Sticking it to Them

Abstract & Commentary

Synopsis: A series of five weekly intra-articular injections of hyaluronic acid produced slightly more improvement than did sham injections plus 500 mg of naproxen twice daily for six months. The benefit of the knee injections for subjects with osteoarthritis lasted throughout the 26 weeks of study and fewer gastrointestinal adverse events were noted in those receiving the hyaluronic acid injections than the naproxen group.

Source: Altman RD, et al. J Rheumatol 1998;25:2203-2212.

A patient with osteoarathritis (oa) of the knees once suggested injecting "grease" into his knees. His layman’s appreciation of the need for better "lubrication" is supported by the finding that the molecule that is principally responsible for the viscosity of synovial fluid, high molecular weight hyaluronic acid (HA), is decreased in OA.1 Physicians have two choices of viscous HA containing fluids that may be injected into knees. Altman and colleagues report their experience with one, hyaluronan (Hyalgan). The other FDA-approved product is hylanG-F 20 (Synvisc), a higher molecular weight product.

The efficacy and safety of hyaluronan was the subject of a multicenter, double-blind, randomized, placebo-controlled study that compared three groups. One group received five weekly injections of 20 mg (2 cc) of HA and twice-daily placebo (HA group, 105 completers). The second received five sham injections and oral placebo (placebo group, 115 completers). The third group received sham injections plus naproxen sodium 500 mg twice daily (naproxen group, 113 completers). The primary outcome measure was the amount of pain experienced during a 50-foot walk. Secondary measures included 50-foot walk time, patient and physician global evaluations, amount of additional analgesics used, and adverse effects. All patients had radiographic evidence of OA.

The mean amount of pain reported at selected times by the subjects after a 50-foot walk is indicated in the figure. (See Figure.) Pain worsened in all three groups during the two-week washout phase. The amount of pain at the baseline evaluation was not statistically different. All three groups reported progressively less pain during the first five weeks of study. When subjects who completed the 26-week study were analyzed, the HA group had, on average, less pain than the other two groups from five weeks to 26 weeks. The mean difference in pain between the HA and naproxen groups at the end of 26 weeks was 8.8 mm and was statistically robust (P = 0.004). However, in a post-hoc, intention-to-treat analysis, which used the last observation recorded for each subject, and included those who withdrew prior to the 26 week visit. Pain means of 27 mm, 28 mm, and 25 mm were recorded at the last visit for the HA, placebo, and naproxen groups, respectively. These were not significantly different. Gastrointestinal (GI) adverse events were more common in the naproxen group (41%) than in the HA (29%) or sham plus placebo (36%) groups. Dropouts were also more frequently due to GI problems in the naproxen group (28%) than the HA (7%) or the sham plus placebo (7%) groups, but total dropouts for all causes were not significantly different for the three groups. In the three groups, 59, 53, and 50 subjects dropped out of the HA, sham plus placebo, and sham plus naproxen groups, respectively.

Comment by Jerry M. Greene, MD

The remarkable thing in this study is the magnitude of the response to the sham injections. From a baseline of mean of about 50 mm on the 100 mm visual analog pain scale, the completers in the sham plus placebo group ended with a mean of about 20 mm, or about a 60% reduction in pain. The completers in the HA injection group, by comparison, achieved about a 70% reduction in pain. The long-lasting effect of the sham injections is also striking, with relatively sustained pain relief from nine weeks to 26 weeks of study. Whether the roughly 10% increment in improvement with HA, beyond that achieved with sham injection alone seen in the "completers" is worth the $500-600 wholesale cost of HA is an interesting question—especially when the post-hoc intention-to-treat analysis showed no efficacy advantage for HA vs. placebo or naproxen. This suggests to me that the role for HA injections is limited. The patient for whom nonsteroidal anti-inflammatory drugs (NSAIDs) are contraindicated or are not tolerated is one example. A patient at high risk of NSAID adverse effects and who needs continuous ulcer prophylaxis would incur the costs of both NSAIDs and antiulcer therapy is another.

Both the higher and lower molecular weight HA injections have been approved under FDA regulations as medical devices, not as medications. As a result, Medicare and some other third party payors cover the cost of the HA preparations. Payors may fear that physicians, while needling their needy patients’ knees, are really tapping the payors’ liquid assets. With appropriate indications and for, and in selected patients, one can stick it to them with a clear conscience, backed up by convincing evidence that it is at least as effective as, and is better tolerated than, continuous use of an NSAID.


1. Balazs E, et al. Semin Arthritis Rheum 1981; 11:141-143.