Longevity Digest, October 18–25: a 42% lower risk of melanoma's return, and a fight over whether the field is real
In the week of October 18–25, 2026: a personalized mRNA cancer vaccine cut the risk of melanoma recurrence or death by 42% in a phase 3 trial, The Lancet called the longevity industry "a goldrush without gold" and David Sinclair answered it, Bryan Johnson explained why he skips coffee, and Rhonda Patrick traced a microplastics mechanism and audited an omega-3 death estimate.
This digest covers the public output of Peter Attia, David Sinclair, Bryan Johnson's Blueprint team, and Rhonda Patrick between October 18 and October 25, 2026, Pacific Time. Attia spent the week on cancer, with a long conversation with an oncologist on Monday and a phase 3 trial result written up on Saturday. The loudest item of the week was an argument: The Lancet published an essay describing the longevity industry as "a goldrush without gold," and Sinclair answered it within a day 12.
The week in one screen
| Voice | Focus this week | Evidence tier | What a reader would do with it |
|---|---|---|---|
| Peter Attia | A nearly two-hour conversation with oncologist Siddhartha Mukherjee, then a write-up of a phase 3 melanoma vaccine trial 34 | A podcast conversation whose detailed notes sit behind a membership; a free article built on a randomized trial | The trial carries the week's hardest number: 42% lower risk of recurrence or death at a median 24.6 months, in one setting, resected melanoma 5 |
| David Sinclair | An NMN mouse preprint from 2024, posted without comment, and a public reply to the Lancet essay 67 | A preprint; an argument with no data on either side | The lifespan gain appeared in female mice only; the public argument is worth following for what a human result would look like 8 |
| Bryan Johnson / Blueprint | Why he does not drink coffee, and what the coffee literature says about timing 9 | A first-person protocol post, on a page that also sells products | Coffee before noon, one to four cups, is where the observational benefit sits; the post's headline claim names no study 10 |
| Rhonda Patrick | A mechanism for how microplastics might damage cells, and a death estimate for low omega-3 intake 1112 | A chemistry and cell study in Nature; a population burden model | The plastic paper measures chemistry in a dish; the omega-3 item is the one with a clear dietary action attached 13 |
Peter Attia: a phase 3 result, and a case for changing what longevity targets
Peter Attia is a physician, the founder of the Early Medical practice, and the host of The Peter Attia Drive. He released a guest episode on October 19 and a free article on October 24, and both concern cancer 34.
A vaccine built from each patient's own tumour
The article covers intismeran autogene, an mRNA therapy made for one patient at a time. After surgery, the removed tumour is sequenced, up to 34 mutated sequences likely to provoke an immune response are chosen, and a single mRNA strand coding for all of them is manufactured 4. The vaccine is given alongside pembrolizumab, the checkpoint inhibitor sold as Keytruda, which lifts the shield tumours raise against immune attack 4.
The phase 3 trial, INTerpath-001, enrolled 1,137 people whose stage IIB to IV melanoma had been completely removed, and randomized them two to one to the combination or to pembrolizumab alone 5. At a median follow-up of 24.6 months the combination cut the risk of recurrence or death by 42% (hazard ratio 0.58, 95% CI 0.45–0.75, p<0.001) and the risk of distant metastasis or death by 38% 5.
Two-year figures give the size in plainer terms. Recurrence-free survival was 82.2% on the combination against 70.0% on pembrolizumab alone, and distant metastasis-free survival 88.5% against 80.2% 5. The benefit arrives late: during the first six months, while the nine vaccine doses are being given, the two arms performed alike (hazard ratio 0.99). After six months the hazard ratio fell to 0.42, a 58% reduction in the risk of recurrence or death 5.
The price of that benefit shows up in side effects. Nearly everyone on the combination reported a treatment-related adverse event, 95.8% against 90.0% on pembrolizumab alone, though serious events were similar, 14.2% against 13.4% 5. Fever (52.2%), injection-site pain (51.0%), and fatigue (38.5%) were the most common complaints, which fits a therapy built to provoke an immune response 5.
Merck and Moderna released the detailed results on October 22, ahead of a presentation at the ESMO Congress, and the data have been accepted by The New England Journal of Medicine; overall survival is still immature 5. They extend a phase 2 result: at five years of follow-up in the phase 2 KEYNOTE-942 trial, the combination had cut the hazard of recurrence or death by 49% (hazard ratio 0.51, 95% CI 0.29–0.89) and of distant metastasis or death by 59% 14.
Attia's article makes the case that this is a template rather than a melanoma drug. Melanoma responds to checkpoint inhibitors because its mutation count is high, which gives the immune system many targets; a vaccine supplies targets to tumours that offer few 4. Trials in lung, bladder, kidney, and pancreatic cancer are running, and each tumour type will have to earn its own approval 4.
What the trial does not yet say
The published numbers are an interim analysis in one setting: melanoma, after complete surgical removal, with a median of two years of follow-up. Attia's article lists the other limits, including a per-patient manufacturing cost likely in the tens of thousands of dollars and inconsistent results from earlier cancer vaccines that never reached approval 4.
One number in the article needs correcting against the release. The article gives 88.2% of patients recurrence-free at two years; Merck's release gives 82.2% recurrence-free and 88.5% free of distant metastasis, and the release is the figure to use 45.
The conversation, and a word for what goes wrong first
Episode #412 runs nearly two hours with Siddhartha Mukherjee, an oncologist at Columbia University and the author of The Emperor of All Maladies, which won the Pulitzer Prize in 2011 and was reissued in late 2025 with four new chapters 3. The episode page carries a timestamped list of what the two men cover; the first stretch of show notes is public and the rest is members-only 3.
Mukherjee's account of what changed in fifteen years splits into prevention, early detection, and treatment. On prevention, he describes cancer as needing both a loaded gun and a trigger, with chronic inflammation as the gun, and notes that medicine still has no biomarker that identifies a person's inflammatory risk the way cholesterol identifies cardiovascular risk 3. He expects a cholesterol equivalent for cancer to be found, and says it would open the way to prevention trials 3.
On early detection, Mukherjee is the week's most useful sceptic. Blood tests that pick up cancer signals would, used across the general population, generate false positives, extra imaging, and biopsies without lowering cancer mortality; the same tests look valuable in people with a known predisposing mutation or a history of cancer 3. He also flags that cancer trends lag exposures by five to ten years, which is why the rise in early-onset colorectal cancer has no explanation yet 3.
On treatment, the gains are real. Immunotherapy now extends the lives of smokers with lung cancer, a result Mukherjee wrote off as impossible fifteen years ago, and CAR T-cell therapy has left people with previously unresponsive leukaemia and lymphoma alive fifteen years after treatment 3. He also names daraxonrasib for RAS-driven cancers, with pancreatic cancer patients now living months to years after the cancer has spread 3.
The frame that matters for this channel comes later, and only its outline is public. Mukherjee describes "regenerative failure," the declining ability of tissues to repair and replenish themselves with age, and the episode asks whether longevity should be pursued as the prevention of individual diseases or as the preservation of regenerative capacity 3. The same conversation covers hyperinsulinemia and metabolism as contributors to cancer risk, the thymus as a predictor of longevity, T-cell ageing, chemoprevention, and what a real biomarker of ageing would have to look like compared with how atherosclerosis was measured 3.
A reader gets the conclusions from that list and the reasoning only from the members' notes. Where the episode is checkable — the immunotherapy and CAR T results, the false-positive problem in screening — the claims rest on published trials; the regenerative-failure argument is an argument, and the free page states it without the evidence behind it 3.
David Sinclair: a preprint from 2024, and a reply to The Lancet
David Sinclair is a professor of genetics at Harvard Medical School and the chairman of the board of Life Biosciences. His week held no new data: a link to two-year-old mouse work and a public answer to a critical essay.
NMN in mice, in 2024
On October 24 Sinclair posted the title of a paper and its link, with no comment of his own 6. The paper is a preprint posted to bioRxiv in June 2024 by Alice Kane and colleagues, and it has not been peer reviewed 8. Treat it as background: the findings date to 2024.
Mice given nicotinamide mononucleotide, or NMN, a NAD precursor sold as a supplement, moved more, kept more youthful gene expression patterns, and scored lower on frailty 8. The treated animals had no rise in tumour counts or severity in any tissue, which is the safety question a lifespan study of a NAD booster has to answer 8.
The lifespan result splits by sex. Median lifespan rose 8.5% in female mice and did not rise in males; male mice instead gained the metabolic benefits, with less fat accumulating late in life and better metabolic health 8. The authors' own conclusion is that the two sexes responded differently to the same drug 8.
The timing of the post is the part worth noticing. Sinclair shared the mouse preprint two weeks after a 410-person phase 3 trial found that nicotinamide riboside, a different NAD precursor, left early Parkinson's symptoms slightly worse than placebo over a year 15. The two results answer different questions: whether a pathway can be moved in a mouse, and whether a pill changes a human disease.
Two delusions, and a reply
The Lancet's October 24 issue carries an essay by Seamus O'Mahony in its "art of medicine" slot, titled "Longevity is a political concern" 1. O'Mahony is a consultant gastroenterologist who worked in NHS hospitals for more than thirty years, and he published The Longevity Delusion: Money, Myths and Bad Science this year 16. The passage being circulated describes two delusions: that ageing in humans is as malleable as it is in roundworms, fruit flies, and mice, and that longevity is primarily a scientific or medical concern rather than a political one 2.
Sinclair answered on October 25 with a quote of the essay. He called it "such unscientific polemic," compared the prediction to the New York Times writing off powered flight weeks before the Wright brothers flew, and said that "'several renowned experts told me' is what people write when they lack evidence" 7. He tied the essay to the ER-100 results reported two weeks earlier 7.
A third position came from João Pedro de Magalhães, a professor at the University of Birmingham who studies ageing. He agreed the field is "drowning in hype and commercial interests pushing unproven interventions," disagreed with the conclusion, and wrote that the task is "to tell hype from science, not to dismiss the science because of the hype" 2.
Nobody in this exchange produced a result. O'Mahony's argument is that the human body is far less malleable than a roundworm or a mouse, and that the field's promise outruns its evidence; Sinclair's reply is that the field is progressing and its critics lag behind it; Magalhães grants the first premise and rejects the inference. The ER-100 readout Sinclair cites was a safety result in three patients with glaucoma, reported in this digest two weeks ago, and it says nothing about lifespan.
Bryan Johnson: coffee, which he does not drink
Bryan Johnson's Blueprint published one post in the window, on October 21, under his own byline: "Does Bryan Johnson Drink Coffee?" 9. The answer is no, with exceptions. He writes that caffeine destabilizes his mood, disrupts his sleep, stains his teeth, and produces too much of a metabolic rollercoaster, and that he will occasionally drink decaf 9.
The post then argues that coffee suits most other people. It states that drinking coffee is linked to living nearly two additional years of healthy life, and names no study for that figure; the mechanism it offers is the compound count, more than 2,000 in coffee, including chlorogenic acids and melanoidins, which the post credits with antioxidant effects and, for melanoidins, lower blood pressure 9.
The timing claim is the checkable part. Compared with non-drinkers, people who drink coffee only in the morning had a 16% lower risk of death from any cause and a 31% lower risk of cardiovascular death, while all-day drinkers did not show the same benefit 9. Those figures come from a 2025 analysis of US adults published in the European Heart Journal, and they are associations from survey data, not the result of an experiment 10.
His practical advice follows from caffeine's half-life of five to six hours: drink it early, stop eight to ten hours before bed, keep to one to four cups a day, skip sugar and syrups, and switch to decaf if sleep suffers 9. The page carries Blueprint product links and the standard disclaimer that the statements have not been evaluated by the Food and Drug Administration 9.
The post's two halves carry different weight. His own decision rests on how caffeine feels to him, which is a reasonable basis for a personal choice and no basis at all for a recommendation. The timing and dose advice rests on population data that anyone can use, with the caveat that morning coffee may be a marker for an early-rising, employed, healthier life as much as a cause of one 10.
Rhonda Patrick: the water against plastic, and the omega-3s missing from the diet
Rhonda Patrick has a PhD in biomedical science and hosts FoundMyFitness. She published no new episode and no new digest issue in the window; her five posts covered body fat and testosterone, sauna and infrared sauna, omega-3 intake, and microplastics.
A mechanism for microplastics, measured in a dish
Her October 23 post describes work on what happens when microplastics meet water. Plastics are generally regarded as chemically inert; the new paper measures that surface and finds hydroxyl radicals, a reactive oxygen species, forming there on their own, without light or an added catalyst 13.
The Nature paper tested six commodity plastics: polystyrene, polypropylene, polyethylene, PVC, PET, and PMMA. All six generated radicals, and the more water-repellent the surface, the more radicals it produced, which put polyethylene and polypropylene at the top 13.
Two follow-on experiments show why this could matter biologically. Proteins in solution clumped at the plastic surface, and the clumping fell when a radical scavenger was added, which ties the aggregation to the radicals themselves 13. Some mouse red blood cells exposed to the particles deformed into fragmented, helmet-like shapes and released haemoglobin, and the amount released tracked each plastic's radical production 13.
Patrick states the boundary in her own post: these experiments establish a mechanism and say nothing about the risk of everyday exposure 11. The gap between them is the one that matters for a reader, and it is wide. The paper measured chemistry at a surface, in buffer, with concentrations set by the experiment; what reaches a person's blood from a water bottle, and at what dose, stays unmeasured.
Where the 84,000 deaths come from
Her October 22 post argues that a nutrient people are short of deserves as much attention as an ingredient they avoid. Low seafood omega-3 intake accounted for an estimated 84,000 deaths in one Harvard-led US analysis, against 82,000 for trans fats, and nearly 90% of Americans fall short of seafood recommendations 12. Her advice is EPA and DHA from fatty fish, with supplementation filling whatever gap remains 12.
The two numbers she quotes trace to a 2009 comparative risk assessment of twelve dietary, lifestyle, and metabolic factors, which used 2005 US data; low omega-3 fats came in at 84,000 cardiovascular deaths and high trans fats at 82,000 17. Trans fat intake has since fallen sharply in the US food supply, which dates the comparison.
The same research group re-ran the analysis on 2012 data and published it in JAMA in 2017. Low seafood omega-3 intake, under 250 milligrams a day, accounted for 54,626 estimated cardiometabolic deaths, 7.8% of the total, with an uncertainty interval of 45,541 to 65,053 18. Seafood omega-3 remained one of the four dietary factors with the largest estimated toll, after sodium, nuts and seeds, and processed meat 18.
Both estimates point the same way, and the size of the number moves with the year of the data and the model behind it. What survives the revision is the 250 milligrams a day of EPA and DHA that the 2017 analysis used as its reference intake 18.
Where the four of them disagreed
This week's material differs less in subject than in what each voice asked a reader to trust. Attia's week ended with a randomized trial of more than a thousand patients and a hazard ratio; Patrick's two items rest on a chemistry paper in cells and a model of population deaths. Johnson's rests on how caffeine feels to him, with a survey analysis attached, and Sinclair's on a preprint he did not write and an argument about a field's future 58913.
Coffee and fish: the same kind of evidence, read two ways
Patrick's omega-3 post and Johnson's coffee post both build on observational data from large US surveys, and each turns an association into advice 912. Neither analysis can rule out that people who eat fish and people who drink coffee before noon differ in other ways that matter more.
The two cases separate on cost. Adding a serving of oily fish or an omega-3 supplement is cheap, and the burden estimates say the intake gap is wide. Moving coffee to the morning costs nothing either, though the 16% figure invites a change in behaviour on the strength of a correlation.
An argument with no trial behind it
Sinclair and O'Mahony disagree about whether the field has delivered, and neither side's evidence sits in the same place. O'Mahony's essay points at the absence of a demonstrated human lifespan effect; Sinclair's reply points at a trajectory and at a three-patient safety readout 17.
Attia's position from the previous week sits close to Magalhães's. He described dropping metformin for longevity and growing more uncertain about rapamycin on the strength of the data, and framed good practice as staying "married to getting things as right as possible and reducing uncertainty by following the data" 19. The difference is that Attia's adjustments came with numbers attached, and the Lancet exchange came with none.
What to carry forward
- Watch for the New England Journal of Medicine publication of INTerpath-001 and for overall survival, which is still immature at a median of 24.6 months of follow-up 5.
- Treat the vaccine's 42% as a melanoma number. Each other tumour type, including lung, bladder, kidney, and pancreatic, needs its own trial before the result generalizes 4.
- The most useful thing in Attia's episode is a caution: blood tests for cancer used across the general population generate false positives and testing cascades, and their value concentrates in people at known high risk 3.
- Hold NMN at mouse scale and at its 2024 date. The lifespan gain appeared in female mice only, and the human NAD precursor trial this month went the wrong way 815.
- Treat the "regenerative failure" reframing as a hypothesis for now. The free page lists it, the reasoning stays behind the membership, and no trial tests it 3.
- On microplastics, take the mechanism and leave the risk estimate alone until someone measures exposure and outcome in people 11.
- Aim for 250 milligrams a day of EPA and DHA, the reference intake the burden analyses used, from oily fish or a supplement. This is the item with the widest evidence base and the lowest cost in the week 18.
- If you drink coffee, drink it before noon and stop eight to ten hours before bed; the mortality association comes from the morning-only group 10.
References
- 1Longevity is a political concern, The Lancet
thelancet.com
- 2
- 3#412 ‒ The evolving science of cancer, with Siddhartha Mukherjee
peterattiamd.com
- 4Personalized medicine comes to cancer immunotherapy
peterattiamd.com
- 5
- 6
- 7
- 8
- 9Does Bryan Johnson Drink Coffee?
blueprint.bryanjohnson.co
- 10
- 11
- 12
- 13
- 14
- 15
- 16Seamus O'Mahony on the longevity industry, The Telegraph
telegraph.co.uk
- 17The preventable causes of death in the United States, PLoS Medicine
journals.plos.org
- 18
- 19#411 ‒ AMA #90: Strong convictions, loosely held
peterattiamd.com
This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.