CDP-Choline (Citicoline) for Cognitive Function
CDP-Choline (citicoline) shows only preliminary, mixed support for cognitive function, with human evidence that is thin, short-term, and dominated by small or industry-funded trials. It cannot yet be called a proven cognitive aid for healthy people.
Why this grade7-layer evidence engine
The C (Weak Evidence) grade rests on just three human studies. A 12-week RCT in healthy older adults (n=100, 500 mg/day; PMID 33978188) found significantly better episodic and composite memory versus placebo, but it was only moderate quality and was funded by Cognizin's maker, Kyowa Hakko Bio, a clear conflict of interest. A smaller RCT of a citicoline-caffeine beverage (n=60; PMID 25046515) reported faster reaction times and better attention, yet was low quality, mixed-funded, and reported no p-values in its abstract.
The single meta-analysis (PMID 36678257) pooled seven trials with uniformly positive results (SMD 0.56 to 1.57), but its own authors judged the overall study quality as poor, and most participants had MCI, Alzheimer's, or post-stroke dementia rather than normal cognition. Extrapolating those clinical findings to healthy users is the scope conflation that pushed this claim down from a borderline B to C.
Regulators and major clinics do not endorse a cognitive benefit. The US FDA has approved no memory or cognition health claim for citicoline, and EFSA formally rejected a memory claim for insufficient evidence. NIH ODS, Mayo Clinic, Cleveland Clinic, and Harvard Health take no supportive position. Small, short, partly conflicted trials plus regulatory silence-to-rejection justify a Weak Evidence tier, published with a warning.
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 = 0.58
- 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 — | B→C 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