This Week at Peak Human: Plasma Exchange's Alzheimer's Data, and the Harvard Happiness Study

A note from Dr. Goel

This week's issue is our regular site digest — everything new on peakhuman.ai over the past seven days. We added a new evidence page on therapeutic plasma exchange and Alzheimer's, published a deep dive on what 85 years of Harvard data say about relationships and health, updated our resveratrol and magnesium glycinate pages with new blood-pressure trial data, and broke down the hormone that makes weight loss so hard to sustain: ghrelin. As always — evidence first, hype never.

— Dr. Sanjeev Goel

What's new at Peak Human this week

Five updates from the past seven days — starting with the one I think matters most.

1. Therapeutic plasma exchange and Alzheimer's: a real signal

We added a new evidence page on therapeutic plasma exchange (TPE) for cognitive decline. In the AMBAR phase 2b/3 trial, 347 Alzheimer's patients who received TPE with albumin replacement showed slower cognitive and functional decline than those on sham treatment. A 2025 real-world study of 32 patients against 194 matched controls found MMSE score decline was roughly 45% slower, with markedly less deterioration in immediate and delayed recall. One caveat worth flagging: a separate 2025 trial found plasmapheresis without albumin or young-plasma replacement did not rejuvenate epigenetic age markers — the replacement fluid appears to matter as much as the procedure itself.

2. What 85 years of Harvard data say about happiness

We published a new deep dive on the Harvard Study of Adult Development, which has followed the same people since 1938. Relationship satisfaction at age 50 predicted physical health at 80 better than cholesterol did. A separate pooled analysis of 148 studies and over 308,000 people found that stronger social ties were associated with a 50% greater likelihood of survival. It's some of the best longevity evidence we have that isn't a pill or a protocol.

3. Resveratrol: real blood-pressure numbers, and a caveat on the "anti-aging" story

We updated our resveratrol page with a 2026 umbrella review of 45 systematic reviews using GRADE methodology. In people with type 2 diabetes, resveratrol lowered systolic blood pressure by roughly 8 mmHg and diastolic by about 3.5 mmHg versus placebo — a high-certainty finding — and modestly reduced waist circumference and total cholesterol. The honest caveat: a 2025 GRADE meta-analysis found no significant effect on human SIRT1, which undercuts the popular "caloric-restriction mimetic" framing.

4. Magnesium glycinate: blood pressure and sleep, quantified

New research block on our magnesium glycinate page. A 2025 meta-analysis of 38 randomized trials in 2,709 adults found an average systolic blood-pressure reduction of about 2.8 mmHg overall — rising to roughly 7.7 mmHg in people with hypertension and 6.0 mmHg in those who were magnesium-deficient, with no significant change in normotensive adults. A separate 2025 trial also found magnesium bisglycinate modestly improved insomnia severity scores over four weeks versus placebo.

5. Ghrelin: the hormone that fights you after you diet

New article on ghrelin, the only hormone shown to directly increase how much people eat — infusing it raised buffet intake by about 28% in controlled studies. It's also sensitive to sleep: two nights of short sleep can raise ghrelin by roughly 28%, and it can stay elevated for a year or more after a diet ends. If you've ever felt hungrier after losing weight, this is a big part of why.

My take: The plasma exchange data is the one I keep coming back to. A roughly 45% slower decline on a validated cognitive scale, in a real-world comparison against matched controls, is a meaningful signal for a therapy most people have only heard of in the context of dialysis units. It's not a cure, and the replacement fluid matters enormously — plain plasmapheresis without albumin replacement did not show the same benefit. If cognitive decline runs in your family, this is worth a conversation with a physician who knows the current evidence, not a DIY biohack.

Longevity news worth your minute

GLP-1 drugs may slow aging itself — in mice. In a new Nature study, semaglutide given late in life to aged female mice extended median lifespan by about 12% (742 to 834 days) and improved memory, motor coordination, and blood-sugar control beyond what calorie restriction alone achieved. It's an animal study only — whether GLP-1 drugs affect human aging trajectories remains unknown.

⚕️ Important disclaimer

This newsletter is for educational purposes only and is not medical advice. Compounds discussed may be investigational and not FDA-approved. Nothing here is a recommendation to start, stop, or change any treatment. Peptide and metabolic therapies carry real risks and should only be used under qualified medical supervision. Always consult your physician before making health decisions.