Original Article: Top POEMs of 2019 Consistent with the Principles of the Choosing Wisely Campaign

Issue Date: December 1, 2020

See additional reader comments at: https://www.aafp.org/afp/2020/1201/p673.html

To the Editor: I was confused by the conclusion of Drs. Grad and Ebell that platelet-rich plasma should not be recommended for knee osteoarthritis. The study summarized in their article reported minimal function and subjective benefit after five years; however, that overlooks the participants' favorable and well-tolerated response to platelet-rich plasma for up to two years.¹ This study was a prospective, double-blind trial, with platelet-rich plasma vials taped to conceal them from the hyaluronic acid comparison group and three injections administered in a weekly series.

Having an additional modality to treat knee osteoarthritis with injections every six to 12 months is important. At my institution, repeat hyaluronic acid or cortisone injections are offered at six months if there is functional improvement after the first injection. With up to an 11% excessive inflammatory reaction rate to hyaluronic acid, platelet-rich plasma offers a better-tolerated alternative to help decrease or postpone the need for knee arthroplasty.²

Drs. Grad and Ebell noted hyaluronic acid to be similar to placebo in effectiveness. However, in the trial, both the hyaluronic acid and platelet-rich plasma groups had subjective and functional improvement scores significantly above baseline for up to two years, suggesting that both treatments were effective.

In Reply: We thank Dr. Kersch for his comments on this topic. Knee osteoarthritis is a chronic painful condition, and studies typically use subjective patient-reported outcomes. A placebo effect is more likely when there is a subjective outcome, such as pain, and an impressive intervention, such as a physician with a large needle. Therefore, we must focus our attention on adequately controlled studies with a true sham intervention and blinded outcome assessment.¹

Many randomized trials of hyaluronic acid injections failed to compare the active drug to a sham or placebo injection or to mask outcome assessors. A systematic review of 71 randomized trials also found clear evidence of publication bias, meaning that trials finding no benefit or finding harm were not published. Benefits were found in 58 studies that did not have adequate allocation concealment, but benefits were not clinically significant in 13 studies with adequate allocation concealment. Similarly, no clinically important benefit was found when outcomes were blindly assessed compared with when they were not. There was also no benefit in larger studies and unpublished studies. To summarize, in the best quality and largest studies, little or no benefit was found; many studies went unreported, and harms were significantly greater in the viscosupplementation groups.²

POEMs (Patient-Oriented Evidence that Matters) have reported on many studies of joint injections for patients with knee arthritis (e.g., hyaluronic acid, platelet-rich plasma, corticosteroids, ozone therapy, prolotherapy). A search of Essential Evidence Plus found a systematic review of six randomized trials comparing platelet-rich plasma injections with hyaluronic acid (five studies) or saline (one study).³ The researchers found no significant difference between groups in pain score or patient satisfaction and more adverse events with platelet-rich plasma (8.4% vs. 3.8%; P = .002). We cannot recommend this treatment to our patients until we have adequately powered studies that compare platelet-rich plasma with saline or placebo injection and adequate masking of patients and investigators. We conclude that the observed score improvements found with platelet-rich plasma injections most likely reflect a placebo effect.

Editor's Note: Dr. Ebell is deputy editor for evidence-based medicine for AFP.