We have previously demonstrated the efficacy of therapeutic exercise for osteoarthritis (OA) of the knee. This study was performed to examine the additive effects of glucosamine or risedronate on the exercise therapy. In this study, 142 female patients with moderate OA of the knee, who had been recommended to undergo home exercise at the first visit to the hospital, were randomly given glucosamine hydrochloride, risedronate, or no additive. Although improvement after 18 months was observed in all groups using individual scales for evaluation of pain and function of the knee, no significant differences were observed between the groups regarding any of the scales, indicating no significant additive effect of glucosamine or risedronate. One reason for the lack of effect of glucosamine or risedronate on OA of the knee may be that the effect of these agents was occluded by the effect of therapeutic exercise to improve pain and function of the knee. This finding means that even if glucosamine and risedronate were to have an effect on OA of the knee, the effect would not be greater than the effect of knee exercise to improve the symptoms.