The efficacy of Curcuma Longa L. extract as an adjuvant therapy in primary knee osteoarthritis: a randomized control trial

Pinsornsak P, Niempoog S · Journal of the Medical Association of Thailand

Adding curcumin to diclofenac changed nothing between the arms: VAS was identical by repeated-measures analysis of variance (F=0.009, p=0.923) and all five KOOS subscales were non-significant, with only a numerical trend on pain and daily function. Both arms improved against their own baselines. The paper reads its daily-living subscale as favourable because the control arm's within-arm improvement did not reach significance while the curcumin arm's did at every timepoint, but the between-arm test on that subscale is null.

Population

Adults with primary knee osteoarthritis at a Thai university hospital, recruited October 2008 to October 2010; 44 randomised per arm and 75 analysed, 83% of them women

Intervention

Diclofenac 75 mg/day plus curcumin 1000 mg/day for 3 months, double blind against an identical-looking placebo

Comparator

Diclofenac 75 mg/day plus placebo

Limitations

Randomisation did not balance the groups: the curcumin arm started significantly worse on the KOOS subscales — 68.55 against 72.91 on symptoms, 67.00 against 76.93 on the next — at p=0.024 and p=0.022, which is what leaves it more room to improve within itself and makes the control arm's flat within-arm result easy to misread. 13 of 88 randomised patients were lost, leaving 75. Unformulated extract at 1000 mg/day may be an under-dose in absorption terms, and an add-on design on top of an effective NSAID has limited room to show a difference. Harms fell in the control arm: two patients with rising urea and creatinine and one with facial swelling, against one report of hair loss on curcumin.