Background: The knee is symptomatically the most frequent affected joint in osteoarthritis and, in the Netherlands and other Western countries, is mainly managed by general practitioners (GPs). An intra-articular glucocorticoid injection is recommended in (inter) national guidelines for patients with knee osteoarthritis as an option for a flare of knee pain and/or for those who are not responding well to pain medication. An innovative approach that could replace the intra-articular injection is an intramuscular gluteal glucocorticoid injection. An intramuscular injection is easier to perform than an intra-articular injection with lesser risk of severe local adverse reactions. We hypothesize that intramuscular gluteal glucocorticoid injection is non-inferior in reducing knee pain compared to intra-articular glucocorticoid injection, with potentially a longer lasting effect than intra-articular injection. Methods/design: The study will be a pragmatic randomized controlled non-inferiority trial with two parallel groups. A total of 140 patients aged 45 years and older with knee osteoarthritis who contacted their general practitioner and have persistent knee pain (score ≥ 3 on 0-10 numerical rating scale; 0 = no knee pain) will be included. Patients will be randomly allocated (1:1) to an injection of 40 mg triamcinolone acetonide intra-articular in the knee joint or intramuscular in the ipsilateral ventrogluteal area. The effect of treatment will be evaluated by questionnaires at 2, 4, 8, 12, and 24 weeks after injection. The primary outcome is patients' reported severity of knee pain measured with the pain subscale of the Knee injury and Osteoarthritis Outcome Score 4 weeks after injection. Statistical analysis will be based on both the per-protocol and the intention-to-treat principle.(Continued on next page)