Vitamin A for Vision / Eye Health
Vitamin A reliably reverses night blindness and dry-eye damage (xerophthalmia) caused by genuine deficiency, but for well-nourished adults it does not improve general vision or prevent age-related eye disease, and high doses or beta-carotene carry real harms.
Why this grade7-layer evidence engine
The evidence splits cleanly by population. In people who are actually deficient, it is strong: Sommer's 1983 Indonesian cohort (PMID 6136744) linked mild xerophthalmia to sharply higher child mortality and showed night blindness and Bitot's spots reverse after repletion, and the Cochrane review of 47 trials (PMID 28282701, ~1.2 million children) found supplementation cut all-cause mortality about 12% (RR 0.88) and night blindness by roughly two-thirds (RR ~0.32). A maternal trial in deficient Bangladeshi women (PMID 10024613) showed the same reversal of night blindness.
Regulators and clinics confirm only this narrow benefit. WHO strongly recommends supplementation for children 6–59 months solely where deficiency is a public-health problem, while NIH ODS and the Cleveland Clinic describe night blindness as an early deficiency sign that vitamin A corrects. None endorse routine supplements for well-nourished adults, and EFSA's authorized claim covers only "maintenance of normal vision," not improvement.
The grade is held at preliminary because the benefit does not generalize and there are documented harms. AREDS2 (PMID 23644932) dropped beta-carotene, a vitamin A precursor, from its eye formula after it raised lung-cancer risk in former smokers, echoing CARET (PMID 8602180; lung-cancer RR 1.28, mortality RR 1.17 in smokers); the American Academy of Ophthalmology cites exactly this. The UK NHS also warns against retinol supplements in pregnancy because of birth-defect risk. For most replete adults seeking sharper vision, vitamin A is unproven and unsafe at high doses.
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 = 0.712
- 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 — A 級條件未達 (需 E1-E3 + L5≥2 supportive + L4 無 against；實際 E1 / L5=1 / L4_against=0)
- detect_disputes — 偵測到 0 個 hard + 0 個 soft dispute
- decide_status — 依 tier + dispute 結果決定 status