Ketogenic diets (KDs)—very-low-carbohydrate and very-high-fat diets—have gained popularity as therapeutic against obesity and type 2 diabetes. However, their long-term effects on metabolic health remain understudied. Here, we show that, in male and female mice, a KD protects against weight gain and induces weight loss but over time leads to the development of hyperlipidemia, hepatic steatosis, and severe glucose intolerance. Unlike mice on conventional high-fat diet, KD-fed mice remain insulin sensitive and display low-insulin levels. Hyperglycemic clamp and ex vivo glucose-stimulated insulin secretion assays revealed systemic and cell-intrinsic impairments in insulin secretion. Transcriptomic profiling of islets from KD-fed mice indicated endoplasmic reticulum (ER)/Golgi stress and disrupted ER-Golgi protein trafficking, which were confirmed by electron microscopy showing a dilated Golgi network consistent with defective insulin granule trafficking and secretion. Together, these results suggest that long-term KD leads to multiple aberrations of metabolic parameters that caution their systematic use as a health-promoting dietary intervention.

Full Paper - https://doi.org/10.1126/sciadv.adx2752

  • jet@hackertalks.comOPMEnglish
    3·
    10 days ago

    TLDR - This study is interesting in mice, but the applicability to humans is minimal, especially as it contradicts already established human studies using a well formulated ketogenic diet (i.e. not 90% fat). It’s statements about glucose intolerance are not meaningful as they performed the test wrong. We also know a high seed oil diet in mice also causes dyslipidemia (studies from the 1960s when the oil companies were still funding research)… but hey, it got published… so we should talk about it.

    This is a very flawed paper, and isn’t well designed, it doesn’t transfer to humans, and it makes claims that are unsupportable for methodological reasons (i’ll get into them).

    This paper was brought to my attention by @farbidden_lands@quokk.au and I appreciate all the literature out there, and I want to thank them again for bringing it up.

    Issues at a glance with the paper:

    • In Mice (Mice are not humans)
    • Mice Metabolism around ketones if fundamentally different than humans
    • 90% Fat
    • Fat is Highly Processed Industrial Seed Oil (Soybean Oil)
    • Glucose Tolerance Tests require a 3 day re-carb-loading window to be accurate. This study had no carb-loading before the GTT.

    Mice - Mice have a different metabolism then humans, their ketone production pathways are different then humans, and have different chemical feedback loops to limit production.

    90% Fat - Humans following Ketogenic Metabolic Therapy do not target anywhere near a 90% Fat rate.

    90% Seed Oils - It has already been demonstrated in mice (and humans) that industrial oils are highly inflammatory and interfere with metabolism, lipids, and insulin. At best this is unintentionally mixing variables.

    Mouse Diet - If we look up the mouse diet used - D16062902

    • 90% of calories from fat
    • Fat source dominated by soybean oil (very high in linoleic acid, omega-6 PUFA)
    • Extremely poor omega-6:omega-3 ratio
    • 10% protein (low for mice)
    • Virtually no carbs and no fiber
    • I don’t think a high PUFA diet really is in the universe of anyone doing nutritional ketosis.

    Glucose Tolerance Test / Impaired Insulin Secretion - We see human’s exhibit similar glucose tolerance issues on a keto diet. Just as with the mice in this study, reintroducing glucose resolve the glucose tolerance sensitivity. In humans, at least, the pancreas keeps a reserve of insulin ready when in demand, and in a persistent ketogenic state this reserve isn’t required - so when exogenous glucose is reintroduced it takes a day or two to refill the reserve and get “normal” glucose response. The authors of this study even demonstrated reversal of glucose tolerance in mice when keto was discontinued (consistent with the behavior in humans) https://www.eurekalert.org/news-releases/1102798 BUT they didn’t document this in the paper… calling this glucose dysregulation is disingenuous - The fact they don’t mention the human equivalent studies on this topic is telling. — In fact here is a paper on this same effect in humans and dogs (known since the 1800s - https://hackertalks.com/post/26465788)… Performing this test this way, and making blanket statements about impaired insulin function is negligent and perhaps malicious. It encourages lay people to run around saying keto will give you diabetes… which is patently false.

    Liver dysfunction - mechanisms we don’t see in humans on keto - in fact human studies show the opposite - total resolution of NAFLD, and the researches replicated the same issue on a non-keto 90% fat diet in mice… that suggests keto is not the key mechanism of action in this mouse study… so the mice are behaving totally different then the humans.

    MD Schur also covered this paper in a 3m easy to watch video - https://youtu.be/-eA42CbEB8k

    summerizer

    Core conclusion

    • Recent headlines link a long-term ketogenic diet in mice with fatty liver, diabetes, and shorter lifespan, but those mouse findings do not mean ketogenic diets are dangerous for people.
    • The mouse paper has major limits for human use, because the core mistake is assuming mice and humans are metabolically interchangeable.

    Mouse study and headline problem

    • Genetically modified mice received a 90% fat, 10% protein, essentially 0% carbohydrate diet for a long period relative to a mouse lifespan.
    • Biochemical markers of organ stress, altered metabolite profiles, increased triglycerides, reduced survival in some groups, and disrupted metabolism occurred.
    • Sustained ketosis may carry hidden risks when maintained continuously over a long duration, especially in mice.

    Why mouse data does not translate cleanly

    • Rodent research is understandable because 20- or 30-year human nutrition studies are not realistic, but the timeline conversion requires a leap of faith.
    • A few months in rodents cannot automatically equal many decades in humans across every health parameter being measured.
    • Human metabolism includes adaptive flexibility, redundancy, repair systems, lifespan buffers, and other features that may keep mouse stresses from matching human experience.

    Diet design does not match human keto

    • The mice were likely under constant rigid ketosis on a very high-fat, low-protein, nearly zero-carbohydrate diet.
    • That diet does not match the ketogenic diet most humans eat, especially a real-food ad-lib keto diet that often becomes nutrient dense and naturally calorie reducing.
    • Synthetic feed components and potential nutrient imbalances make the mouse diet even farther from ordinary human keto practice.

    Mouse genetics and species differences

    • Lab mice are often inbred, genetically uniform, or altered strains with peculiar vulnerabilities.
    • A strain-level weakness in high ketone production or lipid metabolism could dominate one study without applying to healthy humans or even other mice.
    • Humans regulate ketosis through hormonal feedback involving insulin, glucagon, cortisol, thyroid signals, adiponectin, meals, fasting, exercise, stress, and sleep.
    • Mice produce ketones differently and have different rates of substrate use and tissue response, so ketosis in mice is not ketosis in humans.

    Human evidence runs the other way

    • Human clinical trial data supports the safety and efficacy of ketogenic and low-carbohydrate diets in metabolic disease.
    • In type 2 diabetes and insulin resistance, ketogenic diets consistently reduce A1C and insulin levels and can lead to medication discontinuation with disease remission.
    • In fatty liver disease, carbohydrate restriction reduces liver fat and improves liver enzymes, which is the opposite of the mouse finding.
    • Human ketogenic interventions routinely reduce triglycerides, while the mouse data had increased triglycerides.
    • Ketogenic and low-carbohydrate diets also have benefits in PCOS, weight loss, metabolic improvements, epilepsy, psychiatric disorders, and other metabolic conditions.

    How to use the mouse paper

    • The mouse paper can be used as a weak hypothesis generator, but it should not become a guide for human nutrition.
    • Mechanistic findings can be interesting, but human clinical trials measuring real-world health outcomes should carry more weight.
    • Sensational headlines saying keto kills, causes liver disease, or is dangerous should not guide dietary decisions when they rely on rodent data that does not reflect human physiology.
    • When human data contradicts a rodent signal, the human signal should take precedence.
    • A headline based on a mouse study should prompt the questions: does this match human experience, and do human trials show the same pattern or the opposite pattern?
    • Dietary guidance should be based on humans eating real food and living their lives, not on mice living in cages.

    References

    • jet@hackertalks.comOPMEnglish
      3·
      10 days ago

      The paper spends a tremendous amount of space on imaging, imagine techniques, and image analysis - It feels like someone’s PhD thesis is on imaging and this is where they were going to publish it - hell or high water. It’s just odd.

      This paper reads like a hit piece trying to add noise to the keto metabolic discussion. Enough things go the wrong way, and are omitted, or done incorrectly, that this smells of malice rather than naive incompetence. They certainly choose a sensational title.