Longevity Digest - July 19-26, 2026: Lithium, ER-100, and the limits of self-measurement

The week in one screen

Peter Attia's most consequential new result is a negative one: in an 80-person, two-year trial, low-dose lithium did not meet six prespecified primary endpoints in older adults with mild cognitive impairment. David Sinclair's new Lifespan episode put the Phase 1 ER-100 trial at the center of the age-reversal story, but the public material still reports a trial launch and preclinical results, not human efficacy. Blueprint published a skin protocol and a heavy-metals guide; both are more useful as measurement and exposure-reduction advice than as proof of whole-body age reversal. Rhonda Patrick's new episode with Derya Unutmaz is a forecast about AI and medicine, while her practical food post makes a narrower smoothie recommendation.

Coverage window: July 19-26, 2026, Pacific Time.

Peter Attia: a null lithium trial, and a sensible pollution filter

Attia's July 20 AMA, "Environmental pollution and longevity," covers air pollution, noise, light, and electromagnetic fields. The practical frame is an 80/20 one: reduce the exposures with the clearest health relevance, start with indoor air and the local air-quality index, protect sleep and hearing, and do not turn uncertain EMF claims into an expensive avoidance project. 1

That is a useful priority order because pollution is not one exposure with one effect. PM2.5 and other air pollutants have different evidence from nighttime light, traffic noise, or radiofrequency fields. A reader who wants one low-regret move can check outdoor AQI and measure or filter indoor air on bad days. The AMA does not justify trying to create a completely exposure-free life.

His July 25 article on lithium and Alzheimer's disease examines the LATTICE randomized, placebo-controlled trial. The study enrolled 80 adults aged 60 and older with mild cognitive impairment, followed them for two years, and tested six co-primary outcomes: verbal memory, visual memory, a global cognitive composite, hippocampal volume, cortical gray-matter volume, and plasma BDNF. The participants received an average of about 195 mg of lithium carbonate per day, producing an average serum level of about 0.17 mEq/L. The trial did not reach statistical significance on the six prespecified primary endpoints. 2

There were signals, but they are not a prescription. Verbal memory declined more slowly in the lithium group than in the placebo group, and the effect appeared larger among participants with beta-amyloid-positive disease. Attia argues that a low dose, a mixed MCI population, and lithium carbonate may have diluted a disease-specific effect. He points to plasma p-tau217 screening, dose and target-engagement studies, earlier intervention, and a direct comparison with lithium orotate as possible next steps.

The safety boundary is the part to keep. Lithium orotate is sold as an over-the-counter supplement, but its long-term safety is not established. The article describes adverse-event reports, including nausea and tremor after excessive intake, and says current evidence does not support broad lithium prescribing for cognitive protection or self-experimenting above ordinary dietary exposure. A near-significant subgroup result is a reason to design a better trial, not a reason to buy a bottle.

Attia also published a July 22 premium article, "Should you choose organic food?" The public page describes its scope, including pesticides, genetically modified crops, nutrient content, and antibiotic exposure, but the analysis is behind membership access. No specific conclusion from that article is included here. 3

David Sinclair: ER-100 is entering human testing, not proving age reversal

Sinclair's July 23 Lifespan episode, "Inside the World's First Age Reversal Trial," gives the clearest public update from him this week. The episode description says his team restored vision in non-human primates and that a Phase 1 clinical trial of ER-100 has launched to evaluate safety. It also covers eye aging, light, sleep position, alcohol, intraocular pressure, nutrition, and supplements. The important distinction is between a Phase 1 safety study and evidence that a therapy reverses aging or improves vision in humans. 4

The public release contains no human efficacy result and no healthy-adult anti-aging protocol. ER-100 belongs on a clinical-trial watch list. It does not support taking reprogramming compounds outside a trial, and it does not turn an eye-health discussion into a general longevity supplement recommendation.

On July 24, Sinclair shared a Science paper titled "Epigenetic and 3D genome reprogramming during the aging of the human hippocampus." In the accompanying post, he argued that aging may involve loss of epigenetic organization rather than irreversible DNA damage and pointed to mouse work in which reprogramming restored features of aged cells. That is a mechanistic and preclinical argument. The paper link and Sinclair's post do not supply a human treatment result, dose, or safety record. 5 6

This is a sharper translational signal than a supplement update, but it still leaves the reader with nothing safe to copy. The next useful evidence is not another claim that aging is reversible. It is a public human safety result, followed by a controlled efficacy result.

Bryan Johnson's Blueprint: skin metrics and exposure control

Blueprint published "Kate Tolo's Skin Protocol" on July 22. Tolo says a VISIA multispectral scan put her measured skin age at 21, based on eight surface or near-surface indicators: spots, pores, wrinkles, texture, UV spots, brown spots, red areas, and porphyrins. The same page says the scan cannot measure collagen, elasticity, bone structure, or all of facial aging. That limitation changes the claim from "reversed biological age" to "improved a defined set of scanner outputs." 7

The protocol lists sun avoidance, sleep, nutrition, exercise, topical products, red-light therapy, microneedling, IPL, lasers, and other procedures, alongside Blueprint products and supplements. It gives one clear dose: 40 mg of Accutane per week under physician supervision. It gives no dose for the listed vitamin C, copper, vitamin D, hyaluronic acid, collagen, or omega-3 products. Accutane is prescription isotretinoin, not a casual supplement, and the page itself directs readers toward medical supervision.

A face scan can be useful as a baseline if the same device, lighting, and method are used again. It cannot establish that any one product caused a change, and the page explicitly says its products should not be expected to lower biological age on their own. The low-regret part of this post is ordinary sun protection and treatment of diagnosed skin conditions. The high-risk part belongs in a dermatologist's office.

Blueprint's July 24 article, "How to Avoid Heavy Metals," takes a different route. It recommends checking household water lines and testing tap water, using a filter or reverse-osmosis system when appropriate, rinsing and draining rice, rotating rice with oats, quinoa, or barley, choosing lower-mercury fish such as salmon, trout, or cod, and looking for third-party certificates of analysis for foods and supplements. It also gives pipe-flushing instructions that vary by the type of plumbing. 8

Those steps are exposure-reduction measures, not detoxification claims. A filter may reduce some contaminants without removing all of them, and a certificate of analysis is only useful if it covers the relevant batch and contaminant panel. The article's warning that some metals can remain in the body for decades is a reason to avoid unnecessary exposure, not a reason to start an unvalidated chelation or supplement regimen.

Rhonda Patrick: AI forecasts, then one small food decision

Patrick's July 22 episode with immunologist and aging researcher Derya Unutmaz is a 2-hour, 38-minute discussion of AI, cancer, aging, cellular reprogramming, and personal health data. Unutmaz argues that AI could shorten parts of drug discovery and trial design, while Patrick's show notes describe a digital twin as a time-updated model combining health data, biomarkers, behavior, and treatment responses. The episode also makes the needed correction: a consumer system that can reliably simulate disease or treatment response does not yet exist. 9 10

The practical version is much less dramatic than the headline. A dated, structured record of symptoms, labs, medications, sleep, exercise, and interventions can help a person or clinician see trends. It is not a validated digital twin, and an AI summary cannot replace diagnosis or a clinical trial. The episode's claim that the next decade could add decades to lifespan is a forecast, not an outcome estimate supported by a human intervention study.

Patrick also posted a narrow smoothie choice. She said she swaps the banana in a blueberry-kale smoothie for avocado, then explained that avocado also contains polyphenol oxidase but at nearly 130 times lower activity than banana. That is a biochemical rationale for an ingredient swap, not evidence that the swap improves lifespan or even a clinical endpoint. 11 12

The verified Patrick items in this window contain no new supplement dose recommendation. That is useful in itself: the week's most actionable item from her was a food preparation choice, while the larger longevity claims remained speculative.

Where the experts converge, and where they do not

There is no clean same-intervention agreement this week. Attia analyzed lithium as a possible disease-specific intervention but treated the trial as negative and the supplement version as unsafe to generalize. Sinclair discussed epigenetic restoration and ER-100, but neither source supplies a consumer dose. Patrick's interview touches cellular reprogramming through the Yamanaka-factor story, yet it offers no human protocol. The topic overlaps; the evidence does not converge.

The closer agreement is methodological:

Topic	What the week's material supports	What it does not support
Low-dose lithium	A larger, biomarker-selected trial may be justified. 2	Self-prescribing lithium orotate for cognitive protection
ER-100 and reprogramming	Following Phase 1 safety data and the underlying preclinical work. 4	A proven age-reversal treatment for healthy adults
Skin age	Repeated, standardized scans can track selected skin features. 7	Treating eight scanner outputs as whole-body biological age
AI health tools	Keeping a dated personal record can improve organization and questions for a clinician. 9	A consumer digital twin that reliably predicts treatment response

That shared caution is more defensible than a shared protocol. In each case, the useful next step is to improve measurement or reduce a concrete exposure while waiting for evidence that can separate correlation, mechanism, and clinical benefit.

What is reasonable to carry forward

Use AQI and indoor-air measurements to decide when filtration or ventilation matters. Do not spend first on speculative EMF products.
Treat the LATTICE lithium result as a research question. Do not start lithium orotate, lithium carbonate, or a cognitive-protection protocol without a clinician.
If using a skin scan, standardize the follow-up scan and keep the claim limited to the features the device measures. Prescription isotretinoin and cosmetic procedures require medical oversight.
Lower avoidable heavy-metal exposure with water testing, appropriate filtration, varied grains, lower-mercury fish, and batch-specific testing documentation.
Keep a dated health record if an AI tool helps organize it, but do not call the record a validated digital twin or use it to replace clinical care.
Patrick's banana-to-avocado swap is a low-stakes food experiment for people who want to try it. The 130x PPO comparison is not a longevity outcome.