The Therapeutic Effect of Intra-articular Normal Saline Injections for Knee Osteoarthritis: A Meta-analysis of Evidence Level 1 Studies

Abstract

Background: Intra-articular normal saline (IA-NS) injections have been utilized as a placebo in a number of randomized controlled trials pertaining to the management of knee osteoarthritis (OA); however, it is believed that these “placebo” injections may have a therapeutic effect that has not been quantified in the literature. Purpose: To (1) quantify the effect of IA-NS injections on patient-reported outcomes (PROs) and (2) compare postinjection PROs to established minimal clinically important difference (MCID) criteria to demonstrate a potential therapeutic effect. Study Design: Meta-analysis. Methods: A review was conducted to identify all randomized, placebo-controlled trials on injection therapy for knee OA between 2006 and 2016. Patient demographics and PROs before the injection and at 3 and 6 months after the injection were collected for patients in the IA-NS injection group in each study. A random-effects model was used to compare preinjection scores and scores at each postinjection time point in a pairwise fashion. Results: In total, there were 14 placebo cohorts in 13 studies that were analyzed after meeting inclusion criteria for this meta-analysis. This included 1076 patients (Kellgren-Lawrence grade 1-4), with a weighted mean age of 62.53 years and mean body mass index of 28.67 kg/m2. There was only sufficient information to perform analyses of visual analog scale (VAS) pain and Western Ontario and McMaster Universities Arthritis Index (WOMAC) total scores. At 3 months after the IA-NS placebo injection, there was a significant improvement in VAS pain scores (mean difference [MD], 12.10 [95% CI, 3.27 to 20.93]; P = .007), whereas improvement in the WOMAC total scores approached but did not reach statistical significance (MD, 19.75 [95% CI, –0.50 to 40.09]; P = .06). At 6 months, both VAS pain scores (MD, 16.62 [95% CI, 12.13-21.10]; P < .00001) and WOMAC total scores (MD, 11.34 [95% CI, 7.03-15.65]; P < .00001) were significantly improved in comparison to preinjection values. Furthermore, improvements in both the VAS pain and WOMAC total scores at 6 months were clinically significant (MCID, 1.37 and 9, respectively). Conclusion: The administration of an IA-NS placebo injection yields a statistically and clinically meaningful improvement in PROs up to 6 months after the injection in patients with knee OA. This observation supports the notion that the so-called placebo effect for IA-NS injections achieves a clinically meaningful response in patients with OA when provided during comparison studies to an active treatment group (ie, hyaluronic acid).