When calories match, fasting’s metabolic edge mostly disappears

Key takeaways

• In a German trial published in 2025, 31 women ate matched meals in an early or a late eight-hour window for two weeks each, and neither schedule changed insulin sensitivity, blood sugar or blood fats in a meaningful way.
• Fasting does switch the body toward fat and ketones, which start rising 8 to 12 hours into a fast, but the evidence that fasting boosts autophagy in people is thin and exploratory.
• In a 16-week trial of 405 Chinese adults with early type 2 diabetes, a 5:2 meal-replacement plan lowered A1C by 1.9 points, slightly more than metformin or empagliflozin, along with more weight loss.
• In a year-long trial of 88 adults with fatty liver, adding an eating window to the same calorie limit didn’t shrink liver fat any further.
• If you take insulin or a sulfonylurea, fasting can push blood sugar dangerously low, so make a plan with your clinician first; the National Alliance for Eating Disorders helpline is 1-866-662-1235.

In a study published in 2025, researchers in Potsdam and Berlin, Germany, ran a careful test of one of fasting’s favorite ideas: that when you eat matters as much as how much¹⁰.

Thirty-one women with overweight or obesity tried two schedules, two weeks apiece. On one, every bite fell between 8 in the morning and 4 in the afternoon. On the other, eating ran from 1 in the afternoon to 9 at night. The food itself stayed nearly the same in calories and nutrients¹⁰.

Their metabolism shrugged. Neither schedule moved insulin sensitivity, blood sugar, blood fats or inflammation markers enough to matter clinically¹⁰. Their body clocks did respond: after the late schedule, a clock measured in blood cells ran about 40 minutes behind, and the women went to bed and got up later¹⁰.

Olga Pivovarova-Ramich, who led the work, said the gains reported in earlier eating-window studies probably came from people eating less without intending to¹⁰.

That finding captures fasting science in 2026. Its weight-loss record is settled: it roughly ties ordinary calorie cutting, as our intermittent fasting article explains. The open question is whether fasting does something extra for your metabolism. This guide covers what trials show, and who needs a plan before trying it.

Your body really does switch fuels when you fast

In 2019, Rafael de Cabo and Mark Mattson, researchers at the National Institute on Aging, laid out the scientific case for fasting in the New England Journal of Medicine¹. Their central idea was “metabolic switching.”

Between meals, your body runs mainly on glucose, much of it released by the liver. After roughly 10 to 14 hours without food, the liver’s sugar reserve runs low, the review says¹. The body then leans on fat, and the liver turns some of it into ketones, a backup fuel. In people, ketone levels start rising 8 to 12 hours into a fast¹.

That switch is real, well-described physiology. The authors went further. They argued that several of fasting’s benefits, including better blood sugar control and lower blood pressure, don’t depend on weight loss¹.

They were candid about limits, too. Most human trials had lasted only months, mainly in overweight young and middle-aged adults, and no one knew whether people could keep fasting up for years¹. The institute’s own research program supported the review¹.

The test since then: do those extra benefits show up when a trial keeps calories or weight loss equal?

Autophagy is real biology, and in people, it’s barely been measured

Autophagy, from the Greek for “self-eating,” is the cell’s recycling system. It breaks down worn-out parts so they can be reused. Fasting fans often say a long gap between meals switches it on.

The 2019 review said fasting stimulates autophagy, but the studies behind that statement were in animals and cells¹. Measuring it inside a living person is hard.

Researchers in Adelaide, Australia, are trying, using immune cells drawn from the blood. In a 2025 exploratory analysis in the Journal of Physiology, they studied 121 adults with obesity assigned to standard care, daily calorie cutting, or intermittent fasting with an eating window for six months².

At two months, the groups didn’t differ. At six months, the fasting group’s autophagy measure differed from standard care’s, but only in an after-the-fact analysis that barely cleared the usual bar for significance, and partly because the standard-care group’s measure fell². Calorie cutting alone didn’t differ from standard care².

The same team found a surprise in 42 healthy adults: a high-protein meal after an overnight fast didn’t change autophagy in blood cells an hour later, though animal research predicted it would³. The researchers concluded that lab findings can’t simply be assumed to apply to people³.

So fasting might nudge autophagy in people. Nobody yet knows in which tissues, by how much, or whether it improves health. A promise that a 16-hour fast “cleans your cells” runs well ahead of the data.

When calories match, blood sugar and insulin results mostly tie

The TREAT trial, run by the University of California, San Francisco, compared an eight-hour eating window with three meals a day, with no calorie advice for either group⁴. The authors picked a window that mimics skipping breakfast⁵. Our article on extreme fasting covers its weight results.

The blood work is the part that matters here. Fasting glucose, fasting insulin, insulin resistance, A1C, triglycerides and cholesterol didn’t differ between groups, and didn’t even change significantly within either one⁴⁵.

Muscle was a different story. Among volunteers scanned in person, lean tissue in the arms and legs fell 0.64 kilograms (about 1.4 pounds) in the window group, compared with 0.17 kilograms in the three-meal group⁵. The authors urged caution about eating windows for that reason⁴.

Pooled trials point the same way. A 2023 meta-analysis in Frontiers in Nutrition combined 20 trials of 1,511 people with features of metabolic syndrome, a cluster of risk factors such as high blood sugar and a large waist⁹.

Compared with daily calorie cutting, intermittent plans took off about 0.8 kilograms more weight. Fasting glucose, fasting insulin, blood pressure, LDL cholesterol and triglycerides didn’t differ⁹. HDL, the “good” cholesterol, rose very slightly, about 1 mg/dL by our math⁹. Fasting groups also lost about half a kilogram more fat-free mass, which includes muscle⁹.

The American Diabetes Association’s 2026 Standards of Care agree. A 2024 meta-analysis they cite found fasting no better than daily calorie cutting, apart from slightly more waist loss and more loss of fat-free mass¹².

A 5:2 plan beat two diabetes drugs in one trial. Read the fine print

In China’s EARLY trial, published in JAMA Network Open in June 2024, researchers at nine medical centers randomly assigned 405 adults with type 2 diabetes to metformin, to empagliflozin, or to a 5:2 plan⁶⁷.

Participants had been diagnosed within the past year and had an A1C between 7% and 9%⁷. The 5:2 group ate very little on two nonconsecutive days a week, using meal-replacement products. On the other five days, a meal replacement took the place of dinner⁷.

After 16 weeks, A1C fell 1.9 points with the 5:2 plan, 1.6 with metformin and 1.5 with empagliflozin⁶. A1C below 7% was reached by 88.9% of the fasting group, 73.9% of the metformin group and 70.6% of the empagliflozin group⁷. By our math, that’s about 15 more people per 100 reaching the goal than with metformin.

The fasting group also lost 9.7 kilograms (about 21 pounds), compared with about 5.5 to 5.8 kilograms on the drugs⁶.

Now the fine print:

• Weight loss differed. The fasting group lost far more weight, so the trial can’t separate the effect of fasting from the effect of losing more weight.
• It was short. Sixteen weeks of treatment, in Chinese adults early in their diabetes. The authors described the benefit as short-term⁶.
• Lows still happened. Eight of 135 people on the 5:2 plan reported symptoms of low blood sugar, about the same as on metformin⁷.
• Funding is unclear. The summaries we could read didn’t name the trial’s funder or say who made the meal replacements⁷.

For a reader who prefers food over pills, this is encouraging. It’s also one trial, with medical supervision and packaged meals.

Fatty liver responds to eating less, not to the clock

In the TREATY-FLD trial, published in 2023, researchers at Southern Medical University in Guangzhou, China, put 88 adults with obesity and fatty liver on the same calorie limit for a year⁸. About half also had to eat only between 8 a.m. and 4 p.m.⁸.

MRI scans showed liver fat falling by almost exactly the same amount in both groups at six months. At a year, the drop was slightly smaller with the window than without it, a gap small enough to be chance⁸. Fatty liver went away in 33% of the window group and 49% of the calorie-only group, also not a statistically reliable difference⁸.

Both groups lost about 8 kilograms. Blood pressure and cholesterol improved about equally⁸. One measure of insulin resistance did fall more with the window at 12 months⁸, a signal worth watching.

The funding came from Chinese government research programs and the university, and the lead author reported no relevant financial ties⁸. The same team’s earlier weight-loss trial is covered in our extreme fasting article.

Early, late or no breakfast: timing is still an open question

The German trial in our opening is the cleanest test yet of timing alone. With matched meals, an early window and a late window produced the same metabolic results over two weeks¹⁰. Its funding, listed in the published protocol, included the German Research Foundation and a German Diabetes Association award supported by Abbott, a maker of glucose sensors¹¹.

Skipping breakfast hasn’t shown a metabolic payoff either. TREAT’s window, designed to mimic it, didn’t move blood sugar, insulin or blood fats⁴⁵.

Earlier eating may still deserve study. The ADA’s 2026 Standards say some preliminary studies show heart and metabolic benefits when food is eaten earlier, but that the area lacks conclusive evidence¹². Our insulin sensitivity article describes the best-known example, a tiny 2018 trial in men with prediabetes.

For now, choose a window that fits your work, family meals and sleep. The one that helps you eat a little less, without strain, is the one most likely to help your numbers.

Who needs a plan before fasting

People who take insulin or a sulfonylurea. The NIH’s diabetes institute says fasting while taking glucose-lowering medicines raises the risk of lows¹³. The ADA’s 2026 Standards say people on insulin or drugs that push the pancreas to release insulin should be medically monitored while fasting¹².

Don’t stop or change a dose on your own. Your clinician can adjust the plan.

People who fast for faith. The ADA counts fasting for religious or cultural reasons as a major social risk factor for low blood sugar¹⁴. Its 2026 Standards make Ramadan a model of supervised fasting: a risk score from the International Diabetes Federation and the Diabetes and Ramadan International Alliance, and a treatment plan set well before the fast begins¹².

The ADA also notes that about half of lows during Ramadan happen in the last few hours of the fast, so it suggests avoiding hard exercise then¹². Our article on faith and health explains the federation’s risk groups.

People who are pregnant or breastfeeding. Trials such as the German one excluded them¹¹, so they offer no evidence about safety in pregnancy. Ask your prenatal provider first.

People with an eating-disorder history. The German trial excluded this group too¹¹. Fasting rules can blur into restriction, so talk with your treatment team first. The National Alliance for Eating Disorders runs a toll-free helpline at 1-866-662-1235, weekdays from 9 a.m. to 7 p.m. Eastern, staffed by licensed clinicians who offer support and referrals; it isn’t a crisis line¹⁵.

Older adults and anyone protecting muscle. Both TREAT and the pooled trials found extra lean-tissue loss with fasting⁵⁹. Our articles on protein and on strength training explain how to hold on to muscle.

Fasting and your numbers, rated

This tool is our own, built from the trials above. Each row states a claim you’ll hear about fasting and one of your numbers; the rating says how well trials back it.

The claim	What fasting trials show	Rating	Safety note
A 5:2 plan lowers A1C in early type 2 diabetes	5:2 with meal replacements cut A1C 1.9 points in 16 weeks, slightly more than two drugs, with more weight loss67	Promising but early	Medically supervised; lows still occurred7
An eating window improves insulin resistance by itself	No gain over three meals in TREAT; no change with early or late windows when meals matched410	Evidence says no, or unsafe	Fasting on insulin or a sulfonylurea needs a plan1213
A window shrinks liver fat more than a calorie cut alone	Similar drops with or without a window over a year8	Evidence says no, or unsafe	Eating less is what moved it8
Fasting lowers blood pressure more than daily calorie cutting	No difference across 20 trials9	Evidence says no, or unsafe	Pressure improved in both groups of the liver trial8
Fasting improves LDL and triglycerides more than daily calorie cutting	No difference; HDL up about 1 mg/dL9	Evidence says no, or unsafe	Our intermittent fasting article notes one trial where LDL rose
Fasting costs extra muscle	More arm and leg lean loss in TREAT; more fat-free mass lost in pooled trials59	Promising but early	Our protein and strength-training articles explain how to protect muscle

The bigger claims behind fasting rate this way:

Claim	Evidence level	What it rests on
Fasting shifts the body toward fat burning and ketones	Strong evidence	Well-described human physiology1
Fasting improves metabolism beyond what eating less does	Evidence says no, or unsafe	Matched-calorie and matched-weight trials mostly tie8910
Fasting boosts autophagy in people	Not enough evidence	Animal work; one exploratory human analysis123
Eating earlier in the day helps metabolism	Promising but early	Small studies; a 2025 matched-meal trial found no effect1012

An old practice, fairly tested

Fasting is one of the oldest health practices people have, and it costs nothing. The metabolic switch shows the body was built to handle gaps between meals¹. That’s worth respecting.

What the newer trials add is a humbling detail. Much of fasting’s power seems to flow through a simple route: it helps many people eat a bit less. That’s no small thing, and no supplement or app can sell it to you.

Respecting the body also means respecting its limits: protect your muscle, and bring your clinician in if you take medicines that lower blood sugar.

Questions readers ask

The bottom line

Fasting does change how the body burns fuel. But when trials hold calories or weight loss steady, fasting rarely does more for blood sugar, insulin, liver fat, blood pressure or cholesterol than eating less in an ordinary way. Autophagy in people remains mostly unmeasured, and some trials found a small extra cost in muscle.

A fasting plan can still be a good tool if it helps you eat less, and one trial in early type 2 diabetes is encouraging. If you take insulin or a sulfonylurea, are pregnant, or have had an eating disorder, make a plan with a clinician before you start.

Related reading

• Intermittent fasting ties ordinary dieting in trials, which can be good news
• Fasting windows help some people eat less, while long fasts add real risks
• One workout raises insulin sensitivity for about a day, and habits make it last
• Faith shapes how people eat, fast and decide, and your clinician should know
• Type 2 diabetes can go into remission, and staying there is the hard part

Sources

1. de Cabo R, Mattson MP. “Effects of Intermittent Fasting on Health, Aging, and Disease.” New England Journal of Medicine, 2019;381:2541–2551 (copy of the published article). Gwern.net
2. Bensalem J, Teong XT, Hattersley KJ, et al., Heilbronn LK, Sargeant TJ. “Intermittent Time-Restricted Eating May Increase Autophagic Flux in Humans: An Exploratory Analysis.” Journal of Physiology, 2025;603(10):3019–3032 (University of Adelaide repository record). digital.library.adelaide.edu.au
3. South Australian Health and Medical Research Institute. “High Protein Meal Fails to Suppress Autophagy” (on Singh S, Sargeant T, et al., JCI Insight, 2025). SAHMRI news, August 28, 2025. sahmri.org.au
4. PACE-CME. “No Difference in Weight Loss or Metabolic Markers With Time-Restricted Eating” (on Lowe DA, Wu N, Rohdin-Bibby L, et al., TREAT Randomized Clinical Trial, JAMA Internal Medicine, 2020). October 21, 2020. pace-cme.org
5. Agyepong A, Selfridge N. “Time-Restricted Eating, Weight Loss, and Metabolism” (review of the TREAT trial). Relias Media. reliasmedia.com
6. Guo L, Xi Y, Jin W, et al. “A 5:2 Intermittent Fasting Meal Replacement Diet and Glycemic Control for Adults With Diabetes: The EARLY Randomized Clinical Trial.” JAMA Network Open, 2024;7(6):e2416786 (CiteDrive abstract record). citedrive.com
7. Pharmacy Times. “Intermittent Fasting and Meal Replacements Improved Short-Term Glycemia” (on Guo L, et al., JAMA Network Open, 2024). 2024. pharmacytimes.com
8. The Hospitalist (MDedge). “No Added Benefit of Time-Restricted Eating for NAFLD” (on Wei X, Lin B, Huang Y, et al., TREATY-FLD Randomized Clinical Trial, JAMA Network Open, 2023;6(3):e233513). 2023. community.the-hospitalist.org
9. Xu R, Cao Y, Wang P-Y, Chen X-L, Tao D. “Intermittent Energy Restriction vs. Continuous Energy Restriction on Cardiometabolic Risk Factors in Patients With Metabolic Syndrome: A Meta-Analysis and Systematic Review.” Frontiers in Nutrition, 2023;10:1090792. frontiersin.org
10. German Institute of Human Nutrition Potsdam-Rehbrücke and Charité – Universitätsmedizin Berlin. Press release on Peters B, et al., “Intended Isocaloric Time-Restricted Eating Shifts Circadian Clocks but Does Not Improve Cardiometabolic Health in Women With Overweight,” Science Translational Medicine, 2025;17(822):eadv6787. Informationsdienst Wissenschaft, November 4, 2025. idw-online.de
11. Peters B, Koppold-Liebscher DA, Schuppelius B, et al., Pivovarova-Ramich O. “Effects of Early vs. Late Time-Restricted Eating on Cardiometabolic Health, Inflammation, and Sleep in Overweight and Obese Women: A Study Protocol for the ChronoFast Trial.” Frontiers in Nutrition, 2021;8:765543. frontiersin.org
12. American Diabetes Association Professional Practice Committee. “5. Facilitating Positive Health Behaviors and Well-being to Improve Health Outcomes: Standards of Care in Diabetes—2026.” Diabetes Care, 2026;49(Suppl 1) (course copy hosted by the Weitzman Institute). education.weitzmaninstitute.org
13. National Institute of Diabetes and Digestive and Kidney Diseases. “Low Blood Glucose (Hypoglycemia).” NIH, last reviewed July 2021. niddk.nih.gov
14. Guideline Central. “2026 ADA Diabetes Standards of Medical Care Clinical Guideline Summary” (summary of American Diabetes Association, Standards of Care in Diabetes—2026). guidelinecentral.com
15. National Alliance for Eating Disorders. “What Is an Eating Disorder Helpline?” April 14, 2023, accessed October 2026. allianceforeatingdisorders.com