Daniel Tawfik

My wife @ElanaMD is an unstoppable force. A month in a small hospital room—chemo, radiation, and a bone marrow transplant— and she still manages to smile through it all.

Insulin resistance is usually framed as a blood sugar problem. But in 369,447 adults, higher insulin sensitivity tracked with a 33% lower risk of dying from gastrointestinal cancer. It may be a growth problem too.

Researchers at University Hospital Leipzig, led by René Thieme and Chuang Yang, used the UK Biobank, a large study that recruited adults aged 37 to 73 between 2006 and 2010. Everyone in this analysis was cancer-free at the start. Average age was 56. They were followed for an average of 13.5 years, and 4,305 died of a gastrointestinal (GI) cancer: esophagus, stomach, colon and rectum, liver, or pancreas.

The measuring stick is called eGDR, short for estimated glucose disposal rate. It estimates how well your body responds to insulin, the hormone that moves sugar out of your blood and into cells. The gold-standard test, the clamp, is too complex to run on hundreds of thousands of people. So eGDR is calculated from three things: waist circumference, whether you have high blood pressure, and HbA1c (your average blood sugar over about three months). A bigger waist, high blood pressure, and higher HbA1c all push it down.

That simplicity is the point. No special lab needed, which is why it can be applied at this scale.

The team also tested two other insulin resistance markers built from blood fats and glucose: the triglyceride to HDL ratio and the TyG index. This let them ask whether the signal is specific to eGDR or common to any insulin resistance marker.

What they found:

• Each 1 standard deviation increase in eGDR (about 2.5 units) was linked to a 33% lower risk of pooled GI cancer death (sHR 0.67, 95% CI 0.61 to 0.74), after adjusting for age, sex, BMI, smoking, alcohol, diet, physical activity, diabetes, and more
• Site by site, the hazard ratios were 0.67 for colorectal, 0.68 for pancreatic, 0.65 for liver, and 0.66 for esophageal cancer, all still significant after correcting for multiple testing
• People who died of GI cancer started with an average eGDR of 7.1, versus 8.2 in the full cohort
• The link held when the team removed everyone with type 2 diabetes, and when they excluded deaths in the first two years
• The TyG index was only weakly linked to pooled GI cancer death (sHR 1.05), and the triglyceride to HDL ratio was not significant after correction (sHR 1.03)
• For context, the absolute risk is small: the 10-year cumulative risk of dying from any GI cancer was 0.69% across the whole cohort

Why would insulin sensitivity matter for tumors?

First, when cells stop responding well to insulin, the pancreas compensates by making more of it. Think of it like someone knocking on a door that no one answers. They knock louder. High circulating insulin is called hyperinsulinemia.

Second, insulin is not just a sugar hormone. It is also a powerful growth signal. It binds the insulin receptor and the receptor for IGF-1 (insulin-like growth factor 1), which switch on pathways called PI3K-AKT-mTOR and MAPK. In plain terms, these pathways tell a cell to grow, divide, and resist the self-destruct programs that normally remove damaged cells.

Third, many tumor cells carry extra insulin receptors or run these pathways hot. A body with chronically high insulin gives them a steady supply of the signal they want.

That's why this result makes sense to me. A marker of insulin sensitivity tracks with fewer GI cancer deaths, and the biology gives a plausible reason. Keeping insulin low is not only about glucose. The paper also point to chronic inflammation, oxidative stress, and immune dysregulation as additional routes, so insulin is unlikely to be the whole story.

Now the limits, because they matter here.

This is observational. It shows association, not cause. Insulin itself was never measured. eGDR is a surrogate built from waist, blood pressure, and HbA1c, so it also reflects belly fat and blood pressure, which carry their own cancer risks. Adjustment helps, but residual confounding is possible.

Everything was measured once, at baseline, so the study can't say whether improving eGDR over time lowers risk. Cancer stage and treatment were not available. The cohort is 94.8% White and skews healthier than the general population, so results may not travel to other groups.

The result also wasn't uniform. Gastric cancer lost significance after correction, and so did esophageal adenocarcinoma. The pooled number blends biologically different cancers, and the authors warn against reading it as one shared effect. A stronger association in people under 60 was exploratory and uncorrected.

So what does it mean for you?

It does not prove that lowering insulin prevents cancer. This study can't say whether changing your eGDR changes your risk. What it does show is that people who were more insulin sensitive at baseline were less likely to die of these cancers over 13.5 years.

The useful part is that eGDR is built from three numbers you can track and change: waist size, blood pressure, and HbA1c. Here is how I would think about each one. The strategies below come from the broader metabolic literature, not from this paper.

Waist size is a proxy for visceral fat, the fat packed around your organs. It is more metabolically active than fat under the skin and is tightly linked to insulin resistance. If your waist is the biggest drag on your number, fat loss is the most direct lever.

Muscle is the next one. Skeletal muscle is the largest place your body sends glucose after a meal. More muscle, and muscle that is regularly contracted, means more places for sugar to go without the pancreas having to shout. Resistance training builds that tissue. A walk after meals recruits it right when glucose is rising. Both lower the insulin your body needs to do the same job.

Then there is sleep. A few nights of short sleep can make healthy people measurably less insulin sensitive. Poor sleep is easy to overlook because it never shows up on the eGDR formula, but it feeds the same pathway.

Food matters mostly through the load it puts on insulin. Fiber, protein, and fewer refined carbohydrates and sugary drinks flatten the spikes that demand a large insulin response, and they make it easier to keep your waist in check.

One caution. eGDR is a dashboard, not a target. You can raise it by lowering blood pressure with medication, which is worth doing, but that is not the same as improving how your cells respond to insulin. The goal is the underlying biology, and the number is how you check on it.

If you want a practical starting point, a tape measure, a home blood pressure cuff, and an HbA1c test once a year give you the same three inputs the researchers used. In my view, protecting insulin sensitivity is one of the more practical levers available. It costs little, and it is tied to diabetes and heart disease risk too, so the payoff does not depend on the cancer link being causal.

Insulin was built to move sugar. In a body that has stopped listening, it may end up carrying a second message to cells that never needed it.

More exercise time means more fitness. For interval training, most of the gain arrives by about 11 minutes per session. A meta-analysis of 69 trials and 2,387 people maps exactly where the returns flatten.

The study, led by @felipe_mattioni and colleagues at the University Hospital of Tübingen and Liverpool John Moores University, was published in Sports Medicine.

The question is simple. We know high-intensity interval training (HIIT, hard bursts separated by recovery) improves fitness. We know steady moderate exercise (MICT, moderate-intensity continuous training) does too.

But how long does a session actually need to be?

Individual trials struggle to answer that. Each one tests one or two durations. So the authors pooled randomized controlled trials where adults were assigned to HIIT or MICT for at least 4 weeks, then modeled the relationship between minutes per session and the change in VO2max.

VO2max is the maximum rate your body can use oxygen during hard exercise. It's one of the strongest predictors of how long and how well you live.

The design matters because of how they did the math. Instead of asking whether HIIT beats MICT on average, they used a dose-response model. It's kind of like filling a glass. HIIT is a fast tap that fills most of the glass in the first few seconds and then just spills over the rim. MICT is a slow drip that keeps adding water for as long as you stand there.

A dose-response model estimates how much fitness improvement you get at each session length, so it can show where HIIT flattens out and whether MICT ever does.

What they found:

• HIIT improved VO2max by 4.36 mL/kg/min on average versus 2.59 for MICT, a between-group difference of 1.20 mL/kg/min (d = 0.38)
• HIIT reached 80% of its maximal gain at 11.4 minutes per session, while MICT needed 52.4 minutes, a 4.6-fold difference
• At the halfway mark, HIIT needed 5.8 minutes per session and MICT needed 36 minutes
• HIIT plateaued at about 23 minutes per session, with little added benefit beyond that
• MICT showed a straight-line response and never plateaued within the 63 minutes studied
• The curve shape did not differ statistically between healthy adults and people with elevated cardiometabolic risk (obesity, pre-diabetes, pre-hypertension, metabolic syndrome)
• HIIT and MICT produced comparable changes in body fat, blood pressure, triglycerides, and fasting insulin

That last bullet is the part most people skip.

VO2max responded differently to the two types of training.

Cardiometabolic risk factors did not. The same weeks of training lowered blood pressure and body fat by similar amounts whether people did a few hard minutes or a long, easy session.

Why would one outcome split and the others not?

The paper offers a hypothesis:

First, VO2max is largely limited by how much blood your heart can pump at maximum effort. That is cardiac output.

Second, hard intervals push the heart to its peak output and put high mechanical stress on the vessel walls. That stimulus appears to drive the central adaptations that raise VO2max, and it doesn't require much time because the stimulus is intense.

Third, metabolic improvements like insulin sensitivity and body composition are driven more by peripheral changes in muscle and by total energy use. You can get there by working very hard briefly, or by working moderately for a long time. Different routes, similar destination.

That's why the two modalities split on fitness and converge on risk factors.

Now the limitations, because they change how you should read the 11 minutes.

First, session time includes recovery. Across HIIT protocols, recovery made up a median of 50% of each session, with a range of 17 to 90%. When the authors counted only the hard work, 80% of the gain came from about 5 minutes per session. So 11 minutes of HIIT can mean wildly different amounts of actual effort. This is the biggest caveat, and the paper is upfront about it.

Second, the labels are blurry. Most trials set intensity as a percentage of max heart rate or VO2max, not relative to each person's metabolic thresholds. The same percentage can be easy for one person and hard for another. Some MICT was likely closer to heavy exercise than moderate, which may have made the MICT curve look more linear than it really is.

Third, the participants were mostly sedentary or low fitness. Well-trained people were underrepresented, so these curves tell us little about athletes, where the ceiling is probably lower.

Fourth, this is a pooled analysis of studies that each tested only a couple of doses. Nobody tracked the same people across many durations. The authors also note that the HIIT plateau could reflect differences in how longer protocols were designed rather than duration itself.

Fifth, the numbers are population averages. Some confidence intervals are wide, including the HIIT estimate of 11.4 minutes (95% CI 9.5 to 40.2). The subgroup comparison between healthy and at-risk adults also had wide intervals, so treat it as encouraging, not settled.

On the positive side, 59 of the 69 trials had low risk of bias in randomization, and adherence was about 89% in both groups.

So what does this mean for you?

If your goal is raising VO2max and time is short, a short, truly hard session captures most of what a long one would. If your goal is blood pressure, body fat, or insulin, the choice between HIIT and steady cardio matters less than doing one of them consistently. The paper stress that the two are complementary, and that enjoyment and sustainability should guide the mix.

The best session isn't the longest one. It's the one that matches the outcome you're after, and the one you'll still do next month.

Your body burns more energy responding to psychological stress than it does running a mile. That energy comes from mitochondria, and @MitoPsychoBio Martin Picard argues those ancient organelles don't just power the stress response, they actively shape it.

Picard and colleagues published a landmark review in Frontiers in Neuroendocrinology proposing a fundamentally new way to understand stress: as an energy-driven process centered on mitochondria.

The central premise is simple but profound. Every aspect of your stress response requires energy. Gene expression, hormone synthesis, neurotransmitter release, immune activation, tissue remodeling. None of it happens without ATP, and nearly all cellular ATP comes from mitochondria.

But the relationship goes deeper than mitochondria just providing fuel. The review identifies four interconnected ways mitochondria and stress interact.

First, stress responses increase energy demand across multiple biological levels. At the molecular level, turning genes on and off requires energy. Cells need energy to produce proteins. Tissues need energy to remodel. Your entire organism needs energy to mount behavioral and physiological responses to threats.

Think of it this way: when you perceive a stressor, your brain activates the hypothalamic-pituitary-adrenal axis and sympathetic nervous system. Those cascades trigger widespread changes. Your heart rate increases. Blood gets redirected. Immune cells mobilize. Glucose floods your bloodstream. All of that coordination requires substantial energy investment.

Second, mitochondria don't just consume stress hormones, they produce them. This is counterintuitive because we typically think of hormones coming from glands. But mitochondria synthesize and metabolize both glucocorticoids like cortisol and catecholamines like adrenaline.

The enzymes that convert cholesterol into cortisol are located in mitochondria. The same goes for the breakdown pathways. This means mitochondria actively participate in setting the hormonal tone of your stress response, not just responding to it.

Third, the relationship is bidirectional. Stress hormones alter mitochondrial function. Glucocorticoids can enhance mitochondrial energy production in some contexts while impairing it in others, depending on dose, duration, and tissue type.

Acute stress exposure tends to boost mitochondrial capacity, preparing cells for increased energy demand. Chronic stress does the opposite, progressively degrading mitochondrial function and reducing cellular energy capacity. This creates a vicious cycle where impaired mitochondria can't support adequate stress responses.

Fourth, and perhaps most compelling, experimentally manipulating mitochondrial function changes how organisms respond to psychological stress. Picard cites studies where researchers altered mitochondrial DNA or respiratory chain function in mice and observed completely different stress response patterns.

Mice with mitochondrial defects show altered glucocorticoid responses, different behavioral reactions to stressors, and distinct patterns of social behavior. This demonstrates mitochondria aren't passive energy suppliers, they actively regulate stress physiology.

The review also connects mitochondrial function to depression and other stress-related disorders. Depression involves metabolic dysregulation. People with depression often show altered glucose metabolism, insulin resistance, and changes in appetite and energy-seeking behavior.

These metabolic changes may reflect underlying mitochondrial dysfunction. When cellular energy production falters, the brain's capacity to maintain normal mood regulation and stress adaptation declines. The fatigue, cognitive slowing, and motivational deficits characteristic of depression may partly stem from inadequate energy supply to neural circuits.

Picard introduces the concept of mitokines, signaling molecules produced by mitochondria that influence physiology beyond the cell. These include reactive oxygen species, metabolites, peptides, and even mitochondrial DNA fragments that escape into the cytoplasm or bloodstream.

Mitokines can trigger inflammatory responses, alter gene expression in distant tissues, and influence whole-organism metabolism. They represent a signaling channel through which mitochondria communicate cellular energy status to the rest of the body.

Key points from the review:

• Energy flow, not just molecular composition, distinguishes living organisms from dead ones

• Mitochondria are endocrine organelles that both produce and respond to stress hormones

• Acute stress enhances mitochondrial function while chronic stress degrades it

• Mitochondrial defects alter stress reactivity, social behavior, and mental health

• Depression and metabolic disease may share mitochondrial dysfunction as a common mechanism

• Mitokines allow mitochondria to signal energy status across cells and tissues

• Sexual dimorphism in mitochondrial function may explain sex differences in stress responses

The paper also addresses sex differences. Females generally have higher mitochondrial content and better mitochondrial function than males in many tissues. This may contribute to sex-specific patterns in stress-related disorders, with women showing higher rates of depression and anxiety while men show higher rates of metabolic syndrome.

The implications reframe how we think about stress adaptation and pathology. If stress is fundamentally an energetic process, interventions that support mitochondrial health, exercise, adequate sleep, nutritional sufficiency, might protect against stress-related disease not just through psychological mechanisms but through direct effects on cellular energy capacity.

The framework also suggests new therapeutic targets. Rather than only modulating neurotransmitter systems or hormone levels, treatments might aim to restore mitochondrial function and energy metabolism in stress-responsive tissues.

This is one of the most important conceptual papers in stress biology. Picard makes the case that we can't fully understand psychological stress without understanding cellular energetics, and we can't understand cellular energetics without centering mitochondria.

The review calls for more research defining how mitochondria influence human stress responses across the lifespan. Most mechanistic work comes from animal models. Translating these insights into clinical applications requires better tools to measure mitochondrial function in humans and interventional studies testing whether improving mitochondrial health protects against stress pathology.

Stress isn't just neurochemical or hormonal. It's energetic. And the organelles managing that energy don't just keep the lights on, they help determine whether stress makes you stronger or breaks you down.

The Mediterranean diet lowered LDL 7% and triglycerides 22% in 8 weeks. But inflammation didn't move until people added two fasting days a week and weight started coming off.

That is the finding from a new Cell Metabolism trial led by @LuigiFontanaMD's group at the University of Sydney and Washington University in St. Louis.

The question sounds simple. Does the Mediterranean diet lower inflammation because of what's in it, or because people tend to eat less and lose weight on it? Most studies can't separate the two, because the diet and the weight loss arrive together.

Here is how this one pulled them apart.

The researchers randomized 53 adults, and 43 finished (23 Mediterranean, 20 control). Average age was 54.7, 77% were women, and everyone had a BMI of 22 to 28 with at least one metabolic risk factor, like borderline blood pressure or high LDL.

The trial ran 16 weeks in two phases.

Phase 1 (weeks 0 to 8): one group ate a Mediterranean diet. Every meal was cooked in a research kitchen and handed to them. The control group kept eating their usual Western diet. When the Mediterranean group started losing weight on their own, the team added calories (about 300 per day more than controls) to hold weight steady.

Phase 2 (weeks 8 to 16): both groups added two non-consecutive fasting days per week.

Why does this design matter? Because phase 1 tests diet quality with body weight held roughly constant. Phase 2 then lets weight fall. If the diet itself were anti-inflammatory, inflammation should have dropped in phase 1.

It didn't.

The primary outcome was high-sensitivity C-reactive protein (hsCRP), a protein your liver makes when inflammatory signals are running high. Doctors use it as a gauge of background, low-grade inflammation.

What they found:

• Phase 1, hsCRP did not change in either group, even though food diaries showed clearly different diets (the Mediterranean menus delivered 38 to 44 g of fiber a day, versus about 17 g for a typical US adult)
• Phase 1, triglycerides fell 22%, total cholesterol 6%, and LDL 7% relative to controls, with no meaningful weight loss
• Phase 2, hsCRP fell about 31% from week 8 in the Mediterranean group, and the group difference was significant (log-scale change of -0.37)
• From week 8 to 16, the Mediterranean group lost 3.63 kg versus 1.88 kg in controls
• By week 16, systolic blood pressure was 5.2 mmHg lower and diastolic 5.4 mmHg lower than controls
• Fasting glucose was 4.5 mg/dL lower than controls, and the inflammatory signals IL-1β and MCP-1 were also down

What does this say about intermittent fasting?

The fasting days worked as a way to create an energy deficit, not as a special anti-inflammatory signal. Good food at stable weight did not move hsCRP. Losing fat did. Weight loss, not diet quality alone, tracked with lower inflammation.

The trial can't prove that cleanly, because there was no fasting-only arm. But it fits the pattern.

How might this work? Here is the mechanism as the paper lays it out, plus my read on it.

First, cholesterol responds to the food itself. Unsaturated fats, fiber, and plant compounds act on lipid handling directly. In the paper, lipid improvements did not scale with weight loss. People who lost less than 5% of their weight improved about as much as people who lost more.

Second, inflammation seems to track energy balance. When fat cells grow too large, they become insulin resistant and start releasing inflammatory signals. Shrink them and both problems ease. As weight fell, leptin (a hormone fat cells release in proportion to how much fat you carry) dropped 4.2 ug/L more than in controls. That is consistent with this model, though the study did not measure fat cell size or insulin signaling directly.

Third, the immune system in the gut quieted down. The team took biopsies of the colon lining and found 288 genes that differed between groups at week 16. Many were involved in T cell activation, the machinery immune cells use to switch on. Since your liver makes CRP in response to these signaling networks, lower immune activity in the gut fits with lower CRP. The authors call this exploratory and hypothesis generating, and I agree.

One surprise. Gut microbiome composition barely changed (p = 0.24 for group differences in overall community structure). Whatever lowered inflammation here did not require a major reshuffling of gut bacteria.

Think of it like a sink that is overflowing. Better quality water does nothing while the tap is still running at full blast. Cutting the flow is what drains it.

Now the limits, because they matter here.

The design can't tell fasting apart from the energy deficit, the weight loss, or simply eight more weeks on the Mediterranean diet. The phase 2 effect could come from any of them.

The sample is small and narrow. 43 completers, mostly women, 85% White, and 83% with a college degree. That limits how far you can generalize.

Baseline hsCRP was higher in the Mediterranean group (1.43 versus 1.00 mg/L). Regression to the mean, where unusually high values drift back toward average on their own, can't be fully ruled out.

The weight story is also less tidy than the headline suggests. Participants who lost less than 5% still showed an hsCRP drop in phase 2. Adjusting for body weight left the hsCRP result intact but weakened the cytokine findings. So weight loss looks important, but this trial does not prove it is the whole explanation.

Finally, it was 8 weeks per phase, blood was drawn two days after the last fast so ketones weren't measured, and the analysis covered completers only. Four people withdrew, all from the Mediterranean arm.

What does this mean for your own decisions?

Diet quality and energy balance appear to be separate levers. If your goal is LDL and triglycerides, what you eat matters, even without losing weight. If your goal is lower inflammation, blood pressure, and better glucose handling, this trial suggests you may need a calorie deficit on top of good food.

That also means the label matters less than the pattern. The best eating pattern is probably the one you can keep long enough to lower insulin and let fat cells shrink.

A healthy plate and a smaller plate are doing different jobs.