Objectives: The aim of this study was to critically appraise and evaluate effects of low- and high-dose curcuminoids on pain and functional improvement in patients with knee osteoarthritis (OA) and to compare adverse events (AEs) between curcuminoids and non-steroid anti-inflammatory drugs (NSAIDs). Methods: We systematically reviewed all randomized controlled trials (RCTs) on curcuminoids in knee osteoarthritis from the PubMed, Embase, Cochrane Library, AMED, Cinahl, ISI Web of Science, Chinese medical database, and Indian Scientific databases from inception to June 21, 2021. Results: We included eleven studies with a total of 1258 participants with primary knee OA. The meta-analysis results showed that curcuminoids were significantly more effective than comparators regarding visual analogue scale (VAS) and Western Ontario and McMaster Universities Arthritis Index (WOMAC) pain scores. However, no significant difference in pain relief or AEs between the high-dose (daily dose ≥1000 mg or total dose ≥42 gm) and low-dose (daily dose