Quality of prospective controlled randomized trials regarding platelet-rich plasma injection for osteoarthritis of the knee

Abstract

BackgroundRandomized controlled trials (RCTs) are the gold standard for study design and are considered to have the highest level of evidence. The purpose of this study is to evaluate quality of evidence of prospective RCTs that investigated the clinical efficacy of platelet-rich plasma (PRP) treatments for knee osteoarthritis. We hypothesized that a high level of clinical evidence does not correlate with a high-quality scientific study. Material and methodsA systematic literature search to identified RCTs that evaluated the efficacy of PRP treatments for knee osteoarthritis. Inclusion criteria included studies that contained key terms “Platelet-rich plasma,” “PRP”, “knee”, and “osteoarthritis”. Exclusion criteria excluded studies that were not prospective RCTs. Ten RCTs were evaluated by four independent reviewers. The studies were assessed according to the Oxford Levels of Evidence, a modified Coleman Methodology Score, and the revised Consolidated Standards of Reporting Trials (CONSORT) score. ResultsThree of the four Observers rated all 10 studies as Level I, while one Observer rated 80% of the studies as Level I. Using the Coleman Methodology Score, the average score for all studies was 58.0 out of 100 points, designating the studies as “fair.” Areas of deficiency included clinical effect measurement and intent-to-treat analysis. Using the CONSORT checklist, the average score was 16.9 out of 22 points, designating the studies as “good”. Areas of deficiency included inadequate implementation of randomization and interpretation of results. DiscussionClinicians should critically evaluate research studies regardless of study design. A sophisticated study design and high level of evidence designation does not guarantee quality research. We determined that RCTs for PRP treatment of knee osteoarthritis were not as robust in quality despite their Level I Oxford Level of Evidence rating.