NMN reliably raises blood NAD+ (or NAD++NADH) at 250–2,000 mg/day. That is a surrogate. The only gold-standard metabolic win is Yoshino 2021: +25% muscle insulin-stimulated glucose disposal in 13 overweight prediabetic postmenopausal women, with no change in fasting glucose, HbA1c, liver or fat insulin sensitivity, lipids, BP, or muscle NAD+. HOMA-IR trials and the 2025 Zhang meta-analysis of 12 RCTs (n=513) do not show a glycaemic or lipid benefit. Liao improved ventilatory threshold, not VO2max, in already-training runners. Muscle-mass trials are null; function signals (gait, 6-minute walk, afternoon 5-STS) are small, short, and often industry-funded. Pencina’s −19 mg/dL LDL and −7 mmHg diastolic BP appeared only at 2 g/day of a pharmaceutical polymorph for 28 days. Sleep evidence is mixed. **Human longevity evidence is absent.**
**Confidence: weak** for any clinical outcome a person would notice or that would change a hard endpoint; **moderate** for blood-NAD raising and short-term tolerability at 250–1,250 mg/day.
Confidence: moderate