PubMedThe Cochrane database of systematic reviews2026-11-18
Curcuma longa (turmeric) or its active ingredients for osteoarthritis.
Shirinsky Ivan I, Shirinsky Valery V, Filatova Katerina K, Yu Ze Yu ZY et al.
Osteoarthritis is a major cause of disability worldwide. Current treatments have limited efficacy. Curcuma longa (turmeric) contains curcuminoids with anti-inflammatory properties that may modulate osteoarthritis pathophysiology. However, the clinical benefits and safety of Curcuma longa in osteoarthritis remain uncertain.
To assess the benefits and harms of Curcuma longa extracts or сurcuminoids for osteoarthritis.
We used CENTRAL, MEDLINE, five other databases, and two trials registers, together with reference checking, to identify the studies included in the review. The latest search date was 7 November 2025. We applied no restrictions on language, publication date, or publication status.
We included randomised controlled trials (RCTs) comparing Curcuma longa extracts or curcuminoids to placebo, in people with osteoarthritis in any joint. Other comparators included no treatment, non-steroidal anti-inflammatory drugs (NSAIDs), non-NSAID analgesics, or other active comparators. We focused on oral formulations of Curcuma longa, with or without bioavailability enhancers, and excluded combination products containing other active herbal compounds.
Critical outcomes included pain, function, quality of life, participant-reported global assessment of success, adverse events, withdrawals due to adverse events, and serious adverse events. The primary time point was up to three months.
We used Cochrane's RoB 2 tool to assess bias in the RCTs.
We used a random-effects model for meta-analysis. We assessed the trustworthiness of all trials eligible for inclusion using the Cochrane Pregnancy and Childbirth Trustworthiness Screening Tool.
We included nine RCTs involving 1245 participants with knee osteoarthritis. The mean age of participants ranged from 52 years to 62 years; 67% of participants were female. All trials were short (up to 90 days) and investigated different Curcuma longa formulations. One trial evaluated a bioavailability-enhanced Curcuma longa formulation, while others did not report the use of bio-enhancers. Two trials used Curcuma longa extract with essential oils. Four trials used standardised or rhizome extracts of Curcuma longa, with curcuminoid content varying from 23% to 88%. Four trials compared Curcuma longa with placebo (379 participants), four with NSAIDs (673 participants), and one with paracetamol (193 participants). Here, we limit reporting to the primary comparison, Curcuma longa versus placebo. We judged three (75%) placebo-controlled trials to be at low risk of bias for pain, function, and quality of life outcomes. Two trials (50%) were at low risk of bias for safety outcomes.
Low-certainty evidence downgraded for indirectness and publication bias showed that, compared with placebo, Curcuma longa may result in little to no difference in pain, function, or quality of life, and may improve participant-reported global assessment of success. We are very uncertain about the effect of Curcuma longa on adverse events, withdrawals due to adverse events, and serious adverse events due to very low-certainty evidence, downgraded for publication bias, imprecision, and indirectness. By three months, the mean change in pain (0 to 10 scale, 0 = no pain) was -1.57 points with placebo and 0.72 points lower (95% confidence interval (CI) 1.39 points lower to 0.06 points lower) with Curcuma longa (I² = 0%; 4 studies, 349 participants). The effect did not reach clinical significance of 1.5 points. By three months, the mean change in function (0 to 100, 0 = best function) was -5.42 points with placebo and 3.5 points lower with Curcuma longa (95% CI 7.29 points lower to 0.29 points higher; I² = 0%; 4 studies, 348 participants). By three months, the mean change in quality of life (0 to 100 scale, 100 = best quality of life) was 9.83 points with placebo and 2.19 points higher (95% CI 1.56 points lower to 5.94 points higher) with Curcuma longa (I² = 0%; 3 studies, 291 participants). By three months, 107 of 1000 participants receiving placebo reported treatment success, compared with 191 of 1000 receiving Curcuma longa (risk ratio (RR) 1.78, 95% CI 1.19 to 2.67; 2 studies, 170 participants). At the end of follow-up, 224 of 1000 participants in the placebo group reported adverse events compared with 226 of 1000 in the Curcuma longa group (RR 1.01, 95% CI 0.68 to 1.51; I² = 0%; 4 studies, 379 participants). At the end of follow-up, 46 of 1000 participants in the placebo group withdrew due to adverse events compared with 66 of 1000 in the Curcuma longa group (RR 1.45, 95% CI 0.61 to 3.43; I² = 0%; 3 studies, 321 participants). Three trials reported no serious adverse events, while one trial reported two serious adverse events in the placebo group that were unrelated to the study.
Compared with placebo, Curcuma longa may result in little to no improvement in pain, function, or quality of life, and may improve participant-reported global assessment of success in people with knee osteoarthritis. We are very uncertain about the effect of Curcuma longa on adverse events, withdrawals due to adverse events, and serious adverse events. Future studies should last longer than three months; the maximum duration of the included studies was three months. The current evidence is limited to knee osteoarthritis, and future studies should include people with hand and hip osteoarthritis.
No external funding was received for this work.
Protocol (2021): 10.1002/14651858.CD014683.