Phenotypes Identification and Physiotherapy Decision-Making Algorithm for the Conservative Non-Pharmacological Treatment of Knee Osteoarthritis: A Scoping Review

Purpose: Knee osteoarthritis phenotype research is an emerging area of study that has not yet matured or led to any major discoveries with significant implications in the clinical setting. We aimed to systematically summaries knee osteoarthritis phenotypes and recommended non-pharmacologic interventions, and to develop a clinically applicable algorithm that categorizes patients with knee osteoarthritis based on phenotypes, thereby making management easier and more effective. Methods: We conducted a main systematic literature search in publication databases (PubMed and Cochrane). Subsequently, a secondary search in PubMed for recommendations, guidelines, or consensus regarding the non-pharmacologic treatment of knee osteoarthritis was conducted. Our findings were then used to develop an algorithm for the effective, conservative, non- pharmacological treatment of knee osteoarthritis. Results: Five phenotype categories, including neuropathic, local, psychological, high body mass index, and comorbidities were identified. Secondly, we summarized the most recommended interventions, including strengthening, aerobics, education, mind-body, neuromuscular training, flexibility, aquatic exercise, psychological, and walking aids. We subsequently created an algorithm with sixteen phenotypes. Conclusions: We systematically reviewed all eligible studies that aimed to identify knee osteoarthritis phenotypes, and summarized all non-pharmacological interventions mentioned in the relevant literature. Further, we opine that each rehabilitation programme should be tailored to each patient's particularities or phenotype at the least. Further, we provide a unique algorithm based on critical reasoning to assist decision-making in conservative non-pharmacological treatment of knee osteoarthritis pain. In the detailed and well-structured treatment for knee osteoarthritis patients, phenotype identification is essential. We suggest the adoption of this algorithm into daily clinical practice.