Poor sleep and stroke risk, a direct AMPK activator extends lifespan in three species, GLP-1 safety graded honestly and four other things worth your time this week.

Week of October 12th, 2026

This week's theme is the distance between a signal and a recommendation. A senolytic combination reversed liver scarring in a small randomized trial, and a direct AMPK activator stretched lifespan in three invertebrate species, but both come with real asterisks. A Karolinska review sorted the GLP-1 safety conversation into what is established and what is still noise, while a Tufts pilot asked what to do when people stop the drug and the weight returns. A Science paper on 894 dogs showed how epigenetic clocks track lifespan, an umbrella review tied insomnia to stroke and hospital admission, and a midlife cohort hinted that faster epigenetic aging may show up as forgetfulness in your forties. Seven findings, graded honestly. Here's what's worth your attention.

1. A Cancer Drug Plus a Plant Compound Cleared Liver Scarring in a Small Trial

Fibrotic MASH, the scarring stage of metabolic fatty liver disease, has very few treatment options, and cellular senescence is one suspected driver of the inflammation and fibrosis. A team at Amsterdam UMC, with senolytic pioneer James Kirkland among the co-authors, published a phase 2, double-blind, placebo-controlled trial in Nature Metabolism on October 1, 2026. Thirty-one adults (median age 56, 58% with type 2 diabetes) received dasatinib 100 mg plus quercetin 1,000 mg daily, or placebo, three days a week for three weeks, repeated across three 7-week cycles. The primary endpoint, at least one stage of fibrosis improvement without MASH worsening on paired biopsies, was met by 47% on D+Q versus 7% on placebo (P = 0.02). MASH resolution was 53% versus 7%, and single-nucleus RNA sequencing showed lower senescence and fibrosis signatures.

Grade it as a promising proof-of-principle with a small-sample caveat the size of the trial itself. This is a human randomized trial, which is the right design, but 31 people over 21 weeks cannot settle efficacy. Adverse events were more frequent on D+Q (82% versus 43%), though the authors describe them as self-limiting. Dasatinib is a prescription cancer drug, so this is a result for the next, larger trial rather than a supplement stack to copy at home.

Koning M, Kovynev A, Fondevila MF, et al. Senolytics dasatinib and quercetin in metabolic dysfunction-associated steatohepatitis: a proof-of-principle randomized, controlled trial. Nature Metabolism. 2026. DOI: 10.1038/s42255-026-01643-4 https://doi.org/10.1038/s42255-026-01643-4

2. The Seven GLP-1 Safety Worries, Sorted by How Much Evidence Actually Backs Them

With GLP-1 receptor agonists now used by millions, every rare adverse event becomes a headline, and it is hard to tell a real signal from noise. Peter Ueda at the Karolinska Institutet published a narrative review in the Journal of Internal Medicine on October 9, 2026, appraising seven concerns: thyroid cancer, non-arteritic anterior ischaemic optic neuropathy (NAION), suicidal behaviour and self-harm, diabetic retinopathy complications, acute pancreatitis, gallbladder and biliary disease, and intestinal obstruction. Evidence came from mechanistic work, spontaneous reports, randomized trials, and cohort studies.

Grade it as a useful, honest map rather than new data. Gallbladder and biliary disease has the most consistent evidence for a causal link. Retinopathy worsening looks like an early effect of rapid glucose improvement in people who already have retinopathy. For thyroid cancer, NAION, suicidality, pancreatitis, and obstruction, causality remains uncertain, and rare events plus confounding limit what anyone can conclude. The practical point is that even large relative increases can be small absolute risks, but they matter more if your baseline risk is high. Talk through your own history with your prescriber.

Ueda P. GLP-1 receptor agonists and serious adverse events: Evidence, uncertainty and clinical implications. Journal of Internal Medicine. 2026. DOI: 10.1111/joim.70170 https://doi.org/10.1111/joim.70170

3. What Happens After You Stop a GLP-1? A Small Trial Tested Two Ways to Soften the Rebound

Many people stop GLP-1 therapy because of cost, coverage, side effects, or preference, and weight often returns quickly, yet there are few evidence-based strategies for the transition. A Tufts Food is Medicine Institute team, working with UT Southwestern, published a pilot randomized trial in Obesity Science & Practice on October 7, 2026. Thirty-nine adults who had lost weight on a GLP-1 were randomized at the time of stopping to usual care (n=14), a digital wellness app (n=12), or medically tailored meals (n=13), with mean baseline BMI 31.4. Over the first four months, both support arms showed less total weight gain than usual care: -7.16% (95% CI -12.27 to -2.04) for the app and -7.17% (-12.20 to -2.14) for the meals.

Grade it as a hypothesis-generating pilot. The primary outcomes, adherence and satisfaction, did not differ between groups, and weight was an exploratory measure in only 39 people over four months. Still, the direction is encouraging and the interventions are low risk. The takeaway is that a plan for the off-ramp, whether structured meals or a tracking tool, is worth discussing before you stop, and larger and longer trials need to confirm it.

Hollis-Hansen K, Vonderschmidt A, Haskins C, et al. Preventing weight regain after glucagon-like peptide-1 therapy discontinuation: a pilot randomized controlled trial. Obesity Science & Practice. 2026. DOI: 10.1002/osp4.70194 https://doi.org/10.1002/osp4.70194

4. Directly Switching On AMPK Extended Lifespan in Yeast, Worms, and Flies

Metformin is the most famous longevity candidate, but it acts through many pathways, so it is hard to say how much of any benefit runs through AMPK, the cell's energy sensor. A multi-institution group led from the MRC Laboratory of Medical Sciences and Imperial College London, with Queen Mary University of London, published in Aging Cell in October 2026 using compound 991, which binds and directly activates AMPK. It extended lifespan in fission yeast, the nematode C. elegans, and Drosophila. In mice, three weeks of treatment produced a pro-longevity liver proteomic profile, including signs of greater ATP production and mitochondrial biogenesis and suppressed mTOR signaling.

Grade it as a satisfying mechanistic win with a long runway to the clinic. The lifespan effect disappeared in worms and yeast lacking functional AMPK, which supports the mechanism, but mice showed a molecular shift rather than a longer life, and higher doses shortened lifespan in some experiments. This is invertebrate and short-term mouse data. Co-author Charalampos Rallis explicitly cautioned against taking supplements on the strength of these findings, and that is the right reading.

Dos Santos E, Blickling M, Leiper FC, et al. Direct pharmacological activation of AMPK extends lifespan in yeast, worms and flies. Aging Cell. 2026;25(10):e70721. DOI: 10.1111/acel.70721 https://doi.org/10.1111/acel.70721

5. A Dog Epigenetic Clock Explains Why Big Dogs Die Young

Dogs offer a natural experiment in lifespan, since a Great Dane and a Chihuahua share a species but not a life expectancy. Blaise Mariner, Noah Snyder-Mackler, and colleagues at Arizona State University and partner institutions published in Science on October 8, 2026, analyzing 1,640 methylomes from 894 dogs. Their epigenetic clock predicted mortality, and epigenetic aging was fastest early in life. Larger dogs and male dogs, which live shorter lives, showed accelerated molecular aging. Size-related methylation changes were concentrated at transposable elements, while sex-related changes clustered on the X chromosome.

Grade it as a strong, large-scale observational study with a correlative core. It shows that the clock tracks lifespan differences and nominates transposable elements as possible mediators, but it does not prove they cause shorter lives. For human readers the value is conceptual: epigenetic clocks can reflect real differences in mortality risk across a population, and the biology of body size and aging may run through genome regions we rarely discuss. No action for your own routine, but a useful validation of the approach.

Mariner BL, McCoy BM, Greenier A, et al. Epigenetic aging and transposon dysregulation reflect size-related lifespan compression in dogs. Science. 2026;394(6820):eaeb2986. DOI: 10.1126/science.aeb2986 https://doi.org/10.1126/science.aeb2986

6. Insomnia Was Linked to 26% Higher Stroke Risk Across 1.3 Million People

Poor sleep is usually framed as a quality-of-life problem, but the question is whether chronic insomnia also predicts hard outcomes for the brain and body. Luca Vignatelli and colleagues across Italian, Austrian, German, and Swiss centers published an umbrella systematic review in Sleep Medicine Reviews, covered by medical media on October 2, 2026. They combined 16 existing meta-analyses with three new ones. Insomnia consistently predicted dementia, depression, and suicide. The new analyses found a stroke hazard ratio of 1.26 (95% CI 1.05 to 1.53; 9 studies, 1,343,164 participants) and a 28% higher odds of hospitalization for any reason (OR 1.28, 95% CI 1.14 to 1.43; 5 studies, 517,620 participants). Death was not associated.

Grade it as a consistent association from observational data, not proof of cause. Definitions of insomnia varied, confounding is likely, and stroke results differed meaningfully between studies. Whether treating insomnia lowers these later risks is still unanswered. The takeaway is to treat persistent insomnia as a medical issue worth raising with your clinician, not something to endure.

Vignatelli L, Leone MA, Pupillo E, et al. Insomnia as risk factor for brain, mental and general health consequences: an umbrella systematic review. Sleep Medicine Reviews. 2026;102372. DOI: 10.1016/j.smrv.2026.102372 https://doi.org/10.1016/j.smrv.2026.102372

7. Faster Epigenetic Aging in Your Forties May Predict Forgetfulness a Decade Later

Subjective memory complaints, such as forgetting a name or a medication, are an early marker of dementia risk, and epigenetic clocks like DunedinPACE are increasingly used to track aging pace. Heejung Jang, David Almeida, and colleagues at Penn State and other universities linked the two using the National Study of Daily Experiences and MIDUS (N=232), in the Journals of Gerontology Series A, 2026. Across the whole sample there was no main effect. But among adults aged 40 to 49, a faster DunedinPACE was associated with more prospective memory lapses about a decade later, along with greater irritation and interference from those lapses.

Grade it as an early, hypothesis-generating signal. The sample is small, memory lapses were self-reported, and the finding emerged from an age interaction rather than an overall effect, so it needs replication. If it holds, it suggests early midlife may be a sensitive window for epigenetic aging and cognition. For now, it supports focusing on the modifiable basics in your forties rather than ordering a clock test and acting on the number.

Jang H, Holmberg JC, Nicholson JS, et al. Linking epigenetic age acceleration to self-reported daily memory lapses: evidence from the National Study of Daily Experiences. J Gerontol A Biol Sci Med Sci. 2026;81(10). DOI: 10.1093/gerona/glag202 https://doi.org/10.1093/gerona/glag202

Have a great week,

Dr. Smith and Winston

Medical disclaimer

The content above is for informational and educational purposes only and does not constitute medical, nutritional, or professional advice. It is not intended to diagnose, treat, cure, or prevent any disease and is not a substitute for personalized guidance from a licensed physician or other qualified healthcare provider. Individual needs and risks vary. Do not begin, stop, or modify any exercise regimen, dietary plan, medication (including dasatinib, quercetin, GLP-1 receptor agonists, or sleep medications), or supplement based on this article alone. Consult your doctor first, particularly if you have or suspect a cardiovascular, metabolic, liver, gastrointestinal, or neurological condition, take prescription medications, or are pregnant or breastfeeding. Studies summarized here vary in design and quality; preclinical and animal findings often do not translate to humans, and associations reported in observational research do not prove cause and effect.