CBD for Arthritis
For arthritis, the best controlled trials show oral CBD works no better than placebo for joint pain, and at high doses it adds side effects rather than relief. It is graded counter-evidence: the marketed benefit has not held up in rigorous testing.
Why this grade7-layer evidence engine
The grade reflects two adequately powered, double-blind randomized trials that both came up empty. A knee osteoarthritis trial adding 600 mg/day oral CBD to paracetamol found no extra pain relief over placebo (PMID 38033459, WOMAC difference negligible, p=0.80), while a hand osteoarthritis and psoriatic arthritis trial of 20-30 mg/day was likewise null (PMID 34510141, between-group difference 0.23 mm, p=0.96) with no benefit for sleep, anxiety, or depression.
The only positive randomized result was a tiny 18-person topical CBD crossover study of low strength (PMID 35637038), and the often-cited large benefit comes from an uncontrolled survey vulnerable to selection and recall bias (PMID 35999581). Two strong trials failing against placebo outweigh one small positive study and self-reported data, which is why the engine treats oral CBD for arthritis pain as counter-evidence rather than merely unproven.
Authoritative bodies reinforce this caution rather than contradict it. The US FDA has concluded CBD does not fit existing food or supplement frameworks; the UK Food Standards Agency advises healthy adults cap intake near 10 mg/day and warns vulnerable groups, including people on medication, to avoid it; and Mayo Clinic notes the widely promoted pain relief is unconfirmed and that CBD can cause side effects and drug interactions. The 600 mg trial also flagged more adverse events and liver-enzyme elevations, underscoring real risk without demonstrated arthritis benefit.
Scoring transparency
All scores computed by a 7-layer evidence engine — fully auditable▸View the full decision path (audit trail)
- compute_raw_score — 加權公式: L2×0.30 + L3×0.25 + L5×0.25 + L11×0.10 + L1×0.10 = 0.44
- tier_from_score — 依分數區間映射至 tier letter
- apply_hec_rules — 無高階證據可裁決
- tier_strict_requirement_check — | C→D 因 scope.conflation_risk=true 且 L11 獨評較低 (B7-2 tier cap)
- detect_disputes — 偵測到 0 個 hard + 0 個 soft dispute
- decide_status — 依 tier + dispute 結果決定 status