ibuprofen on a total of 606 gonarthrosic out-patients. Movement limitation and pain were scored according to the Lequesne index, and the efficacy goals were strictly pre-determined. Access to other medications was not allowed. Glucosamine was significantly more effective than placebo, while no difference was detected in comparison with the NSAID (p < 0.025 and p = 0.77, respectively: Fisher's two-tailed exact test). On the other hand, glucosamine was as well tolerated as placebo, while the percentage of patients suffering adverse drug reactions was higher in the ibuprofen group (37% vs 7%: p < 0.001). Long-term trials are in progress and several aspects are to be considered in their design: they must be double-blind, placebo-controlled, randomized, continued for a period of years and (most importantly) with the careful use of imaging and biochemical techniques capable of generating objective evaluation criteria.