eBioMedicine · Jun 2026 · New
Omega-3 (brain) demoted to LORE: DHA reached the brain and cognition didn't move.
The definitive CNS target-engagement RCT (Yassine et al., eBioMedicine, Jun 2026): n=365, 2 g/day DHA, 24 months. CSF DHA/AA ratio rose 0.19 (p<0.0001), independent of APOE ε4 — the supplement clearly reaches the brain. But no treatment differences in brain volumes or cognitive performance over two years. This is exactly the trial you'd design to settle the question, and it settles it. The cardiovascular/triglyceride claim is separate and unchanged; the brain claim is now LORE.
Lilly · May & Jul 2026 · New
Retatrutide readouts: −28.3% weight over 80 weeks, with a directional MACE signal.
Lilly's triple agonist (GLP-1/GIP/glucagon) reported all three TRIUMPH Phase 3 trials in 2026. TRIUMPH-1 (n=2,339, 80 weeks): weight change −19.0% / −25.9% / −28.3% (4/9/12 mg) vs −2.2% placebo; 45% of the 12-mg arm hit ≥30% loss. TRIUMPH-3 five-component MACE HR 0.82 (0.55–1.22) — directional, underpowered. Board adds retatrutide as REAL for metabolic, MIXED for hard longevity endpoints until a proper outcomes trial reports. Same class caveats: pair with resistance training and ≥1.2–1.6 g/kg protein.
Aging · Apr 2025 · New
PEARL: rapamycin missed its healthy-aging primary endpoint at p=0.942.
PEARL — the only dedicated healthy-aging rapamycin RCT — published in Aging (Albany NY): 48 weeks, weekly 5 mg or 10 mg vs placebo. Primary endpoint (visceral adiposity) not met, p=0.942. Secondary: lean tissue mass ↑ (ηp²=0.202, p=0.013), pain ↓ in women on 10 mg. Interesting biology; no hard-outcome case for healthy adults. Verdict moves from MIXED to LORE for the healthspan indication.
Three 2025 RCTs · New
Nicotinamide riboside demoted to LORE: NAD⁺ doubled, everything else was null.
Three 2025 RCTs on the same molecule tell the same story: NAD⁺ raising is confirmed, downstream endpoints are not. aMCI phase-II (Alz Dement, Dec 2025, n=42, 12 wk, 1 g/d) — blood NAD⁺ 23→48 µM (p<0.001); cognition not improved. NR + exercise for hypertension (GeroScience, Dec 2025, n=54): daytime systolic BP +5.2 vs −2.7 mmHg — NR was worse. SCD/MCI crossover (Alz Dement TRCI, Jan 2025) — no RBANS change; only pTau-217 moved. Textbook "don't fool yourself."
BJSM · Jun 2026 · New
Resistance training joins evidence-5: 147k people, 30 years, HR 0.55 combined.
The largest pooled dose-response analysis to date (BJSM, Jun 2026): n=147,374, up to 30-year follow-up, 35,798 deaths. 90–119 min/week of resistance training → all-cause mortality HR 0.87 (0.81–0.95); cardiovascular mortality HR 0.81; neurological-disease mortality HR 0.73. Combined with aerobic activity: HR 0.55 (0.50–0.60). The sweet spot is 90–120 min/week and the biggest gain is stacking with aerobic work. Now the highest-evidence non-drug lever on the board — and even more essential now that GLP-1s are pulling weight down at the cost of lean mass.
IJBNPA · Jan 2026 · New
VILPA: 1–2 minute exercise "snacks" 5–8×/day map to HR 0.46 for mortality.
Accelerometry cohort (IJBNPA, Jan 2026, n=3,293 US adults, 6.7-year follow-up, 290 deaths): 5.3 vigorous-intermittent-lifestyle-activity bouts/day → all-cause mortality HR 0.56 (0.39–0.82); ~8 bouts/day → HR 0.46 (0.28–0.77). Backed by an 11-RCT meta on VO₂max and 60-second sit-to-stand. Confounding by baseline capacity is plausible — but for people who won't do structured training, this is the highest-yield entry we've ever added. New on the board in v1.4.
Cell Rep Med + Neurology · 2026 · New
Hearing aids upgrade: three converging datasets, RR 0.67 for 7-year dementia.
Pooled 7-cohort analysis (Cell Rep Med, May 2026, n=61,089, 6.5y, 8,911 events): hearing-aid use HR 0.91; effective hearing improvement HR 0.86; poor correction HR 0.98 (i.e., wearing them matters). Target-trial emulation (Neurology, Feb 2026, 7-year): dementia risk 5.0% vs 7.5% (RR 0.67, 0.37–0.97), inverse dose-response by use frequency. ACHIEVE secondary analysis: 61.6% slower decline in the highest-risk quartile. Ownership without use doesn't count.
Clin Nutr + BMJ · 2026 · New
Vitamin D is the archetype of a conditional REAL: works only if you're deficient.
Target-trial emulation of RCTs (Clin Nutr, Mar 2026, n=237,502 and 185,809, 5.3–5.7y): overall all-cause mortality NULL (HR 0.97 and 1.02); in insufficiency HR 0.85 and 0.81; in deficiency HR 0.79 and 0.75. Fractures (BMJ, May 2026, 69 trials / 153,902 participants): vitamin D alone RR 1.00, high certainty. Measure 25(OH)D, correct if low, do not megadose a sufficient level. B-vitamins behave the same way — real but small, and only in the elevated-homocysteine framing.
CJASN · Feb 2026 · Updated
Semaglutide splits by endpoint: REAL for cardiometabolic + kidney mortality, LORE for dementia.
FLOW severity analysis (n=3,533 T2D+CKD, 3.4y): all-cause death HR 0.80 (0.67–0.95); UACR ≥2000 mg/g subgroup HR 0.47 (0.31–0.70). SELECT hospitalization ratio 0.90 (p<.001). But EVOKE / evoke+: oral semaglutide 14 mg did not slow cognitive or functional decline in early Alzheimer's. File the AD claim under "does not work for."
JAMA · Aug 2025 · Anchor
US POINTER: structured lifestyle beats self-guided — statistically robust, clinically small.
US POINTER (n=2,111, 2 years): the structured MIND-style diet + supervised exercise + cognitive training + cardiometabolic monitoring bundle improved global cognition 0.243 vs 0.213 SD/year self-guided — between-group Δ 0.029 SD/year (95% CI 0.008–0.050, p=0.008). Both arms improved. The active ingredient is the bundle. MIND diet upgraded to evidence-4; effect size honestly labeled Small.
BMJ + Lancet HL · 2026 · New
Menopausal HRT: small mortality signal, dementia claim now formally null.
Register-based cohort (BMJ, Feb 2026, n=876,805, 14.3y): adjusted death HR 0.96 (0.93–0.98); 3–4.9 y of use HR 0.90 (0.84–0.95). Dementia (Lancet Healthy Longev, Dec 2025, n=1,016,055 across 10 studies): no association between MHT and MCI/dementia — subgroup analyses by timing, duration, formulation also null. REAL for symptom control and possibly small mortality benefit; LORE for the "HRT prevents dementia" narrative.
Centenarian resilience · 2026
Cognitively healthy centenarians are genetically protected against Alzheimer's — not just lucky.
A five-cohort analysis shows centenarians across the US, Europe, and Asia carry a higher Alzheimer's polygenic protective score than controls — beyond APOE. Protection increases with age: supercentenarians (≥110) carry the most protective alleles. Complementary work on centenarian microglia finds they overexpress neprilysin, clearing amyloid 1.5–2× faster than young adult microglia. Implication: reaching 100 cognitively intact is a resilience phenotype, not an absence-of-risk one.