Longevity Digest — August 2–9, 2026: Protein in CKD, sauna evidence, and the protocol stopping points

The week in one screen

From August 2 through August 9, 2026, the most useful updates were about where a protocol stops being evidence-based. Peter Attia questioned routine protein restriction in chronic kidney disease (CKD), Rhonda Patrick separated Finnish sauna evidence from infrared devices, and FoundMyFitness summarized a photobiomodulation meta-analysis without offering a universal dose. Bryan Johnson shared a striking cancer claim that he explicitly labeled preclinical and observational. David Sinclair posted two brief signals about epigenetic restoration and AI-assisted biology, but no new supplement dose, human endpoint, or consumer protocol.

The practical pattern is simple: a measurement or mechanism can be interesting without being a treatment instruction. That distinction matters more this week than any new item in a supplement stack.

Coverage window: August 2–9, 2026, Pacific Time. Dates below refer to the source's displayed publication or update date, or to the local Pacific date of a social post.

Peter Attia: protein restriction is not a default CKD prescription

Attia's most actionable item was an August 8 article written with Taylor Yeater and Nat Pedley. Its question is narrow: does restricting protein still add enough benefit for people with CKD to justify the nutritional cost, now that kidney-protective drugs are available? 1

The article separates three situations that are often collapsed into one. For people without CKD, Attia states a general target of roughly 1.6–2.2 g/kg/day. For adults with CKD stages 3–5, the 2024 KDIGO guideline cited in the article suggests about 0.8 g/kg/day and advises avoiding intake above 1.3 g/kg/day in patients at risk of progression. The strongest historical signal for slowing kidney failure appeared in carefully selected patients with advanced CKD who followed very-low-protein diets around 0.3–0.4 g/kg/day, usually with amino-acid or keto-acid supplements. 1

That last number is not a DIY target. The same article notes that the largest MDRD trial did not clearly show that more restriction slowed kidney-function loss, and that long-term follow-up of its very-low-protein group found higher mortality. The authors do not treat that observational follow-up as proof that the diet caused the deaths, but it makes severe restriction difficult to call harmless. 1

The newer question is whether an old dietary intervention adds anything on top of ACE inhibitors or ARBs, SGLT2 inhibitors, finerenone, and—where appropriate—GLP-1 receptor agonists. Attia's answer is skepticism, not a new dosing rule: no large randomized trial has tested the additive value of a low-protein diet on top of optimized contemporary CKD care. That leaves a gap between physiological plausibility and a modern clinical recommendation. 1

His August 3 episode with biophysical chemist Jim Otvos supplied the measurement side of the same problem. The episode explains why LDL particle number (LDL-P) and apoB can add information beyond LDL cholesterol, and discusses NMR-derived measures including LP-IR for insulin resistance, GlycA for inflammation, and the experimental Metabolic Vulnerability Index (MVX). None of those topics produces a supplement protocol; the useful point is that a familiar lab panel can leave clinically relevant variables unmeasured. 2

What changed for a reader: protein intake now needs a disease-state label. A high-protein target for a healthy person, a CKD guideline range, and a very-low-protein diet near dialysis are not interchangeable. If CKD is part of the picture, the relevant next step is a clinician or renal dietitian who can combine eGFR, albuminuria, nutritional status, muscle mass, and current kidney-protective therapy—not copying a number from a longevity article.

David Sinclair: enthusiasm without a new protocol

Sinclair's in-window output was brief. On August 8, he quoted a post describing epigenetic reprogramming as one of the first technologies that makes age reversal feel like an engineering problem, adding only: "🎯 epigenetic restoration." The post names no study, intervention, dose, safety result, or human endpoint. 3

On August 9, he thanked collaborators at St. Jude's while quoting a third-party post about AI helping researchers do "160 years of biology work in two months." That is a signal about research capacity, not evidence that an age-reversal treatment works. It also does not establish what St. Jude's collaboration produced. 4

The absence of a dose or endpoint is the important fact here. Epigenetic restoration is a research direction, not a reason to start an NAD-raising product, fasting schedule, or reprogramming regimen. Those interventions require their own human evidence; a mechanistic slogan cannot substitute for it.

What changed for a reader: nothing in these two posts supports a new consumer action. They are worth tracking as indicators of where Sinclair's attention is moving, but they do not meet the threshold for a protocol update.

Bryan Johnson: a recipe, a cancer claim, and no new stack

The dated Blueprint blog listing shows one in-window article: an August 7 whipped protein yogurt recipe written by Vanessa Gibbs. The page is editorial Blueprint content, not a direct statement from Johnson. It combines one scoop of Blueprint Longevity Protein, two-thirds of a cup of plant-based Greek yogurt, 10 macadamias, and half a cup of blueberries. The page estimates about 40 g of protein, 25 g of fat, 8 g of fiber, and 465 calories, but notes that the numbers vary with the ingredients and brands used. 5

That is a food idea, not a newly announced supplement-stack change. The ingredient list shows how Blueprint is packaging protein, nuts, berries, and a branded powder; it does not show that this bowl improves longevity, and it is not evidence for a particular daily protein target.

Johnson's own August 3 post was more consequential—and more speculative. He wrote that combining sildenafil with statins blunts cancer spread, citing 18% lower colorectal-cancer mortality and 15% lower risk of metastasis. The same post says the study was preclinical in cell and mouse models, while the human data were observational. 6

The caveat changes the meaning of the post. A cell or mouse mechanism can generate a trial question; observational human data can show an association. Neither establishes that a person should take sildenafil, tadalafil, or a statin to prevent or treat cancer. The post's mention that the mechanism might extend from sildenafil to tadalafil is also a mechanistic extrapolation, not a clinical result.

The dated Blueprint output therefore contains a recipe with transparent quantities but no new dose for a longevity supplement, and Johnson's social output contains a high-interest cancer hypothesis with an explicit evidence warning. Those belong in different mental folders.

What changed for a reader: if a protocol includes a branded powder, separate the recipe's nutrition estimate from the product's health claims. If a post combines a prescription drug with a cancer outcome, stop at the research question unless a clinician and a relevant clinical trial support moving further.

Rhonda Patrick: the useful distinction is dose versus device

Patrick's posts and FoundMyFitness updates had more practical detail than Sinclair's this week, but they also drew sharper boundaries around what the data can support.

Exercise snacks and movement breaks

On August 3, Patrick suggested accumulating 10 minutes per day of activity vigorous enough to make breathing harder. She listed stairs, uphill walking, carrying something heavy, and short "exercise snacks" as examples, while saying a formal HIIT workout is not required. This is her public suggestion, not a dose established by the post itself. 7

Her FoundMyFitness brief, updated August 3, gives a more bounded finding. A systematic review and three-level meta-analysis pooled 21 randomized crossover trials involving 433 participants. Brief standing or activity breaks during prolonged sitting produced small acute improvements in executive function and memory, but no clear benefit for global cognition, information processing, or attention. The evidence certainty was low for most outcomes and very low for global cognition. 8

The page says walking breaks showed a relatively consistent signal, but its exploratory frequency analysis does not establish a universal rule to move every 20–30 minutes. That is the part worth keeping: interrupting long sitting is a modest, reversible experiment; it is not a proven dementia-prevention protocol.

Patrick also posted that older adults doing moderate-intensity aerobic exercise three times per week for one year saw a 2% increase in hippocampal volume. Her post does not include the paper title, sample size, or enough protocol detail to turn that number into a personal prescription. 9

Finnish sauna versus infrared

On August 6, Patrick gave a clear device comparison. She said she would choose a traditional Finnish sauna when both options are available because the long-term observational evidence is attached to Finnish sauna use, not to infrared devices. She also summarized a 2025 head-to-head study in which traditional sauna raised core temperature, whereas a 45-minute far-infrared session raised skin and body temperature and cardiac output without raising core temperature. There is no validated time-for-time conversion between the two. 10

That is a useful correction to the common phrase "sauna is sauna." Infrared may be the option a person can tolerate and use consistently, but the evidence for it is smaller and the thermal dose is not demonstrably equivalent. The honest action is to choose the device you can use safely, without claiming that 45 minutes in one modality equals a particular number of minutes in the other.

Photobiomodulation: a signal, not a shopping specification

FoundMyFitness updated a photobiomodulation brief on August 7. The underlying meta-analysis included 13 randomized trials comparing red or near-infrared light treatment with sham treatment, mainly in healthy active men and lower-limb muscles. Peak torque was higher with photobiomodulation immediately after exercise and at 24, 48, 72, and 96 hours; self-reported soreness was lower immediately and at 24 hours, but not clearly later. 11

The page also reports exploratory favorable doses around 180–300 joules for some outcomes, while stressing that treatment parameters varied across studies and that there was no consistent dose–response relationship. That number is not a standardized consumer regimen. Device wavelength, power, treatment area, timing, and the muscle being treated all change the intervention.

Two other FoundMyFitness briefs updated August 7 show the same evidence boundary. A prenatal omega-3 trial used 2.4 g/day of long-chain omega-3 from pregnancy week 24 to one week after birth and later found small differences in brain blood flow and oxygen use in children, without clear differences in intelligence or several cognitive and mental-health measures. A cross-sectional HRT analysis of 1,374 women aged 60 or older found associations with some cognitive measures, but did not record HRT type, dose, timing, or duration and cannot show causation. 12 13

What changed for a reader: Patrick's strongest actionable suggestions this week are the least glamorous: brief movement, a manageable amount of vigorous activity, and a realistic sauna choice. The device-based and hormone-related findings are reasons to refine a question—not reasons to infer a dose from a headline.

Where the four did—and did not—converge

There is no clean same-week agreement on a supplement or treatment protocol among these four sources. That absence is itself useful. Attia discussed protein in CKD and lipid testing; Sinclair posted about epigenetic restoration; Johnson discussed a protein recipe and a preclinical cancer hypothesis; Patrick discussed movement, sauna, and photobiomodulation. These are adjacent longevity topics, not a shared intervention.

The closest apparent overlap is protein, but it is not a real disagreement. Attia's article asks whether protein restriction adds benefit in CKD under modern medical care. Blueprint's recipe supplies one meal with an estimated 40 g of protein. Sinclair's stronger anti-protein comments appeared in the previous coverage window, not this one. Treating those three facts as a current expert showdown would create a disagreement the sources did not make.

Question	What this week's sources actually support	What they do not support
Should protein be restricted?	CKD status changes the question; the additive benefit of restriction on modern therapy remains unproven. 1	A universal high- or low-protein target for every adult.
Can brief activity help?	Small acute cognitive benefits from movement breaks, plus Patrick's suggestion to accumulate short vigorous bouts. 7 8	A proven replacement for structured exercise or a fixed dementia-prevention dose.
Is infrared sauna equivalent to Finnish sauna?	Patrick says the evidence and thermal response are not equivalent; no time conversion is validated. 10	That infrared is useless, or that a 45-minute session has a known Finnish-sauna equivalent.
Should sildenafil or tadalafil be used against cancer?	Johnson's post describes a preclinical and observational signal. 6	Self-prescribing a prescription drug for cancer prevention or treatment.
Does a longevity mechanism equal a protocol?	Sinclair's posts provide no dose, endpoint, or safety data; PBMT's reported doses are exploratory. 3 11	Turning a mechanism or a single study number into a personal regimen.

What is reasonable to carry forward

If CKD is relevant, do not copy either side of the protein debate. Bring your kidney function, albuminuria, nutritional status, muscle goals, and current medications to a clinician or renal dietitian. The evidence does not justify treating a general longevity target as a CKD prescription.
Use movement as a low-cost test, not a magic number. Try brief breaks during long sitting or short vigorous bouts that fit your current capacity. Track whether you can sustain them; do not infer long-term disease prevention from an acute cognitive effect.
Treat sauna modality as a real variable. Finnish sauna has the stronger long-term evidence base in Patrick's summary. Infrared can still be the practical choice, but there is no validated conversion that lets you borrow Finnish-sauna claims by multiplying minutes.
Keep photobiomodulation in the experimental column. The meta-analysis supports a recovery signal under varied protocols, not a universal home-device dose. A joule count without wavelength, power, area, and timing is incomplete.
Do not turn Johnson's cancer post into a medication experiment. The post itself says the evidence is preclinical and observational. That is a reason to watch for clinical trials, not to add sildenafil, tadalafil, or a statin.
Read the person, date, endpoint, and dose together. A recipe, a mouse study, a cross-sectional association, a social post, and a clinical recommendation are different kinds of evidence even when they use the same language of "protocol" or "longevity."

The week did not produce a new consensus stack. It produced several useful stopping points: before restricting protein in CKD, before equating sauna devices, before buying a light-treatment dose, and before treating a mechanistic cancer result as care.