Background: Hyaluronan, or hyaluronic acid (HA), is the major hydrodynamic nonprotein component of joint synovial fluid (SF). Its unique viscoelastic properties confer remarkable shock absorbing and lubricating abilities to SF, while its enormous macromolecular size and hydrophilicity serve to retain fluid in the joint cavity during articulation. HA restricts the entry of large plasma proteins and cells into SF but facilitates solute exchange between the synovial capillaries and cartilage and other joint tissues. In addition, HA can form a pericellular coat around cells, interact with proinflammatory mediators, and bind to cell receptors, such as cluster determinant (CD)44 and receptor for hyaluronate-mediated motility (RHAMM), where it modulates cell proliferation, migration, and gene expression. All these physicochemical and biologic properties of HA have been shown to be molecular weight (MW) dependent. Objective: Intra-articular (IA) HA therapy has been used for the treatment of knee osteoarthritis (OA) for more than 30 years. However, the mechanisms responsible for the reported beneficial clinical effects of this form of treatment remain contentious. Furthermore, there are a variety of pharmaceutic HA preparations of different MW available for the treatment of OA, but the significance of their MWs with respect to their pharmacologic activities have not been reviewed previously. The objective of the present review is to redress this deficiency. Methods: We reviewed in vitro and in vivo reports to identify those pharmacologic activities of HA that were considered relevant to the ability of this agent to relieve symptoms and protect joint tissues in OA. Where possible, reports were selected for inclusion when the pharmacologic effects of HA had been studied in relation to its MW. In many studies, only a single HA preparation had been investigated. In these instances, the experimental outcomes reported were compared with similar studies undertaken with HAs of different MWs. Results: Although in vitro studies have generally indicated that high MW-HA preparations were more biologically active than HAs of lower MW, this finding was not confirmed using animal models of OA. The discrepancy may be partly explained by the enhanced penetration of the lower MW HA preparation through the extracellular matrix of the synovium, thereby maximizing its concentration and facilitating its interaction with target synovial cells. However, there is accumulating experimental evidence to show that the binding of HAs to their cellular receptors is dependent on their molecular size; the smaller HA molecular species often elicits an opposite cellular response to that produced by the higher MW preparations. Studies using large animal models of OA have shown that HAs with MWs within the range of 0.5 × 106-1.0 × 106 Da were generally more effective in reducing indices of synovial inflammation and restoring the rheological properties of SF (visco-induction) than HAs with MW > 2.3 × 106 Da. These experimental findings were consistent with light and electron microscopic studies of synovial membrane and cartilage biopsy specimens obtained from OA patients administered 5 weekly IA injections of HA of MW = 0.5 × 106-0.73 × 106 Da in which evidence of partial restoration of normal joint tissue metabolism was obtained. Conclusions: By mitigating the activities of proinflammatory mediators and pain producing neuropeptides released by activated synovial cells, HA may improve the symptoms of OA. In addition, HAs within the MW range of 0.5 × 106-1.0 × 106 Da partially restore SF rheological properties and synovial fibroblast metabolism in animal models. These pharmacologic activities of HA could account for the reported long-term clinical benefits of this OA therapy. However, clinical evidence has yet to be described to support the animal studies that indicated that HAs with MW > 2.3 × 106 Da may be less effective in restoring SF rheology than HAs of half this size. Semin Arthritis Rheum 32:10-37. Copyright 2002, Elsevier Science (USA). All rights reserved.