Spermidine for Aging
Despite a plausible mechanism (autophagy induction), spermidine has not been shown to slow human aging: the largest, highest-quality trial missed its primary endpoint, and no human studies test lifespan or healthspan. The grade is Weak (Tier C).
Why this grade7-layer evidence engine
The human evidence rests on four small RCTs and points more negative than positive. The largest and best-conducted trial (SmartAge, PMID 35616942; n=100, 52 weeks) found no benefit on its primary memory endpoint (treatment effect -0.03, p=0.47); only an exploratory inflammatory marker moved in a high-compliance subgroup. A smaller 12-week pilot (PMID 30388439; n=30) reported a memory gain (Cohen's d=0.79), but it is underpowered and was contradicted by the larger study. A safety trial (PMID 29315079) showed good tolerability but did not test efficacy.
A key mechanistic problem undercuts the whole case: a pharmacokinetic RCT (PMID 37111071; n=12) found that 15 mg/day of oral spermidine did not raise plasma or salivary spermidine, challenging the assumption that supplements reach the levels animal studies rely on. Critically, no human RCT has ever measured lifespan or healthspan; longevity claims rest only on dietary-intake epidemiology.
Regulators and clinicians reinforce the caution. The US FDA 'ceased to evaluate' the GRAS notice at the maker's request, EFSA's 6 mg/day limit is a food-safety ceiling (wheat-germ extract only) with no approved health claims, and Health Canada bars its use in natural health products without more safety data. The American Federation for Aging Research is blunt: 'There are no proven anti-aging interventions,' and Mayo, Cleveland Clinic and Harvard Health say nothing at all. Together this supports a Weak (Tier C) grade, not an endorsement.
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 = 0.485
- 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 — Tier 條件達標，未降階
- detect_disputes — 偵測到 1 個 hard + 0 個 soft dispute
- decide_status — 依 tier + dispute 結果決定 status