← Back to clock

THE RESEARCH LIBRARY

Fasting - the evidence.

What has been observed. What remains uncertain. A readable review of daily, intermittent and multiday fasting, with a closer look at cancer and autophagy.

Selected sources checked 24 October 2026 · Search coverage and independent review remain incomplete

Read the overview Explore 109 study records

FOLLOW THE EVIDENCE · 1969 to 2026

Explore the research timeline.

Follow a subject through time. Select a study to see who participated, what it tested, and how much its results can tell us.

109 publication records · 19 publication years

  1. 1969

    1. PeopleDoes your body still make sugar during a long fast?The liver and kidneys continued producing blood sugar as ketone levels rose.Owen et al.

  2. 1991

    1. PeopleCan fasting followed by a vegetarian diet ease arthritis?Symptoms improved in a small trial. The effects of fasting and the later diet could not be separated.Kjeldsen-Kragh et al.

  3. 2001

    1. PeopleCan a supervised water-fasting program lower blood pressure?Blood pressure fell in a residential program, but fasting’s effect could not be separated from other changes.Goldhamer et al.

  4. 2004

    1. Animals & labDoes fasting affect cell recycling equally in every organ?In mice, responses differed across tissues and changed over time. There was no single body-wide pattern.Mizushima et al.
    2. PeopleWhat happens to blood sugar during Ramadan fasting?People with diabetes reported more severe low-blood-sugar episodes during Ramadan than in other months.Salti et al.

  5. 2011

    1. Animals & labCan normal cell recycling help prevent liver tumors?Mice missing a recycling gene developed benign liver tumors. This was not a fasting experiment.Takamura et al.

  6. 2013

    1. Animals & labCan cell recycling also help lung tumors survive?Disabling a recycling gene reduced malignant tumors in mice, showing why more recycling is not always better.Guo et al.

  7. 2014

    1. People, animals & labCan fasting help the body rebuild immune cells?Most evidence came from mice. The small human observations did not demonstrate an immune-system reset.Cheng et al.
    2. PeopleWhat happens to muscle during a three-day fast?Signs of muscle protein breakdown increased. The study did not establish a peak time for cell recycling.Vendelbo et al.

  8. 2015

    1. PeopleCould a short fast reduce chemotherapy side effects?A thirteen-person breast cancer pilot found changes in some blood tests, not broad proof of protection.de Groot et al.

  9. 2016

    1. People, animals & labCould a fasting-style diet help with multiple sclerosis?A small human pilot found quality-of-life signals. Later diet changes make fasting’s contribution unclear.Choi et al.
    2. Animals & labCan brain tumors use ketones as fuel?Brain tumors in rats used ketones. A ketogenic diet did not clearly improve survival.De Feyter et al.
    3. PeopleHow well do selected patients tolerate fasting during chemo?A small supervised pilot tested feasibility. It could not establish general safety or cancer-treatment benefits.Dorff et al.
    4. PeopleCan men maintain strength while eating in a shorter window?Strength and lean mass were maintained in a small study of men who lifted weights.Moro et al.
    5. Animals & labCan cell recycling help cancer treatment work?Fasting improved chemotherapy response in mice, but the benefit depended on cell recycling and immune cells.Pietrocola et al.

  10. 2017

    1. People, animals & labCan lung tumors use fuels other than sugar?Human lung tumors used lactate for energy, showing that their fuel choices can vary.Faubert et al.
    2. People, animals & labBlood-cell recycling depended on how it was measuredSmall human experiments detected a fasting response with a special laboratory test, without establishing a universal clock.Pietrocola et al.
    3. PeopleIs eating less every other day easier or more effective?Weight loss was similar to daily calorie cutting, and more people dropped out.Trepanowski et al.
    4. PeopleCan a five-day fasting-style diet improve health markers?Weight and some risk markers improved. The trial did not test disease prevention or longer life.Wei et al.
    5. Animals & labCould a ketone help some skin cancers grow?One ketone promoted growth in a specific melanoma model. The result does not apply to every cancer.Xia et al.

  11. 2018

    1. PeopleCould short fasts ease chemotherapy fatigue?A small study suggested some symptom benefits, but many participants were not included in the final analysis.Bauersfeld et al.
    2. PeopleCan eating less twice a week help manage diabetes?Blood sugar results were similar to daily calorie cutting under medical supervision.Carter et al.
    3. PeopleWhat can go wrong during a supervised water fast?Medical records included serious problems requiring hospital care. Supervision did not remove all risk.Finnell et al.
    4. PeopleIs the 5:2 diet better than cutting calories every day?Both approaches produced similar weight changes after nearly a year.Schübel et al.
    5. PeopleCan eating earlier help without losing weight?A study of eight men found better insulin response and blood pressure. Larger trials are needed.Sutton et al.

  12. 2019

    1. PeopleDo diabetes improvements last after diet support ends?Blood sugar rose again in both diet groups. This was follow-up of the same trial, not a new study.Carter et al.
    2. PeopleDoes eating earlier affect the body’s cell recycling?Blood gene activity changed, but this did not show how much material cells actually recycled.Jamshed et al.
    3. PeopleWhat do people experience during a long modified fast?A large supervised group reported weight and health changes. There was no nonfasting comparison group.Wilhelmi de Toledo et al.

  13. 2020

    1. People, animals & labA sugar-storage protein helped some kidney tumors growReducing GYS1 slowed laboratory tumors, but the protein’s signaling role complicates the fuel explanation.Chen et al.
    2. PeopleDoes a fasting-style diet help breast cancer treatment?Some tumor-response measures improved, but complete tumor clearance at surgery did not.de Groot et al. · DIRECT
    3. PeopleDoes eating from noon to 8 p.m. help with weight loss?A twelve-week trial found no clear extra weight loss compared with three structured meals a day.Lowe et al.
    4. PeopleDoes keto plus fasting improve brain cancer treatment?The diet raised ketones but did not clearly improve cancer control or survival.Voss et al. · ERGO2

  14. 2021

    1. PeopleCan people stick with the 5:2 diet for a year?Fewer people kept following the plan over time. Weight loss was similar to conventional advice.Hajek et al.
    2. PeopleCould gut bacteria help explain blood pressure changes?This smaller part of a larger trial found promising changes, but it was not independent confirmation.Maifeld et al.
    3. People, animals & labDo kidney tumors always need stored sugar?Removing the ability to make or use glycogen did not stop these kidney-cancer models growing.Xie et al.

  15. 2022

    1. People, animals & labDoes fasting change cell recycling in muscle?Muscle and liver measurements responded differently. The findings do not establish a whole-body effect.Chaudhary et al.
    2. PeopleDoes a short fast add to a healthy-lifestyle program?Some weight benefits appeared, but the two main blood pressure and insulin tests found no clear advantage.Cramer et al.
    3. People, animals & labCould a ketone slow bowel cancer growth?One ketone reduced tumors in mice through cell signaling. This was not proof that ketones starve cancer.Dmitrieva-Posocco et al.
    4. PeopleDoes combining fasting and exercise improve blood sugar?Some measures improved, but the main blood sugar test found no clear advantage.Haganes et al.
    5. PeopleDoes eating earlier help with weight loss?An early eating window added weight loss, but the extra body-fat loss was uncertain.Jamshed et al.
    6. PeopleDoes an eight-hour eating window add to weight loss?Adding a daily eating window did not clearly beat calorie cutting alone over one year.Liu et al.
    7. PeopleCan an eating window fit around 24-hour shifts?Selected firefighters shortened their eating windows. Some health measures improved, especially in higher-risk groups.Manoogian et al.
    8. PeopleWhat happens after a long water fast and refeeding?Some health markers improved and others worsened after people started eating again.Scharf et al.
    9. PeopleHow does a fasting-style diet affect immunity in cancer?Blood and immune changes were observed. Better cancer outcomes were not established.Vernieri et al.
    10. Animals & labKetogenic food helped chemotherapy in mice, even as tumors burned ketonesThe diet alone did not extend survival; the combination with chemotherapy did.Yang et al.

  16. 2023

    1. Animals & labCan a fasting-style diet help leukemia treatment?More mice survived without leukemia when diet and chemotherapy were combined. Reduced cell recycling helped explain the result.Buono et al.
    2. PeopleCan a daily eating window replace calorie counting?Both approaches helped with weight loss over a year. Neither clearly outperformed the other.Lin et al.
    3. People, animals & labSlower-growing cancer cells could still survive treatmentFasting-like conditions left some bowel-cancer cells dormant; an experimental drug combination increased killing.Liu et al.
    4. People, animals & labKetones helped these stressed lung-cancer cells growBlocking a ketone-processing enzyme reduced growth in laboratory and mouse experiments.Lu et al.
    5. PeopleCan an eating window help with weight and type 2 diabetes?It helped compared with usual eating, but did not clearly beat a calorie-cutting diet.Pavlou et al.
    6. PeopleDo blood sugar benefits from fasting last?An advantage at six months was no longer clear at eighteen months.Teong et al.
    7. PeopleDoes an eating window help reduce liver fat?Liver fat fell with calorie cutting. Adding an eating window did not clearly help further.Wei et al.
    8. PeopleCan a structured diet put type 2 diabetes into remission?More people met the study’s remission criteria. That does not mean a permanent cure.Yang et al.

  17. 2024

    1. PeopleA fasting-style diet showed promise during breast chemotherapyA 44-person trial found better tumor response and some fewer severe side effects, with important caveats.Bahrami et al.
    2. Animals & labDo fasting mice live longer?Fasting and calorie cutting extended lifespan in female mice. This does not establish longer life in people.Di Francesco et al.
    3. PeopleLymphoma improved in three reported fasting casesScans improved after prolonged fasting and dietary changes, but three cases cannot establish the cause.Gabriel et al.
    4. PeopleHow does a meal-replacement fasting plan compare with medication?A structured food program improved early diabetes measures, but the benefit cannot be assigned to fasting alone.Guo et al.
    5. People, animals & labA natural molecule helped connect fasting with cell recyclingSpermidine rose during fasting; experiments in cells and other organisms tested its role in recycling and longevity.Hofer et al.
    6. Animals & labWhat happens in the gut when eating resumes after a fast?Refeeding boosted gut repair in mice, but also tumor growth in a model with an altered cancer-related gene.Imada et al.
    7. PeopleLonger overnight fasting improved sleep, but not clearly memoryA small trial separated a favorable sleep result from uncertain cognitive benefits.James et al.
    8. PeopleLonger survival was observed in a small fasting-diet groupFourteen women receiving diet plus chemotherapy lived longer than a separate comparison group; selection could explain part of the difference.Ligorio et al.
    9. PeopleDoes intermittent calorie cutting work better for teenagers?It did not outperform daily calorie cutting in a year-long trial with intensive clinical support.Lister et al.
    10. PeopleCan a shorter eating window improve blood sugar?A three-month trial found a small improvement compared with standard care alone.Manoogian et al.
    11. PeopleDoes meal timing matter when calories stay the same?With food provided, an earlier eating window did not produce extra weight loss or better blood sugar.Maruthur et al.
    12. PeopleWhat changes during a week without food?Blood proteins changed during a seven-day water fast. Those changes do not prove health benefits.Pietzner et al.
    13. PeopleA ten-hour eating window did not clearly improve weight lossThe main comparison in this 100-person trial found little added benefit, including after follow-up.Quist et al.
    14. PeopleMonthly low-calorie cycles reduced diabetes medication needsA 100-person trial found better glucose management after a year of supervised fasting-style diet cycles.van den Burg et al.

  18. 2025

    1. PeopleA very long water fast was followed by serious vitamin deficiencyOne man developed confusion, walking problems and double vision after a 40-day fast.Aweer et al.
    2. PeopleSome weight loss lasted two years; the glucose advantage did not clearly lastFollow-up separated a lasting weight difference from uncertain diabetes benefits.Azhar et al.
    3. PeopleFasting did not beat a healthy diet on the main MS outcomeAn 18-month trial found no clear advantage for new brain lesions, despite some promising secondary results.Bahr et al.
    4. PeopleDoes intermittent fasting increase cell recycling?A small blood-cell study found an uncertain difference after six months, not a body-wide fasting timetable.Bensalem et al.
    5. PeopleFor night workers, fasting did not clearly beat daily calorie cuttingAll three programs helped some participants lose weight, with no clear advantage on the main comparisons.Bonham et al.
    6. PeopleDoes eating less three days a week help with weight loss?People lost a little more weight than with daily calorie cutting, with intensive support.Catenacci et al.
    7. PeopleAre short eating windows linked to heart-related deaths?An observational study found a link, but could not show that intentional fasting caused it.Chen et al.
    8. People, animals & labCan stored sugar help lung tumors grow?Blocking sugar storage suppressed tumor development in mice. Fasting as a human treatment was not tested.Clarke et al.
    9. PeopleDoes a long fast always reduce inflammation?Inflammation markers rose during this small supervised study, even as people lost weight.Commissati et al.
    10. PeopleDoes the timing of an eating window change belly fat?Early, late and self-chosen windows did not clearly add to loss of fat around internal organs.Dote-Montero et al.
    11. PeopleCan a fasting-style diet change cell recycling?Blood-cell results differed between diet formulations. The small study did not establish a health benefit.Espinoza et al.
    12. PeopleHow much weight lost during a long fast is body fat?Early weight loss included substantial lean mass, which partly returned after eating resumed.Gabriel et al.
    13. PeopleCan eating less on some days help during breast cancer care?An early trial suggested longer cancer control, but the result remains uncertain.Harvie et al.
    14. People, animals & labA ketone helped pancreatic cancer spread in miceThe effect involved a growth-related protein; patient observations did not establish a diet effect.Jiang et al.
    15. Animals & labSome cancers used ketones to build new cellsTracing experiments found a route from ketones to the building materials needed for tumor growth.Kaluba et al.
    16. PeopleA fasting-style diet was tested before breast-cancer surgeryThirty patients received the diet with chemotherapy; tumor changes were promising but the abstract provides no usual-diet comparison.Ligorio et al.
    17. PeopleDo people with higher ketones develop less cancer?A large observational study did not establish a protective relationship or test a fasting diet.Luo et al.
    18. Lab cellsA ketone made some breast-cancer cells harder to treatCells exposed to BHB became more resistant to one chemotherapy drug in laboratory cultures.Mashinchian et al.
    19. Lab cellsKetones could not replace glucose for these childhood-cancer cellsThree cell lines failed to keep growing in glucose-free conditions when ketones were supplied.Muto-Fujita et al.
    20. PeopleA shorter eating window reduced liver fat about as much as calorie cuttingA larger trial found benefit compared with usual care, with similar results from daily calorie restriction.Oh et al.
    21. PeopleA liver study measured a cell-recycling markerA blood protein changed after eight weeks; this did not reveal when autophagy started or how much recycling occurred.Ozlu Karahan et al.
    22. PeopleCould an eating window help older people with memory problems?Some test scores improved in a small pilot. Dementia prevention was not established.Wang et al.
    23. PeopleA fasting-like diet during chemotherapy was feasible, but symptoms matteredA small study recorded severe symptoms during some diet cycles and did not test whether cancer outcomes improved.Xue et al.

  19. 2026

    1. PeopleAn early breast-cancer trial checked safety firstFourteen patients tried nightly fasting with metformin; cancer-effectiveness results were not reported.Briata et al.
    2. PeopleA small cancer trial found immune changes, not proven tumor controlKetogenic meals before surgery changed blood markers and immune-cell measurements in endometrial cancer.Dantas et al.
    3. Lab cellsPreprint · not peer reviewedCan hungry cell layers push out damaged cells?Lab-grown cells expelled damaged neighbors. This early finding has not been demonstrated by fasting in people.Gates et al.
    4. PeopleCould people fast before bowel-cancer chemotherapy?Most of the six fasting participants followed the plan; the trial was too small to establish treatment benefits.Herbert et al.
    5. PeopleHow practical are eating windows for older adults?A small trial tested feasibility and function. It could not establish which diet worked best.Houston et al.
    6. PeopleCould a keto diet help pancreatic cancer treatment?A small trial alongside chemotherapy had encouraging but uncertain survival results.Jameson et al.
    7. PeopleA five-day modified fast left some weight change, while many blood changes fadedA small controlled study followed healthy adults for twelve weeks after fasting.Kaufhold et al.
    8. PeopleLonger follow-up did not show a clear cancer-survival benefitPatients who chose ketogenic food during radiation were followed for about six years.Klement et al.
    9. PeopleDoes a shorter eating window improve memory and focus?Healthy women improved with repeated testing in both groups. Fasting added no clear benefit.McLeod et al.
    10. People, animals & labCan fasting make breast cancer hormone treatment work better?Mouse findings suggest a possible mechanism. Small human samples do not establish a treatment benefit.Padrão et al.
    11. PeopleCan an eating window work as well as dietitian advice?An early blood sugar comparison met the study’s target. The one-year comparison did not.Parr et al.
    12. PeopleDoes eating less protein increase cell recycling?Blood-cell measurements showed no clear increase. This was a protein study, not a fasting trial.Singh et al.
    13. People, animals & labCan fasting help immune cells attack bowel cancer?Mouse tumors grew less, with changes in a nutrient and immune activity. Human tissue observations did not test a fasting treatment.Song et al.
    14. People, animals & labPancreatic tumors could switch on ketone useA cancer mutation and additional signals together helped cells turn ketones into usable energy.Wang et al.
    15. Animals & labA ketone slowed thyroid-cancer growth, with cell recycling involvedThe experiments showed growth arrest, without a clear increase in programmed cell death.Wang et al.
    16. PeopleCan a ten-hour eating window add to weight loss?People lost more weight, but a late change to the trial’s main outcome makes the result less certain.Wilkinson et al.
    17. PeopleTwo diets helped men under prostate-cancer surveillance lose weightBoth 5:2 fasting and daily calorie reduction worked in a 20-person pilot; cancer progression was not established.Yeary et al.
    18. Animals & labKetones helped bladder-cancer treatment in miceA ketogenic diet or BHB slowed tumors and strengthened an immune-based treatment in this model.Zeng et al.
    19. Animals & labA cancer protein mattered for more than storing sugarRemoving GYS1 slowed breast-cancer growth, while other ways of lowering glycogen did not.Zhuo et al.

Headlines summarize the research in plain language; full records retain the original titles. Publication years are grouped, not spaced to a calendar scale. Records include companion reports; they are not a count of independent trials.

Research findings describe groups under specific conditions. They do not predict your results or establish a safe fast duration. Fasting or ketosis should not replace cancer treatment; dietary changes during treatment belong with the oncology team. [121]

01 / Start here

The evidence in a minute

The clearest human evidence concerns weight loss and some metabolic outcomes. Evidence for treating cancer, extending human life, or timing autophagy is much less settled. The labels below are editorial assessments of this selected evidence, not formal certainty grades.

Better studied in humans

Weight & metabolic health

Some supported fasting programs help adults lose weight or improve glucose measures. Superiority over other calorie-reduction approaches is inconsistent. [1] [2] [5]

Early, mixed clinical evidence

Cancer treatment

Small trials suggest possible adjunctive benefits. They do not establish that fasting or ketones alone starve and kill human cancers. [20] [21] [23] [24]

Mechanism under investigation

Autophagy & longevity

Human cellular measurements are emerging. They do not provide a universal onset, peak, duration, or demonstrated lifespan benefit. [27] [28] [34]

02 / Fasting approaches

Different approaches, different evidence

“Fasting” is not one intervention. These descriptions explain the protocols in the cited studies; they are not instructions or recommended targets.

What the studies actually tested
ApproachWhat changesEvidence boundary
Daily time-restricted eatingFood is eaten within a daily window; calories may or may not be prescribed.A six-hour early window is not interchangeable with a multiday fast.
Alternate-day / 5:2 / 4:3 restrictionWhole days alternate between ordinary intake and zero or reduced calories.Many trials allow food on “fast” days. Read the actual protocol.
Multiday water-only fastingNo food energy for consecutive days; water is permitted.Selected supervised cohorts cannot establish home safety or an optimal duration.
Modified multiday fastingSmall energy intake, often juice or broth, for consecutive days.Results cannot automatically be assigned to water-only fasting.
Fasting-mimicking diet (FMD)Formulated low-energy, often low-protein diets in repeated short cycles.Diet composition, calorie deficit and refeeding all contribute.
Ketogenic dietCarbohydrate restriction with ongoing food intake.Ketosis is not synonymous with fasting; cancer trials using it test a different intervention.

Sources [3] [7] [10] [6] [24]

Religious fasting, prolonged dry fasting, fasting before procedures and continuous calorie restriction also differ. Religious fasting is addressed separately below; procedural and prolonged dry fasting are not comprehensively evaluated. Evidence from water-permitted trials cannot validate withholding fluids.

03 / Potential benefits

Where human studies are most informative

Weight loss: a useful option for some people

A year-long 4:3 trial found a modest advantage over daily restriction. Other year-long trials found no clear advantage for daily eating windows or alternate-day restriction. This supports choice of approach, not a rule that more fasting produces more benefit. [1] [2] [3]

Glucose regulation: promising, with important context

A small controlled-feeding experiment suggests meal timing can influence insulin sensitivity without intended weight loss. A separate diabetes trial reported remission with a structured restricted-calorie diet. These answer different questions; neither isolates a universal fasting “reset.” [4] [5]

Blood pressure and other risk markers: mixed

FMD studies reported favorable risk-marker changes. In a multiday-fasting program, an encouraging microbiome substudy should be read alongside the larger trial’s null primary blood-pressure and insulin-resistance results. Changes in a calculated risk score are not observed prevention of disease. [6] [8] [7]

Symptoms and inflammation: condition-specific

An older rheumatoid-arthritis trial reported symptom improvements after fasting followed by a year of dietary change. That combined program cannot tell us how much fasting contributed. Conversely, one prolonged-fasting cohort showed increased inflammatory markers. “Fasting reduces inflammation” is too broad. [9] [12]

Multiday fasting: physiology is better mapped than lasting benefit

Water-only studies report weight and biochemical changes, but small uncontrolled cohorts and changed refeeding diets limit causal claims. The studies reviewed here do not establish that a longer fast is better than a shorter one, or that short-term weight loss represents only fat. [10] [11]

Longevity, brain health and regeneration: unanswered clinical questions

A large mouse experiment supports lifespan effects in that species. The selected human studies measure risk markers and cellular processes, not added years of life, dementia prevention or a whole-body rejuvenation outcome. Those clinical benefits should remain hypotheses. [34] [6] [27]

More active-control trials help put the benefits in perspective

In a 337-person fatty-liver trial, time-restricted eating and daily calorie restriction produced similar relative liver-fat reductions over sixteen weeks. In 250 night-shift workers, fasting did not clearly outperform daily restriction on the main weight and insulin-resistance comparisons. A two-year follow-up in insulin-treated diabetes found a persistent weight difference, but not a statistically clear advantage for HbA1c or insulin dose. These studies support practical options while keeping each outcome and comparator distinct. [102] [103] [101]

Multiday fasting now has another small controlled study

LEANER assigned 38 healthy adults to an immediate five-day modified fast or a waiting period before the same intervention. Participants consumed 200–250 kcal/day from juice and broth. Some weight and fat reduction remained at twelve weeks, while many metabolic changes returned toward baseline. Mild, temporary symptoms were common. Without an active calorie-restriction comparison, the study cannot establish that prolonged fasting is the better approach. [109]

04 / Weight, glucose & liver

The comparator changes the conclusion

Weight loss: useful options, inconsistent superiority

A program can help people lose weight without outperforming another supported diet. Twelve-month comparisons by Liu and Lin and the 50-week HELENA trial did not establish a clear advantage over daily restriction. The 4:3 trial did find a modest additional benefit. These different results should remain visible, rather than be compressed into a promise that fasting always works better. [2] [36] [38] [1]

What happens when calories are held similar?

The 41-person TRIM trial provided isocaloric food and found no additional weight or glucose benefit from earlier restricted eating. Sutton’s much smaller controlled-feeding experiment found some insulin and pressure benefits without weight loss. Together they support testing specific timing effects, not assuming they explain every fasting-associated improvement. [42] [4]

Fat loss is not interchangeable with scale weight

The early-eating trial by Jamshed improved its primary weight outcome but did not clearly improve the main body-fat outcome. A later 197-person trial found no significant advantage for its primary MRI visceral-fat outcome across early, late or self-selected windows. Water, glycogen and lean tissue can contribute to scale changes. [39] [40] [68]

Diabetes: improvement, remission and durability

Pavlou found glycemic improvements with both time and calorie restriction. Carter demonstrated planned HbA1c equivalence at twelve months, but the later follow-up did not preserve that result. Teong’s early glucose advantage disappeared at eighteen months; Parr’s noninferiority criterion was met at four but not twelve months. These studies strengthen the case for follow-up, not permanent-cure language. [43] [44] [45] [46] [49]

Dietary packages and medication comparisons

The EARLY trial tested meal replacements and intermittent restriction against two medications. Its results cannot isolate the fasting schedule from supplied foods, calorie reduction or other dietary changes. The smaller CMNT trial supports a remission signal in a selected group, with ongoing clinical interpretation and monitoring. [48] [5]

Monthly fasting-style diets: a year-long diabetes trial

The 100-person FIT trial added five-day monthly low-calorie cycles to usual primary care. At twelve months, medication needs, HbA1c and weight favored the program. Participants initially used diet or metformin; there was no equally supported calorie-restriction arm. This supports a supervised treatment option in that population, without isolating fasting from weight loss or showing a universal diabetes cure. [93]

A practical eating window can also have a null result

In RESET, 100 adults at high diabetes risk tried a ten-hour eating window or continued habitual living. The main weight difference was small and uncertain after three months, with no sustained advantage at six months. A positive analysis restricted to completers should not replace that main finding. [94]

Fatty liver: improvement is not universal elimination

In TREATY-FLD, both calorie-restricted groups reduced MRI liver fat over twelve months. Adding an eight-hour window did not clearly improve the result. That supports a narrower claim about measured liver fat, not guaranteed disease elimination or proven prevention of cirrhosis. [50]

Adherence is part of effectiveness

In the pragmatic 5:2 study, reported adherence fell from 74% at six weeks to 22% at one year. Controlled-feeding experiments explain mechanisms; supported long-term trials better address whether a regimen remains usable. Both matter. [41]

05 / Multiday fasting

Longer fasting needs its own evidence

Water-only, juice/broth modified fasting and fasting-mimicking diets should not be pooled as though they were the same exposure. The large Buchinger cohort included calories and supervised activity; the hypertension water-fast cohort also included preparation and low-sodium plant-based refeeding. Neither had a randomized nonfasting control. [67] [64]

A dramatic before/after change is not a treatment effect

Residential selection, medication changes, rest, calorie deficit, reduced sodium and the refeeding diet can all influence outcomes. In an uncontrolled cohort, the change attributable specifically to fasting cannot be separated from those factors. Small proteomic experiments add mechanistic detail, not proof of fewer future diseases. [64] [10] [11]

Refeeding belongs in the result

The 2025 DXA analysis found marked early fat-free-mass loss and partial later recovery. Hydration and glycogen complicate interpretation, and a plant-based follow-up diet was part of the program. Another supervised cohort showed an inflammatory-marker increase during prolonged fasting. Benefit and harm should be assessed across fasting, refeeding and follow-up. [68] [12]

No validated “best length” emerges

The selected studies do not establish that seven, fourteen or twenty-one days maximizes net clinical benefit. A biochemical shift after three days is not a benefit threshold. Selected clinic observations also cannot establish safety for an unsupervised person. [10] [13] [67]

Rare serious harms belong in the evidence too

A case report describes vitamin-B1-related brain injury after a 40-day water-only fast. A lymphoma case series also recorded low potassium requiring a change to the fasting regimen. Case reports cannot estimate incidence, but favorable averages in selected clinics do not erase these possible harms. [77] [89]

06 / Cancer research recheck

What did the latest recheck change?

The 24 October 2026 assessment now covers all 928 candidates retained from the wider search: 840 evidence assessments, 36 exclusions after source checks, and 52 records whose original results were inaccessible. Thirteen additional studies are featured in the timeline. The searchable assessment appendix preserves findings and limitations for the full group. Most records were assessed from original abstracts; a completed assessment pass does not mean every full paper, supplement or registry history was verified.

Ketones can help a treatment without starving the tumor

In pancreatic-cancer mice, a ketogenic diet increased tumor ketone use. Diet alone did not extend survival; adding it to chemotherapy did. Other lung-cancer experiments found that ketones supported growth under nutrient stress. Conversely, three childhood-cancer cell lines could not use ketones to rescue growth without glucose. These different results support a cancer-specific explanation, rather than a rule that every cancer can or cannot use ketones. [98] [100] [104]

Slowing a cancer cell is different from removing it

A 2026 thyroid-cancer experiment linked a ketone to recycling markers and slower growth, but found no clear increase in apoptosis, a form of programmed cell death. In another study, fasting-like conditions left some bowel-cancer cells dormant and drug-tolerant; adding experimental drugs that trigger iron-related cell death improved killing. Autophagy was involved, but neither experiment demonstrated that fasting alone clears cancer in people. [107] [99]

Human cancer studies still separate feasibility from benefit

A small randomized endometrial-cancer trial found metabolic and immune changes before surgery, not proven tumor shrinkage or longer survival. Extra nonrandomized tissue controls complicate one immune analysis. Long-term KETOCOMP follow-up found no statistically clear survival advantage for ketogenic eating. A separate breast-cancer feasibility study recorded five severe symptom events across 84 diet cycles; cycles must not be mistaken for patients. [108] [106] [105]

New evidence for immune effects, not proof of sugar starvation

Song’s October 2026 study found less bowel-tumor growth in fasting mice, with changes in taurine, cell stress and immune activity. Its main intestinal-tumor comparison used 11–15 mice per group; human tissue observations did not test fasting treatment. Chen’s 2026 study also reports immune changes alongside treatment in mice and patients, but its full human sample, control groups and clinical effect sizes could not be verified here. Its abstract is a lead for further assessment, not confirmation of a survival benefit. [73] [116]

Cell recycling can help treatment or help a cancer survive

The added Pietrocola and Buono experiments point in different directions: functioning autophagy was needed for fasting to improve chemotherapy in one mouse setting, while suppressing it helped the treatment combination in another. The newer Gates preprint describes expelling damaged cells from laboratory cell layers. Together, these findings call for a more precise explanation than either “autophagy kills cancer” or “autophagy cannot help kill cancer.” [70] [71] [72]

Original reports requiring further extraction
Report and evidence levelWhat was checked; what remains open
Liu, 2026 - preclinical CAR T-cell experiments and healthy volunteersAbstract reports stronger immune-cell metabolism and tumor control in models. Human component measured immune-cell energetics, not cancer efficacy. Sample sizes and full controls require extraction.
Xing, 2026 - breast-cancer models and human tissue observationsAbstract reports reduced tumor progression with BHB or ketogenic diet in a PAK5-related mechanism. Full model counts and controls remain unverified; human observations were not a dietary treatment trial.
Wu, 2025 - lung-cancer metabolic experimentsAbstract reports ketone use supporting tumor-initiating cells under low glucose, while creating vulnerabilities to metabolic inhibitors. Neither benefit nor harm from diet alone in patients is established; full denominators remain unverified.
Clarke, September 2026 - preprint, cells/mice and human observationsLow oxygen and tumor glycogen; a tissue comparison used 5 patients with sleep apnea and 5 without. Complete mouse denominators and full appraisal remain pending. This was not a patient fasting trial.
Chen, 2026 - mouse experiments and human treatment observationsOriginal abstract reports immune-cell changes with a 16-hour fast and immunotherapy. Human numbers, assignment, controls and clinical effect sizes are not yet verified here.
Stringer, December 2024 - human feasibility preprintAbstract reports 15 patients with chronic lymphocytic leukemia: 10 followed for 3 months, 5 for 6. Seven had stable or reduced cancer-cell counts; reductions accompanied lower autophagy. Full comparator and assay checks remain pending. This does not establish a causal treatment effect.
Li, 2025 - preclinical treatment-combination studyPublisher summary describes fasting with sorafenib suppressing mitochondrial recycling and increasing cell stress. Complete sample sizes, control groups and full-text appraisal remain pending.

Sources [110] [111] [112] [115] [116] [117] [118]

07 / Cancer & ketones

Does switching to ketones starve cancer?

First, separate glycogen from glucose

Glucose is a small sugar used in metabolism; glycogen is a stored chain of glucose units within cells. Blood glucose, liver glycogen and tumor glycogen are distinct pools. Depleting liver reserves is not the same intervention as disabling a tumor’s own glycogen synthesis. [15] [14]

What the glycogen study supports

The 2025 lung-adenocarcinoma work links glycogen to tumor progression and shows an effect of genetic glycogen-synthesis disruption in mice. It is biologically important evidence for a particular vulnerability. It did not test whether a patient fasting clears tumor glycogen or cures cancer. [14]

The 2026 follow-up does not close that gap

A September preprint links low oxygen with greater use of glycogen in lung-cancer models. One human tissue comparison included five patients with sleep apnea and five without it. This extends the case for studying tumor glycogen, but it is not a fasting treatment trial. The manuscript is not peer reviewed, and the complete study extraction remains pending here. [115]

Does every cancer need glycogen? Direct tests disagree

Xie’s kidney-cancer experiments disrupted both glycogen storage and release without substantially slowing growth. An earlier kidney study found that altering the GYS1 protein did change growth, but explicitly did not prove dependence on glycogen as fuel. A 2026 breast-cancer study offers another distinction: removing GYS1 slowed growth, while other ways to lower glycogen did not. The protein’s signaling role mattered. Cancer type and the exact intervention cannot be skipped. [78] [79] [80]

The proposed starvation chain, checked step by step
Proposed stepWhat the evidence permits
Less dietary carbohydrate → no glucose for the tumorUnsupported. During prolonged fasting, the body continues producing circulating glucose.
High glucose use → glucose is the only usable fuelUnsupported as a universal rule. Human lung tumors have been observed using lactate; metabolic pathways vary.
More ketones → cancer cells cannot obtain energyContradicted in some models. Gliomas oxidized ketones; pancreatic and melanoma models used ketone carbon for growth.
A tumor marker changes → cancer is killedNot established. Growth, radiological response, pathological response, progression and survival are different outcomes.

Sources [15] [16] [17] [81] [84] [20]

The Warburg effect is not proof of a broken energy system

High glucose uptake and lactate production in oxygenated tumors describe a metabolic pattern. They do not demonstrate that mitochondria are universally nonfunctional or that removing dietary sugar selectively deprives every cancer of energy. Direct tracing studies reveal more metabolic flexibility. [16] [17]

Ketones can act as signals as well as fuel

Beta-hydroxybutyrate (BHB) suppressed colorectal-tumor growth through a signaling pathway in one preclinical program. Acetoacetate promoted growth in a BRAF-mutant melanoma model. These are different ketones and contexts. Neither observation can be converted into a universal dietary treatment rule. [18] [19]

The newer ketone studies strengthen the case for specificity

A 2025 tracing study showed how some tumors turn BHB into building material. Other experiments linked BHB to pancreatic-cancer spread and breast-cancer drug resistance. Conversely, a 2026 bladder-cancer study found that a ketogenic diet or BHB improved tumor control and helped immunotherapy in mice. These opposing results are reasons to investigate particular tumors and treatments; neither side proves a universal patient diet effect. [81] [82] [83] [84] [85]

Promising 2026 immune-cell and breast-cancer findings also matter

Liu’s study reports that BHB can fuel engineered cancer-fighting immune cells and improve tumor control in experimental models. Its healthy-volunteer component measured immune-cell metabolism, not cancer survival. Xing reports benefits in a particular breast-cancer pathway. These primary abstracts support further investigation; complete sample sizes and controls remain unverified here. Benefits to immune cells and effects on tumor cells can coexist, so the net clinical result must be tested. [110] [111]

A more plausible research question: can diet help a treatment work?

Investigators are testing whether nutrient and hormone changes alter treatment sensitivity or immune responses. Glucocorticoid signaling in hormone-sensitive breast-cancer models is one example. These mechanisms go beyond a simple “sugar off, cancer dies” story. [25] [22]

The human trials deserve both attention and restraint

DIRECT found some response signals but no improvement in pathological complete response. ERGO2 produced ketosis without meeting its primary cancer endpoint. New breast and pancreatic screening trials are encouraging, but their wide confidence intervals and one-sided 20% significance thresholds call for larger confirmatory trials. [20] [21] [23] [24]

Positive human results should be included with their controls

In Bahrami’s 44-person randomized breast-cancer study, 16 of 22 diet-group patients had a high pathological response, compared with 9 of 22 controls. The diet group also received omega-3, so fasting’s independent contribution is uncertain. Ligorio’s advanced-cancer comparison found longer survival among 14 diet participants versus 76 others, but was not randomized. BREAKFAST studied 30 patients; its registered arms both included the fasting-style diet, so it cannot isolate diet versus ordinary eating. [86] [87] [88] [114]

Case reports and feasibility trials answer smaller questions

Three reported lymphoma cases improved after prolonged fasting and dietary changes, but there was no control and the disease can fluctuate. The 2026 SWiFT trial randomized only 12 people and primarily tested whether fasting before chemotherapy was practical. A prostate-cancer pilot found weight loss with two diets, and a breast-cancer interim report examined safety. None of these designs establishes cancer eradication by fasting. [89] [90] [92] [91]

08 / Autophagy

Cellular recycling has no universal countdown

Autophagy describes cellular processes that deliver material to lysosomes for breakdown and reuse. In the macroautophagy studied here, cellular cargo is enclosed, delivered and degraded. It is not synonymous with killing an entire cell or selectively removing cancer. [30] [32]

When does it start?

No validated fasting hour

Basal autophagy exists in fed cells. The question is when fasting changes its rate in a particular tissue, not when a body-wide switch turns on. [30] [27]

When does it peak?

No established human peak

Sampling blood or muscle after a chosen interval cannot locate a universal maximum. Popular 18-, 24-, 48- or 72-hour milestones are not validated by these studies. [26] [29] [28]

How long does it last?

No established fixed duration

Responses vary with tissue and feeding state. Available human trials do not show a whole-body benefit window or a precise time when recycling stops after food. [27] [28] [30]

Why measuring it is difficult

More LC3 gene expression suggests pathway regulation. More LC3-containing structures could mean increased production or a bottleneck in degradation. Autophagic flux measures movement through the pathway, often by comparing samples with and without lysosomal inhibition. Even flux in blood immune cells is not a measurement of liver, brain or tumor flux. [26] [29] [27]

What the newer human-related findings add

Hofer’s work connects fasting, spermidine and autophagy through experiments across species. Human blood samples showed metabolite changes; the experiments establishing pathway dependence and longevity effects were not human lifespan trials. In a 48-person fatty-liver trial, an autophagy-related blood protein changed after eight weeks, but actual recycling flow was not measured. Neither study supplies a reliable hour-by-hour human countdown. [74] [95]

What the human timepoints actually mean
ObservationWhat it cannot establish
4 days of early eating-window assignment; morning/evening blood samples (Jamshed)LC3A gene expression cannot determine the first fasting hour with increased flux.
Overnight versus 72-hour fasting; muscle biopsies (Vendelbo)Ambiguous static markers cannot identify a 72-hour peak or prove whole-body clearance.
Blood flux assessed after 2 and 6 months of repeated restriction (Bensalem)A chronic intervention comparison cannot locate the onset during one fast.
FMD samples on days 4, 6 and 8 (Espinoza)Sparse blood-cell measurements cannot define a universal start, persistence or clinical benefit.

Sources [26] [29] [27] [28]

What might it do for the body?

Cellular recycling helps manage damaged proteins and organelles and makes material available during stress. Its importance is supported by genetic experiments. But showing that the pathway is necessary is different from showing that pushing it higher through fasting improves a person’s health. [31] [32]

Why cancer makes the story more complicated

Autophagy can restrain early cellular damage: disabling it produced liver adenomas in one mouse model. It can also support tumor metabolism: disabling it reduced malignant progression in a Kras lung model. Tumor type, mutations, stage and treatment matter. “More recycling” is not automatically “less cancer.” [31] [32]

Can autophagy contribute to killing cancer? Yes, in some experimental settings

In a 2016 mouse study, fasting helped chemotherapy control tumors only when tumor-cell autophagy and immune cells could function. That is evidence for a useful role in a treatment response. It is not evidence that autophagy simply swallows entire cancers or that fasting alone cures people. Conversely, a 2023 leukemia study linked better survival with a fasting-style diet plus chemotherapy to reduced tumor-cell autophagy. The direction of the effect depends on the setting. [70] [71]

A newer finding: pushing damaged cells out

A September 2026 preprint reports that nutrient-deprived cell layers expelled damaged and gene-altered cells. This used some autophagy-related machinery but did not require the usual digestion of cell contents. It was studied in laboratory cell lines, not people fasting. The short wave observed over one to four hours is a laboratory result, not a human autophagy schedule. Whether fasting triggers this process in living animals or people remains unanswered in that study. [72]

Refeeding is part of the biology

Nutrient reintroduction changes growth and repair signals. In mice, post-fast refeeding increased intestinal regeneration but also amplified tumors after engineered Apc loss. This does not establish human harm; it shows why regeneration alone cannot be equated with cancer protection. [33]

What would answer the timing question?

We would need repeated validated flux measurements across relevant human tissues during fasting and refeeding, with suitable controls, followed by trials linking those changes to benefits and harms. Ketone strips, hunger and elapsed time have not been validated as substitutes in the studies reviewed here.

A measurement at 24 hours is not a start time

Pietrocola’s small human blood-cell experiments illustrate the measurement problem. A test performed outside the body with degradation blocked detected increased recycling after 24 hours of fasting, while direct measurements after 48 hours did not clearly detect it. The key panels used four volunteers per condition. This does not mean recycling was absent before 24 hours, stopped by 48 hours, or followed the same pattern in the liver, muscle or a tumor. [97]

A tissue-specific negative result matters

Chaudhary measured fasting-related markers in mouse liver and muscle and in human muscle. Responses differed by tissue; the muscle findings did not support a universal increase. Neither static gene expression nor a chosen 24-hour sampling point defines an activation threshold. [69]

09 / Muscle, brain & populations

Who was studied matters

Muscle and function

Resistance-trained men maintained lean mass during a short time-restriction trial with exercise. That does not generalize to a 72-hour complete fast or to frail older adults. An older-adult feasibility trial adds useful functional data, but was not powered to rank treatments or establish frailty prevention. [51] [29] [55]

Cognition and neurological disease

A 46-person trial in mild cognitive impairment reported a MoCA improvement; a 43-person study of healthy women found no intervention-specific cognitive benefit. The multiple-sclerosis FMD pilot combined one cycle with six months of Mediterranean eating. Test scores and quality-of-life signals are not evidence of dementia prevention or neurological cure. [56] [57] [63]

Larger or longer follow-up changes the picture

An 18-month, 105-person multiple-sclerosis trial found no advantage for fasting or ketogenic eating over a healthy diet on its main brain-lesion outcome. A separate 58-person overnight-fasting pilot improved sleep relative to health education, without a clear cognitive advantage. These specific results should remain separate from broader claims about brain repair. [75] [76]

Immunity and regeneration

Cheng’s animal work connects fasting-related IGF-1/PKA changes to blood stem-cell biology. Its tiny preliminary human observations came from the same chemotherapy program later reported by Dorff. Neither provides a validated 72-hour immune reset, fewer infections in healthy people, or a general human rejuvenation outcome. [62] [60]

Adolescents and shift workers

The adolescent trial used intensive multidisciplinary care and did not find intermittent restriction superior to continuous restriction. The firefighter trial suggests feasibility in selected shift workers, with several benefits concentrated in secondary or higher-risk subgroup analyses. Neither is a blanket recommendation for its entire population. [54] [53]

Religious fasting has a different exposure

Ramadan can combine daytime food and fluid abstention with changes in sleep, medication and nighttime meals. EPIDIAR reported more severe hypoglycemia during Ramadan in people with diabetes. Its observational comparison cannot quantify the effect of intentional, water-permitted TRE in healthy adults. Evidence on prolonged dry fasting is not established by this library. [65]

Mortality headlines require a design check

A published NHANES analysis associated short eating duration with cardiovascular mortality. Two dietary recalls do not reliably identify a long-term intentional fasting practice; illness and other behaviors may explain the association. This does not prove that fasting causes cardiovascular death, and favorable biomarker trials do not prove that it prevents it. [58] [7]

Pregnancy, breastfeeding, underweight, frailty, eating-disorder histories, advanced kidney or liver disease, and medication-dependent fasting remain important evidence gaps in this selected review. Absence of a study record is not evidence of either safety or harm. Safety decisions require condition-specific clinical assessment.

10 / Dr. Jamnadas

Dr. Pradip Jamnadas, MD, in context

Dr. Pradip Jamnadas, MD, is included through his practice’s Fasting for Survival lecture link and his original 2022 Processed Food Addiction slides. A lecture is an educational perspective; the cited experiments determine the strength and scope of a clinical claim. The glycogen assertion above is not presented as his quotation. [120] [119]

Claims in the original slides, checked against primary studies
Original materialWhat the evidence supports
Pages 46–47: weight, insulin sensitivity and diabetes claimsSeveral supported programs improve weight and glucose. Remission occurred in a small selected trial; permanent cure or medication discontinuation for everyone is not established.
Page 47: induction of autophagy and mitophagyHuman evidence depends on tissue and assay. A pathway diagram or static marker cannot establish a universal start time or clinical benefit.
Page 47: elimination of fatty liverA randomized liver-fat trial found similar reductions with calorie restriction with and without a daily eating window. Universal elimination is unsupported.
Pages 46–47: neural, mood, sleep, pain and toxin-related claimsSelected cognition and symptom studies cannot establish every listed benefit. The selected studies do not establish general detoxification or elimination of sleep apnea. Those outcomes require dedicated assessment and were not comprehensively reviewed here.
Page 49: autophagy diagram including cancerA biological relationship is not proof that fasting kills cancer. Autophagy can suppress early damage or support established tumors, depending on context.

Sources [119] [5] [50] [27] [69] [56] [57] [31] [32]

Source boundary: the identified slides were visually checked, including pages 46, 47 and 49. The complete lecture transcript was not independently verified, so exact lecture quotations and numerical promises are not attributed to the speaker. Broad slide claims should not be substituted for trial results. [119]

11 / Harms & applicability

Benefit includes what can go wrong

The longer-fasting evidence comes largely from selected, supervised adults. Reported problems include dehydration, electrolyte disturbance, weakness and other symptoms; short-term falls in scale weight are not a guarantee of muscle preservation or lasting benefit. [13] [12] [29]

Cancer-related weight loss and malnutrition can make further restriction particularly consequential. During cancer care, discuss any fasting or ketogenic plan with the treating team and an oncology dietitian. The ASCO guidance cited here predates the newer small trials; those trials still do not establish a general replacement for standard treatment. [121] [24] [23]

Reintroducing nutrition after prolonged inadequate intake is a distinct clinical issue. NICE describes assessment, monitoring and gradual nutrition support for people at risk of refeeding problems. An elapsed-time threshold or feeling well does not establish that a fast or refeeding at home is safe. [122]

EVIDENCE PROFILES

How strong is each conclusion?

Confidence applies to a specific question, outcome and comparator, rather than to fasting as a whole. These are provisional editorial assessments. They are not independently adjudicated GRADE ratings.

Does fasting help adults lose weight?

Supported for several structured programs

Moderate, provisional

Repeated randomized comparisons support weight loss relative to usual behavior. Average advantages over active calorie restriction are usually small or uncertain; support, energy intake and adherence differ. Most follow-up is one year or less. [3] [38] [36] [2] [1] [41]

Is fasting better than calorie restriction?

No consistent overall advantage

Moderate, provisional

Several larger or longer active-control trials are null; some supported regimens produce additional loss. A single pooled number would conceal different schedules, calorie targets and comparators. “Not significant” does not mean proven equivalent. [2] [36] [38] [42] [1] [39]

Can it improve diabetes measures or remission?

Possible benefit under clinical management

Low to moderate, provisional

HbA1c improvements and remission signals are clinically relevant, but small studies, dietary packages, medication management and limited durability reduce confidence in a fasting-specific effect. Remission is not permanent cure. [5] [43] [44] [45] [46] [48] [49] [93]

Does it prevent heart attacks or extend human life?

Not established by this trial library

Insufficient direct trial evidence

Blood pressure, lipids and glucose are intermediate outcomes. The selected trials do not establish fewer cardiovascular events or longer human survival. Observational mortality associations cannot settle causality; mouse lifespan cannot answer the human question. [7] [4] [58] [34]

Does multiday water-only fasting improve long-term health?

Short-term changes; comparative benefit unresolved

Very low, provisional

Mostly uncontrolled, selected residential cohorts with refeeding and subsequent dietary changes. Fluid/lean loss, adverse markers, attrition and limited follow-up prevent strong conclusions about net benefit, optimal duration or home safety. [64] [10] [11] [12] [68] [13]

Do ketones starve and kill cancers?

Universal claim is not supported

Direct mechanistic counterexamples; clinical efficacy unproven

Some tumors use alternative fuels, including ketones; effects differ by tumor and ketone. A specific glycogen pathway is not a universal nutrient dependency. No selected human trial establishes fasting or ketosis alone as a cancer cure. [14] [78] [80] [81] [84] [18] [85] [21]

Can fasting improve cancer treatment?

Investigational adjunct, not established treatment

Low to very low, provisional

Small heterogeneous trials, adherence problems and wide intervals. Some response or symptom signals coexist with null primary outcomes. New survival screening trials use permissive one-sided thresholds and need confirmatory replication. [20] [21] [23] [24] [86] [87] [90]

When does beneficial autophagy begin and end?

No validated universal human schedule

Insufficient direct timing/outcome evidence

Genes, static proteins and tissue-specific flux are different measurements. Sparse or chronic sampling cannot identify a universal onset, peak, duration or health benefit. Basal recycling already occurs in fed cells. [26] [27] [28] [29] [69] [30] [74] [95]

Can autophagy contribute to an anticancer response?

Yes in some models; the direction depends on context

Experimental evidence; human treatment implications unresolved

Tumor autophagy was necessary for a fasting-plus-chemotherapy benefit in one mouse model, while inhibiting it helped another treatment combination. Cell extrusion in the Gates preprint is distinct from ordinary intracellular digestion. None establishes universal cancer clearance through fasting-induced autophagy in people. [70] [71] [32] [72] [99] [107]

Does fasting preserve muscle?

Depends on regimen and population

Low, provisional

Short resistance-training studies and supervised multiday studies test very different exposures. Fat-free mass includes water; neither a stable DXA result nor later fluid recovery establishes preserved muscle function in every setting. [51] [37] [29] [68] [55]

Does it improve cognition or reset immunity?

Preliminary and population-specific

Very low, provisional

Small cognitive pilots conflict across populations. Animal regeneration, immune-cell changes and multi-component dietary trials do not establish general human rejuvenation, infection prevention or a timed immune reset. [56] [57] [62] [63] [9] [75] [76]

Does it eliminate fatty liver?

Liver fat can improve; universal elimination unproven

Low to moderate for measured liver fat

A 12-month active-control trial found similar liver-fat reductions with and without time restriction. It did not establish that fasting invariably resolves the disease or prevents liver complications. [50] [102]

THE STUDY LIBRARY

Read the study behind the claim

109 original-publication records, grouped by question. Open a record for the sample, controls, protocol, outcomes and limitations. Some reports share participants; this count is not a count of independent trials.

Weight, glucose & cardiovascular markers

The most informative human comparisons include active dietary controls and null results.

01Catenacci et al. · 2025Human randomized trialA modest additional weight benefit in an intensive support program.Study design, results & limitations

The Effect of 4:3 Intermittent Fasting on Weight Loss at 12 Months: A Randomized Clinical Trial

What happened
The three-day group lost about 2.9 kg more, on average, after a year.
Who took part
165 adults with overweight or obesity started; 125 finished the one-year study.
What they compared
84 people ate much less on three days each week. The other 81 cut calories every day. Both groups received substantial coaching and exercise advice.
What this means
This approach helped in a supported program. The study cannot show whether meal timing itself, eating fewer calories, or sticking to the plan explains the difference.
Population & sample
165 adults with overweight or obesity randomized; 125 completed 12 months.
Arms & controls
2 arms: 4:3 restriction (84) and daily calorie restriction (81). One active control; both received intensive behavioral support and an exercise goal.
Protocol & duration
80% energy restriction on 3 nonconsecutive days each week versus 34% daily restriction; 12 months. Restricted days still included food.
Reported results
Primary weight outcome: 2.89 kg greater loss with 4:3 (95% CI 0.14–5.65 kg; P=.040), in the intention-to-treat analysis.
What limits the inference
Not a water-fast trial. Attrition, intensive support, and differences in achieved calorie restriction limit generalization and a calorie-independent explanation.
Source check & disclosures
Original indexed abstract; protocol and primary outcome. NIH funding reported. Registry NCT03411356.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [1] DOI: 10.7326/ANNALS-24-01631

Link to this study record
02Liu et al. · 2022Human randomized trialAn eating window did not clearly improve on calorie restriction.Study design, results & limitations

Calorie Restriction with or without Time-Restricted Eating in Weight Loss

What happened
Both groups lost weight. The study did not find a clear extra benefit from the eight-hour eating window.
Who took part
139 adults with obesity; 118 completed the one-year follow-up.
What they compared
Both groups cut calories. One group of 69 also ate only between 8 a.m. and 4 p.m.; the other 70 kept their usual meal timing.
What this means
An eating window may be one way to organize a diet, but this study did not show that it works better than calorie cutting alone.
Population & sample
139 adults with obesity randomized; 118 completed the 12-month visit.
Arms & controls
2 arms, one active control: calorie restriction with an 08:00–16:00 eating window (69) versus calorie restriction alone (70).
Protocol & duration
Both groups received calorie targets for 12 months; only one also restricted meal timing.
Reported results
Mean weight change was −8.0 versus −6.3 kg. Adjusted between-group difference: −1.8 kg (95% CI −4.0 to 0.4; P=.11). Body-fat and metabolic outcomes did not establish additional benefit.
What limits the inference
A confidence interval spanning zero leaves benefit and little difference plausible; it does not prove the approaches identical. No mortality endpoint.
Source check & disclosures
Original abstract and publisher PDF enrollment table. Registry NCT03745612.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [2] DOI: 10.1056/NEJMoa2114833

Link to this study record
03Trepanowski et al. · 2017Human randomized trialAlternate-day restriction was not superior and had more dropouts.Study design, results & limitations

Effect of Alternate-Day Fasting on Weight Loss, Weight Maintenance, and Cardioprotection Among Metabolically Healthy Obese Adults: A Randomized Clinical Trial

What happened
The two dieting groups lost similar amounts of weight. About 38% left the alternate-day group, compared with 29% in the daily-diet group.
Who took part
100 adults with obesity, followed for one year.
What they compared
34 alternated low-food days with higher-food days, 35 cut calories every day, and 31 were asked to keep their usual habits.
What this means
Alternate-day eating was not clearly more effective or easier to maintain. The low-food days still included calories; this was not a water-only fast.
Population & sample
100 adults randomized: 34 alternate-day fasting, 35 daily restriction, 31 no-intervention control.
Arms & controls
3 arms; two comparators for fasting: an active calorie-restriction program and usual behavior.
Protocol & duration
6 months of weight loss, then 6 months of maintenance. During weight loss, alternate days supplied 25%/125% of energy needs; daily restriction supplied 75%.
Reported results
At 12 months, weight loss relative to control was about 6.0% versus 5.3% for alternate-day versus daily restriction, without a significant difference. Dropout: 38% versus 29% versus 26%.
What limits the inference
Adherence differed from assigned intake. “Fasting” meant modified fasting with calories. No evidence here of superior cardiovascular event prevention.
Source check & disclosures
Original publication abstract, design and results.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [3] DOI: 10.1001/jamainternmed.2017.0936

Link to this study record
04Sutton et al. · 2018Human randomized crossoverA small controlled-feeding experiment supports a timing effect.Study design, results & limitations

Early Time-Restricted Feeding Improves Insulin Sensitivity, Blood Pressure, and Oxidative Stress Even Without Weight Loss in Men with Prediabetes

What happened
Earlier eating improved how their bodies responded to insulin, the hormone that helps control blood sugar. Morning blood pressure also improved.
Who took part
Eight men with blood sugar above the normal range completed both eating schedules.
What they compared
Each man tried a six-hour eating window ending before 3 p.m. and a twelve-hour window, for five weeks each. Food was provided to keep weight steady.
What this means
This suggests timing might matter even without weight loss. Eight men are too few to show that the approach prevents diabetes or heart disease.
Population & sample
8 men with prediabetes completed both conditions; this is a completer count, not a large clinical trial.
Arms & controls
2 conditions in randomized order. Each participant was his own control: early 6-hour versus 12-hour eating window.
Protocol & duration
5 weeks per condition; meals provided and intake adjusted to maintain weight. Early eating ended before 15:00.
Reported results
Insulin sensitivity and morning blood pressure improved with early eating; glucose measurements did not show the same benefit.
What limits the inference
Male-only, very small, short study. Different fasting lengths before testing complicate interpretation. Surrogate changes do not establish prevention of diabetes or heart attacks.
Source check & disclosures
Full text: study design, Results and limitations; meal provision verified.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [4] DOI: 10.1016/j.cmet.2018.04.010

Link to this study record
05Yang et al. · 2023Human randomized trialRemission was more frequent with a structured dietary program.Study design, results & limitations

Effect of an Intermittent Calorie-restricted Diet on Type 2 Diabetes Remission: A Randomized Controlled Trial

What happened
After the program and three more months of follow-up, 17 people in the diet group met the study’s diabetes-remission criteria, compared with one in the usual-diet group.
Who took part
72 adults with type 2 diabetes, split into two groups of 36.
What they compared
One group received a three-month program of supplied meals and periods of reduced eating. The other continued its usual diet.
What this means
Here, remission meant blood sugar stayed below the diabetes threshold without diabetes medicine for at least three months. It is not a permanent cure, and this result does not justify stopping medication on your own.
Population & sample
72 adults with type 2 diabetes; 36 assigned to each group.
Arms & controls
2 arms: Chinese Medical Nutrition Therapy (CMNT) versus usual diet/control. One control; no calorie-matched alternative dietary program.
Protocol & duration
3-month intermittent restricted-calorie program with supplied meals, followed by follow-up. This was not water-only fasting.
Reported results
After intervention plus 3 months of follow-up, remission occurred in 17/36 (47.2%) versus 1/36 (2.8%). Remission required HbA1c below 6.5% for at least 3 months without diabetes medication. At 12-month follow-up, 16/36 intervention participants remained in remission.
What limits the inference
Small, selected population; diet composition, calorie reduction and medical management cannot be separated. Remission is not permanent cure and does not authorize stopping medication.
Source check & disclosures
Original abstract and publisher results/participant-flow text. Long-term HbA1c values differ between abstract and body; those values are not repeated here.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [5] DOI: 10.1210/clinem/dgac661

Link to this study record
06Wei et al. · 2017Human randomized trialRisk markers improved; disease prevention was not tested.Study design, results & limitations

Fasting-mimicking diet and markers/risk factors for aging, diabetes, cancer, and cardiovascular disease

What happened
Weight, body fat, blood pressure and a hormone linked with growth improved in the diet group.
Who took part
100 generally healthy adults started: 52 assigned to the special diet and 48 to usual eating.
What they compared
For three months, one group ate a low-calorie, low-protein diet for five days each month. The other initially kept its usual diet; some tried the special diet later.
What this means
These measurements may be useful clues, but the trial did not show fewer cancers, fewer heart attacks or longer life. Not everyone finished, and later diet users were not a separate trial.
Population & sample
100 generally healthy adults randomized: 52 FMD, 48 control. Initial randomized comparison had 39 FMD and 43 control completers; 71 eventually completed FMD after crossover.
Arms & controls
2 initial arms; one usual-diet control, later offered FMD. The 71 pooled completers are not a second independent trial.
Protocol & duration
5-day low-calorie, low-protein fasting-mimicking diet each month for 3 months; ordinary eating between cycles.
Reported results
Randomized results favored FMD for weight, body fat, blood pressure and IGF-1. Larger improvements in higher-risk participants were post hoc analyses.
What limits the inference
Attrition, short duration and surrogate endpoints. Lower IGF-1 is not demonstrated cancer prevention; no human lifespan outcome.
Source check & disclosures
Full text: Results, participant flow and Tables 1–2. Commercial FMD interests, including L-Nutra relationships, disclosed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [6] DOI: 10.1126/scitranslmed.aai8700

Link to this study record
07Cramer et al. · 2022Human randomized trialSome benefits, but the two primary outcomes were not positive.Study design, results & limitations

Effects of Fasting and Lifestyle Modification in Patients with Metabolic Syndrome: A Randomized Controlled Trial

What happened
The main blood-pressure and insulin-response tests showed no clear extra benefit after twelve weeks. A later weight measurement favored the fasting group.
Who took part
145 adults with a combination of risk factors such as high blood pressure, raised blood sugar or excess weight.
What they compared
73 began with five days of very little food, including juice and broth, before a lifestyle program. The other 72 received the lifestyle program without that fast.
What this means
The results were mixed. A positive weight finding should not hide the lack of a clear difference in the two main tests.
Population & sample
145 adults with metabolic syndrome randomized; 73 fasting plus lifestyle, 72 lifestyle alone. 15 per arm were lost during follow-up.
Arms & controls
2 arms; one active control with modified DASH diet, exercise and mindfulness.
Protocol & duration
5-day modified fast before 10 weeks of lifestyle intervention; follow-up to 24 weeks. The fast included approximately 300–350 kcal/day of juice and broth.
Reported results
Neither 12-week ambulatory systolic pressure nor HOMA insulin resistance, the co-primary outcomes, differed significantly. At 24 weeks, secondary weight difference favored fasting by 2.7 kg (95% CI 0.5–4.8).
What limits the inference
Short-term improvements were not uniformly sustained. Multi-component intervention and secondary outcomes limit fasting-specific claims. Uric acid rose initially.
Source check & disclosures
Full text: Methods, Results, Tables 2–3. Parent trial of the Maifeld microbiome substudy; not independent replication.
Cohort relationships
Maifeld reports a 71-person substudy of the 145-person Cramer program (NCT02099968).

Read original publication [7] DOI: 10.3390/jcm11164751

Link to this study record
08Maifeld et al. · 2021Human randomized substudyA promising subgroup result needs its parent-trial context.Study design, results & limitations

Fasting alters the gut microbiome reducing blood pressure and body weight in metabolic syndrome patients

What happened
This smaller group showed improvements in blood pressure and weight, alongside changes in gut bacteria.
Who took part
71 people from the larger Cramer trial: 35 in the fasting group and 36 in the comparison group.
What they compared
Researchers compared five days of very low food intake followed by a healthy-lifestyle program with the lifestyle program alone. They also studied gut bacteria.
What this means
That does not prove the bacteria caused the improvement. These were some of the same participants as the larger trial, whose main blood-pressure result was not clearly positive.
Population & sample
71 participants: 35 fasting plus DASH, 36 DASH alone; a subset of the 145-person trial reported by Cramer.
Arms & controls
2 arms, one active DASH/lifestyle comparator; repeated microbiome and immune measurements.
Protocol & duration
5 days of 300–350 kcal/day juice/broth, then 3 months of modified DASH and lifestyle support.
Reported results
The substudy reported sustained blood-pressure and weight improvements and changes in microbiome composition. Antihypertensive medication reductions were reported in 43% versus 17%.
What limits the inference
Exploratory subgroup and many molecular comparisons. Microbiome associations do not establish mediation. The larger trial did not meet its blood-pressure or HOMA co-primary endpoints.
Source check & disclosures
Full text: Table 1, Figure 2 and Methods; registry NCT02099968. Read with Cramer, not as a second confirming trial.
Cohort relationships
Maifeld reports a 71-person substudy of the 145-person Cramer program (NCT02099968).

Read original publication [8] DOI: 10.1038/s41467-021-22097-0

Link to this study record
09Kjeldsen-Kragh et al. · 1991Human randomized trialA combined fasting/diet program improved symptoms in a small trial.Study design, results & limitations

Controlled trial of fasting and one-year vegetarian diet in rheumatoid arthritis

What happened
Joint symptoms, movement and several blood measurements improved in the diet-program group.
Who took part
53 people with rheumatoid arthritis, a condition in which the immune system attacks joints.
What they compared
27 began with seven to ten days of very little food, then followed vegetarian diets for the rest of a year. The other 26 continued an ordinary diet.
What this means
The fast and the year of dietary changes were tested together. The study cannot tell us how much benefit came from fasting itself.
Population & sample
53 patients: 27 assigned to the dietary program, 26 to control; examiner blinded to assignment.
Arms & controls
2 arms, one ordinary-diet control. Both began with 4-week residential stays, in different facilities.
Protocol & duration
7–10-day subtotal fast, then 3.5 months of individually adjusted gluten-free vegan diet, followed by lactovegetarian eating for the remainder of the year.
Reported results
Joint symptoms, function and several inflammatory markers improved in the intervention group, with differences reported at one year.
What limits the inference
Small, older trial with many outcomes and a sustained dietary change. Cannot isolate fasting from the subsequent diet or setting; not evidence for replacing arthritis treatment.
Source check & disclosures
Original publication abstract. Full participant-flow tables and effect-size confidence intervals were not extracted.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [9] DOI: 10.1016/0140-6736(91)91770-U

Link to this study record
36Lin et al. · 2023Human randomized trialA practical alternative to calorie counting, without a demonstrated advantage.Study design, results & limitations

Time-Restricted Eating Without Calorie Counting for Weight Loss in a Racially Diverse Population : A Randomized Controlled Trial

What happened
Both supported dieting approaches helped people lose weight. The eating-window group did not clearly outperform the calorie-counting group.
Who took part
Ninety adults with obesity, thirty in each group. Seventy-seven completed a year.
What they compared
One group used a daily eating window, one counted and cut calories, and one kept its usual diet. The eating window widened during the second half of the year.
What this means
An eating window may be a useful alternative to counting calories. This trial does not establish that it has a special weight-loss effect independent of eating less.
Population & sample
90 adults with obesity randomized; 77 completed 12 months.
Arms & controls
3 arms, 30 each: time restriction, calorie restriction, usual diet. Two comparators; counseling in both active arms.
Protocol & duration
12:00–20:00 eating for 6 months, then a 10-hour maintenance window; comparator prescribed a 25% calorie deficit.
Reported results
At 12 months, time restriction versus control: −4.61 kg (95% CI −7.37 to −1.85). Versus calorie restriction: +0.81 kg (−3.07 to 4.69; P=.68).
What limits the inference
Small trial; active arms received more support than control. Secondary comparisons were not all multiplicity-adjusted.
Source check & disclosures
Full text, participant flow and outcomes. NCT04692532.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [36] DOI: 10.7326/m23-0052

Link to this study record
37Lowe et al. · 2020Human randomized trialA midday eating window did not produce extra weight loss.Study design, results & limitations

Effects of Time-Restricted Eating on Weight Loss and Other Metabolic Parameters in Women and Men With Overweight and Obesity: The TREAT Randomized Clinical Trial

What happened
The eating window did not produce clearly greater weight loss. Some scan results raised questions about loss of lean tissue, which includes muscle and water.
Who took part
141 adults were assigned to eating plans. Only 116 supplied weight data, and 105 completed the study. A smaller group had body-composition scans.
What they compared
For twelve weeks, one group ate only between noon and 8 p.m. The comparison group ate three structured meals a day. Neither was given a calorie-cutting target.
What this means
Missing results and the small scan group make firm conclusions difficult. This study does not support assuming that a shorter eating window automatically burns extra fat or preserves muscle.
Population & sample
141 randomized; 116 supplied weight data (59 time restriction, 57 control); 105 completed. Body-composition subgroup: 50.
Arms & controls
2 arms: 12:00–20:00 eating versus three structured meals daily; one active meal-pattern control.
Protocol & duration
12 weeks, without prescribed calorie reduction or exercise changes.
Reported results
Between-group weight difference: −0.26 kg (95% CI −1.30 to 0.78; P=.63). An appendicular lean-mass loss signal appeared in the smaller subgroup.
What limits the inference
25 randomized people contributed no weight data. Subgroup composition findings require replication; lean mass is not identical to muscle protein.
Source check & disclosures
Full text and participant flow. November 2020 Figure 2C correction and 2021 participant-message clarification identified. NCT03393195.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [37] DOI: 10.1001/jamainternmed.2020.4153

Link to this study record
38Schübel et al. · 2018Human randomized trial5:2 and daily restriction produced similar longer-term weight changes.Study design, results & limitations

Effects of intermittent and continuous calorie restriction on body weight and metabolism over 50 wk: a randomized controlled trial

What happened
The two dieting groups lost similar amounts of weight by the final follow-up.
Who took part
150 adults with overweight or obesity, followed for about a year.
What they compared
One group ate much less on two days a week, another cut calories every day, and a third received no advice to cut calories. Both dieting plans targeted a similar weekly calorie reduction.
What this means
The two-day approach did not clearly work better. Similar results in one trial do not prove that every version of the diets will have exactly the same effects.
Population & sample
150 adults with overweight or obesity; arm-specific allocation and final missingness were not extracted from the original full text.
Arms & controls
3 arms: 5:2 restriction, daily calorie restriction, no advice to restrict calories. Two comparators.
Protocol & duration
12 weeks intervention, 12 maintenance and 26 follow-up; both active prescriptions targeted approximately 20% weekly restriction.
Reported results
Week-50 weight change: −5.2%, −4.9%, −1.7%; 5:2 versus daily restriction P=.89. No advantage in the 82 selected adipose-tissue genes.
What limits the inference
A nonsignificant difference is not formal proof of equivalence. Gene expression and biomarkers do not establish disease prevention.
Source check & disclosures
Original indexed abstract; full-text extraction incomplete. NCT02449148.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [38] DOI: 10.1093/ajcn/nqy196

Link to this study record
39Jamshed et al. · 2022Human randomized trialEarly eating added weight loss, but the primary fat-loss result was uncertain.Study design, results & limitations

Effectiveness of Early Time-Restricted Eating for Weight Loss, Fat Loss, and Cardiometabolic Health in Adults With Obesity: A Randomized Clinical Trial

What happened
The early-eating group lost about 2.3 kg more weight on average. Its extra loss of body fat was less certain.
Who took part
Ninety adults with obesity, 45 assigned to each group. Only 59 completed every assessment.
What they compared
Both groups cut calories for fourteen weeks. One also ate between 7 a.m. and 3 p.m.; the other spread eating across at least twelve hours.
What this means
Scale weight and body fat are different outcomes. Missing assessments and the short study make it harder to know how much benefit would last.
Population & sample
90 adults with obesity randomized, 45 per arm; 59 completed all assessments.
Arms & controls
2 arms: early 8-hour window plus energy restriction versus at least 12-hour eating plus energy restriction.
Protocol & duration
07:00–15:00 eating; 14 weeks with weight-loss support in both groups.
Reported results
Additional weight change: −2.3 kg (95% CI −3.7 to −0.9; P=.002). Body-fat difference: −1.4 kg (−2.9 to 0.2; P=.09).
What limits the inference
20 withdrawals and 11 pandemic-related missing assessments. Positive completer-only fat results should not replace the main analysis.
Source check & disclosures
Full text and flow diagram; primary outcomes distinguished. NCT03459703.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [39] DOI: 10.1001/jamainternmed.2022.3050

Link to this study record
40Dote-Montero et al. · 2025Human randomized trialChanging eating-window timing did not clearly reduce visceral fat further.Study design, results & limitations

Effects of early, late and self-selected time-restricted eating on visceral adipose tissue and cardiometabolic health in participants with overweight or obesity: a randomized controlled trial

What happened
After twelve weeks, scans showed no clear extra reduction in fat around internal organs from any of the eating windows.
Who took part
197 adults with overweight or obesity, divided among four groups.
What they compared
Everyone received Mediterranean-diet education. Three groups also used an eight-hour eating window: early, late, or at a time they chose. The fourth had no added window.
What this means
This trial did not establish a best time of day for losing this type of belly fat. Changes on the scale cannot substitute for the scan result.
Population & sample
197 adults: 49 usual care, 49 early, 52 late and 47 self-selected time restriction.
Arms & controls
4 arms; one usual-care comparator with Mediterranean-diet education. Three 8-hour windows added to that care.
Protocol & duration
12 weeks; primary outcome was MRI-measured visceral adipose tissue.
Reported results
Between-group differences versus usual care: −4% (95% CI −12 to 4), −6% (−13 to 2), −3% (−11 to 5), respectively. All nonsignificant.
What limits the inference
Short duration. A favorable weight result does not establish an advantage for the primary visceral-fat endpoint.
Source check & disclosures
Original publisher abstract; five mild adverse events, no serious events reported. NCT05310721.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [40] DOI: 10.1038/s41591-024-03375-y

Link to this study record
41Hajek et al. · 2021Human randomized trialSimple 5:2 advice was feasible, but adherence fell substantially.Study design, results & limitations

A randomised controlled trial of the 5:2 diet

What happened
The 5:2 plan and ordinary advice produced similar average weight loss. Reported use of the 5:2 plan fell from 74% at six weeks to 22% at a year.
Who took part
300 adults with obesity, with 100 in each group.
What they compared
Researchers compared ordinary brief weight-loss advice, a self-help 5:2 plan, and the same 5:2 plan with six group-support sessions. They followed people for a year.
What this means
A diet has to remain workable over time. Extra early support did not establish a lasting advantage in this trial.
Population & sample
300 adults with obesity, 100 per arm.
Arms & controls
3 arms: brief conventional advice, 5:2 self-help, or 5:2 with six group sessions. Two active comparators.
Protocol & duration
One-year follow-up; 5:2 restricted intake on two days each week.
Reported results
At one year, self-help 5:2 and conventional advice produced about −1.9 and −1.8 kg. Reported 5:2 adherence fell from 74% at six weeks to 22% at one year.
What limits the inference
Additional early group support did not establish a sustained advantage. Self-reported adherence and attrition limit interpretation.
Source check & disclosures
Full text; ISRCTN79408248. Public funding and no competing interests reported.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [41] DOI: 10.1371/journal.pone.0258853

Link to this study record
42Maruthur et al. · 2024Human randomized trialA controlled-feeding trial found no calorie-independent weight advantage.Study design, results & limitations

Effect of Isocaloric, Time-Restricted Eating on Body Weight in Adults With Obesity : A Randomized Controlled Trial

What happened
Both groups lost some weight. The early eating window did not clearly add weight loss or improve blood sugar.
Who took part
41 adults with obesity and raised blood sugar or diet-controlled diabetes. All finished the study.
What they compared
Researchers supplied food for twelve weeks, matching calories and nutrients. One group ate within ten hours, mostly earlier in the day; the other ate over a longer period, mostly later.
What this means
When calories were similar, changing the clock did not produce a clear extra benefit here. The trial was small, so it cannot rule out every possible timing effect.
Population & sample
41 adults with obesity and prediabetes or diet-controlled diabetes; all completed. 93% women and 93% Black.
Arms & controls
2 isocaloric feeding arms, allocated 1:1; exact arm counts not extracted. Early 10-hour versus usual extended eating.
Protocol & duration
12 weeks; matched nutrients and calories. Early arm ate 80% before 13:00; comparator at least 50% after 17:00.
Reported results
Weight changes −2.3 versus −2.6 kg; difference +0.3 kg (95% CI −1.2 to 1.9). Glucose measures did not differ.
What limits the inference
Small, single-site study with baseline weight imbalance; estimates do not rule out all timing effects.
Source check & disclosures
Original indexed abstract; NCT03527368. American Heart Association funding.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [42] DOI: 10.7326/m23-3132

Link to this study record
43Pavlou et al. · 2023Human randomized trialTime restriction helped versus control, without clear superiority to calorie restriction.Study design, results & limitations

Effect of Time-Restricted Eating on Weight Loss in Adults With Type 2 Diabetes: A Randomized Clinical Trial

What happened
The eating-window group lost more weight than the usual-diet group. Both dieting approaches improved blood sugar, without a clear difference between them.
Who took part
75 adults with type 2 diabetes and obesity, 25 in each group. Sixty-nine completed six months.
What they compared
One group ate between noon and 8 p.m., one cut calories by counting them, and one continued its usual diet. Blood sugar and medication needs were monitored.
What this means
Beating usual eating does not mean beating another supported diet. These results came with medical monitoring and should not be used to adjust diabetes medicine independently.
Population & sample
75 adults with type 2 diabetes and obesity, 25 per arm; 69 completed.
Arms & controls
3 arms: 8-hour eating, 25% calorie restriction, usual-diet control.
Protocol & duration
Six months; 12:00–20:00 window. Medication management and glucose monitoring accompanied the intervention.
Reported results
Weight difference versus control: −3.56% (95% CI −5.92 to −1.20). Direct comparison with calorie restriction: −1.78% (−4.09 to 0.53). HbA1c improved similarly in both active arms.
What limits the inference
Small study. One arm being significant versus control and another not does not prove the active arms differ.
Source check & disclosures
Full text, outcome table and participant flow; NCT05225337.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [43] DOI: 10.1001/jamanetworkopen.2023.39337

Link to this study record
44Carter et al. · 2018Human randomized trialIntermittent and continuous restriction met the planned HbA1c equivalence test.Study design, results & limitations

Effect of Intermittent Compared With Continuous Energy Restricted Diet on Glycemic Control in Patients With Type 2 Diabetes: A Randomized Noninferiority Trial

What happened
The two diets produced similar long-term blood-sugar test results under the study’s rules. Some people taking insulin or certain other diabetes medicines had episodes of high or low blood sugar.
Who took part
137 adults with type 2 diabetes: seventy assigned to intermittent calorie cutting and 67 to daily calorie cutting. Ninety-seven finished a year.
What they compared
One group ate very little on two days each week. The other reduced food every day. The program included planned medication adjustments and glucose monitoring.
What this means
The result supports a possible supervised alternative, not fasting without medication review. Many people did not finish, and later follow-up did not preserve the same conclusion.
Population & sample
137 adults with type 2 diabetes: 70 intermittent, 67 continuous; 97 completed 12 months.
Arms & controls
2 active dietary arms; no untreated comparator.
Protocol & duration
500–600 kcal on two nonconsecutive days versus daily restriction. Medication-reduction protocol in place.
Reported results
HbA1c difference +0.2 percentage points (90% CI −0.2 to 0.5), within the prespecified ±0.5 margin. Early hypo/hyperglycemic events occurred in 16/46 insulin or sulfonylurea users.
What limits the inference
Substantial attrition. HbA1c equivalence does not prove weight equivalence or safety without medical monitoring.
Source check & disclosures
Full text; ACTRN12615000383561. Read with the 24-month follow-up.
Cohort relationships
Carter’s 12- and 24-month reports follow the same randomized cohort (ACTRN12615000383561).

Read original publication [44] DOI: 10.1001/jamanetworkopen.2018.0756

Link to this study record
45Carter et al. · 2019Human trial follow-upThe same trial’s later follow-up did not preserve the equivalence conclusion.Study design, results & limitations

The effect of intermittent compared with continuous energy restriction on glycaemic control in patients with type 2 diabetes: 24-month follow-up of a randomised noninferiority trial

What happened
Average blood sugar rose above starting levels in both groups. The later results no longer met the study’s rule for saying the diets worked similarly.
Who took part
The same people originally enrolled in the 137-person Carter trial. This paper is follow-up, not another independent trial.
What they compared
Researchers checked the two diet groups at two years, including a year after the supported program ended.
What this means
A short-term improvement may not last after support ends. This follow-up does not establish permanent diabetes control from either schedule.
Population & sample
Follow-up of the original 137-person Carter cohort; exact 24-month completer count not extracted.
Arms & controls
Same two originally randomized dietary groups; no new independent trial.
Protocol & duration
24 months from baseline, including 12 months after the supported intervention ended.
Reported results
HbA1c rose above baseline in both groups. The 90% CI for their difference (−0.31 to 0.83 percentage points) exceeded the ±0.5 equivalence margin.
What limits the inference
Durability and missing follow-up matter. Earlier short-term improvement is not permanent diabetes remission.
Source check & disclosures
Original indexed follow-up abstract; linked to ACTRN12615000383561.
Cohort relationships
Carter’s 12- and 24-month reports follow the same randomized cohort (ACTRN12615000383561).

Read original publication [45] DOI: 10.1016/j.diabres.2019.03.022

Link to this study record
46Teong et al. · 2023Human randomized trialA glucose-tolerance advantage at six months was absent at eighteen.Study design, results & limitations

Intermittent fasting plus early time-restricted eating versus calorie restriction and standard care in adults at risk of type 2 diabetes: a randomized controlled trial

What happened
The intermittent plan improved the body’s handling of a test meal more at six months than daily calorie cutting. That advantage was no longer clear at eighteen months.
Who took part
209 adults at increased risk of type 2 diabetes, divided among three groups.
What they compared
One group ate very little before noon on three days a week, one cut calories daily, and one received a weight-loss booklet. Diet support lasted six months, with a further year of follow-up.
What this means
The early difference did not establish a lasting advantage. The plan changed both food timing and amount, so their individual contributions are uncertain.
Population & sample
209 adults at increased diabetes risk, allocated 2:2:1. Exact arm totals not extracted from the original full text.
Arms & controls
3 arms: intermittent early eating, daily calorie restriction, standard-care booklet.
Protocol & duration
Three days weekly at 30% energy needs between 08:00–12:00 followed by 20 hours without food; six months support and 12 further follow-up.
Reported results
Primary mixed-meal glucose comparison favored intermittent early eating at six months (P=.03), but not at eighteen months. More fatigue was reported.
What limits the inference
Timing and energy restriction were combined. Long-term adherence and loss of the initial difference temper the claim.
Source check & disclosures
Original abstract; NCT03689608. Parent trial of the Bensalem autophagy substudy.
Cohort relationships
Bensalem is an autophagy substudy of the Teong trial (NCT03689608).

Read original publication [46] DOI: 10.1038/s41591-023-02287-7

Link to this study record
47Manoogian et al. · 2024Human randomized trialA small HbA1c improvement accompanied a personalized eating window.Study design, results & limitations

Time-Restricted Eating in Adults With Metabolic Syndrome : A Randomized Controlled Trial

What happened
The eating-window group had a small improvement in a blood test reflecting average blood sugar.
Who took part
122 adults with several diabetes or heart-health risk factors; 108 completed the study.
What they compared
Half added an eight- to ten-hour eating window to their usual care. The other half continued usual care alone. The comparison lasted three months.
What this means
The effect was small, and the study did not measure whether people avoided diabetes or heart disease. Changes to the study’s main test and missing results also need consideration.
Population & sample
122 adults with metabolic syndrome randomized, 61 per arm; 108 completed (54 each).
Arms & controls
2 arms: standard care plus an 8–10-hour eating window versus standard care.
Protocol & duration
Three months; medications could continue. Personalized windows shortened habitual eating duration.
Reported results
HbA1c difference −0.10 percentage points (95% CI −0.19 to −0.003).
What limits the inference
Small effect, complete-case analysis and unadjusted co-primary tests. Primary outcome changed from fasting glucose in December 2021; authors report this preceded analysis.
Source check & disclosures
Full text and reporting history; NCT04057339. Historical registry versions not exhaustively audited.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [47] DOI: 10.7326/m24-0859

Link to this study record
48Guo et al. · 2024Human randomized trialAn intensive meal-replacement program improved early diabetes measures.Study design, results & limitations

A 5:2 Intermittent Fasting Meal Replacement Diet and Glycemic Control for Adults With Diabetes: The EARLY Randomized Clinical Trial

What happened
The food-program group lost more weight and had a larger fall in average blood sugar during the trial.
Who took part
405 adults with early type 2 diabetes, 135 in each of three groups. 332 completed the study.
What they compared
A sixteen-week program combining meal replacements and two low-calorie days per week was compared with two diabetes medicines. The program also changed meals on the other days.
What this means
This was a whole food program, not a test of fasting alone. Its short-term results do not show permanent cure or justify changing medication without clinical advice.
Population & sample
405 adults with early type 2 diabetes, 135 per arm; 332 completed.
Arms & controls
3 arms: 5:2 meal replacement, metformin, empagliflozin. Two medication comparators; no matched dietary control.
Protocol & duration
16 weeks plus eight-week follow-up; intervention also replaced dinners on the other five days.
Reported results
HbA1c fell about 1.9 versus 1.6 versus 1.5 percentage points; weight about 9.7 versus 5.5 versus 5.8 kg.
What limits the inference
Tests a substantial dietary package, not fasting alone. Short duration cannot establish cure or lasting superiority to medication.
Source check & disclosures
Full text and flow. ChiCTR2000040656. Commercial funding from Beijing MetabolicControl Technology.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [48] DOI: 10.1001/jamanetworkopen.2024.16786

Link to this study record
49Parr et al. · 2026Human randomized trialNoninferiority at four months did not persist at twelve months.Study design, results & limitations

Time-restricted eating versus dietetic guidance on glycaemic outcomes in adults at risk of type 2 diabetes: a non-inferiority randomised clinical trial

What happened
At four months, the eating window met the study’s rule for working about as well on blood sugar. At twelve months, the results were too uncertain to reach that same conclusion.
Who took part
247 adults at increased diabetes risk: 124 in the eating-window group and 123 receiving individual diet advice.
What they compared
One group used a nine-hour eating window ending by 7 p.m. The other worked on its diet with professional guidance. Both received five support calls early in the study.
What this means
Blood-sugar changes were small in both groups. An early comparison does not establish lasting diabetes prevention.
Population & sample
247 adults at diabetes risk: 124 time restriction, 123 dietetic guidance.
Arms & controls
2 active programs, each with five telehealth contacts in the first three months.
Protocol & duration
Nine-hour eating window ending by 19:00 versus individualized dietary guidance; 12 months.
Reported results
HbA1c difference −0.02 percentage points (95% CI −0.07 to 0.03) at four months met the 0.10 noninferiority margin. At twelve months +0.05 (−0.02 to 0.11) did not.
What limits the inference
Both programs produced small glycemic changes. Noninferiority at one timepoint is not superiority or durable disease prevention.
Source check & disclosures
Full text; NCT04762251. Public research funding.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [49] DOI: 10.1007/s00125-026-06762-x

Link to this study record
50Wei et al. · 2023Human randomized trialLiver fat improved with calorie reduction; an eating window added no clear benefit.Study design, results & limitations

Effects of Time-Restricted Eating on Nonalcoholic Fatty Liver Disease: The TREATY-FLD Randomized Clinical Trial

What happened
Liver fat fell in both groups. Scans did not show a clear extra benefit from the eating window.
Who took part
88 adults with obesity and excess fat in the liver: 45 in the eating-window group and 43 in the other diet group.
What they compared
Both groups followed calorie targets for a year. One also ate only between 8 a.m. and 4 p.m.; the other kept its usual meal timing.
What this means
Improvement in average liver fat does not mean fasting eliminates fatty liver in everyone. The trial did not establish fewer serious liver complications.
Population & sample
88 adults with obesity and nonalcoholic fatty liver disease: 45 time restriction, 43 daily calorie restriction.
Arms & controls
2 arms with the same calorie prescription; one active dietary control.
Protocol & duration
08:00–16:00 window versus habitual meal timing for 12 months; MRI-measured liver triglyceride was primary.
Reported results
At twelve months, liver fat fell 6.9 versus 7.9 percentage points; between-arm difference +1.0 (95% CI −1.6 to 3.5; P=.45).
What limits the inference
Changes in average liver fat do not establish universal elimination of fatty liver, fibrosis reversal or fewer liver complications.
Source check & disclosures
Original abstract and full text; NCT03786523 / NCT04988230.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [50] DOI: 10.1001/jamanetworkopen.2023.3513

Link to this study record
66Wilkinson et al. · 2026Human randomized trialA recent trial found extra weight loss, with a reporting-history caveat.Study design, results & limitations

Time-Restricted Eating Promotes Weight Loss and Favorable Changes in Adipose in Obesity: The TREAD Randomized Control Trial

What happened
The eating-window group lost about 2.9 kg more weight on average.
Who took part
61 adults started; 52 completed the twelve-week study, 26 in each group.
What they compared
Researchers compared a ten-hour eating window with standard lifestyle advice.
What this means
Researchers changed the trial’s main outcome to weight after the final participant visit, though they said analysis had not begun. That makes the finding less secure than a clearly planned test.
Population & sample
61 randomized; 52 completed, 26 per arm.
Arms & controls
2 arms: 10-hour eating window versus standard lifestyle counseling.
Protocol & duration
12 weeks; behavioral support and self-monitoring.
Reported results
Weight changes −4.59 versus −1.68 kg; difference −2.91 kg (95% CI −4.21 to −1.60).
What limits the inference
Primary outcome changed from adipose catecholamine sensitivity to weight after the last participant visit; authors report the change preceded analysis. This reduces confidence in a strictly prespecified claim.
Source check & disclosures
Full text and linked protocol; NCT04916730. Outcome change dated July 26, 2023; final visit May 26, 2023.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [66] DOI: 10.1002/oby.70228

Link to this study record
93van den Burg et al. · 2024Human randomized trialPeriodic FMD improved diabetes management compared with usual care in a selected population.Study design, results & limitations

Integration of a fasting-mimicking diet programme in primary care for type 2 diabetes reduces the need for medication and improves glycaemic control: a 12-month randomised controlled trial

What happened
After a year, the diet group needed less medication on average and had modestly better blood-sugar control and weight loss.
Who took part
100 people with type 2 diabetes managed with diet or metformin.
What they compared
51 added a five-day low-calorie diet each month; 49 continued usual care.
What this means
Useful evidence for a supervised program in this group. It does not show a cure or prove an advantage over other supported weight-loss diets.
Population & sample
100 adults with type 2 diabetes and BMI at least 27: 51 FMD, 49 control. Eight and ten, respectively, were lost to follow-up.
Arms & controls
2 arms: usual primary care plus monthly FMD versus usual care alone. Participants used diet and/or metformin, not a broad mix of diabetes medicines at baseline.
Protocol & duration
Five-day FMD cycles monthly for 12 months, with clinical medication management. Outcome assessors were masked.
Reported results
Adjusted differences favored FMD: HbA1c −0.3 percentage points (95% CI −0.6 to approximately 0), weight −3.6 kg (−5.2 to −2.1), and lower medication-effect score. Improved glycemic management: 53% versus 8%.
What limits the inference
No equally intensive calorie-restriction comparator; cannot isolate fasting from weight loss and support. Results do not generalize automatically to insulin-treated patients.
Source check & disclosures
Original full text and participant flow checked. NCT03811587. L-Nutra co-funded the study; related MRI and cost reports concern this same trial.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [93] DOI: 10.1007/s00125-024-06137-0

Link to this study record
94Quist et al. · 2024Human randomized trialThe RESET trial did not establish a clinically important weight advantage.Study design, results & limitations

Effects of 3 months of 10-h per-day time-restricted eating and 3 months of follow-up on bodyweight and cardiometabolic health in Danish individuals at high risk of type 2 diabetes: the RESET single-centre, parallel, superiority, open-label, randomised controlled trial

What happened
The main analysis did not show a clear weight advantage after three months, or after the later follow-up.
Who took part
100 middle-aged or older adults at high risk of type 2 diabetes.
What they compared
Half limited eating to ten hours daily; half kept their usual schedule.
What this means
A shorter eating window is not guaranteed to produce meaningful extra weight loss.
Population & sample
100 adults at high diabetes risk, 50 per group; 46 in each completed the three-month intervention.
Arms & controls
2 arms: self-selected ten-hour eating window between 06:00 and 20:00 versus habitual living.
Protocol & duration
Three months of intervention followed by three months of follow-up. External randomization; open-label behavior, masked statistical outcome assessment.
Reported results
Weight difference at three months: −0.8 kg (95% CI −1.7 to 0.2; P=.099). At six months: −0.2 kg (−1.6 to 1.2). A favorable completer analysis did not persist.
What limits the inference
Specific population and schedule. A nonsignificant result does not exclude every possible benefit, but does not support a strong weight-loss promise.
Source check & disclosures
Original indexed abstract checked; NCT03854656. Full-text extraction incomplete.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [94] DOI: 10.1016/s2666-7568(24)00028-x

Link to this study record
101Azhar et al. · 2025Two-year observational follow-up of randomized trialWeight and diabetes control are different outcomes. Missing participants and self-chosen eating after the trial make long-term effects less certain.Study design, results & limitations

Sustained weight reduction following 12 weeks of intermittent fasting intervention in people with insulin-treated type 2 diabetes-Two-year follow-up of the randomised controlled InterFast-2 trial.

What happened
Weight fell about 4.2 kg versus 0.4 kg. Long-term blood-sugar and insulin-dose differences were not statistically clear.
Who took part
39 returning participants from a 46-person trial in insulin-treated type 2 diabetes: 17 originally assigned to fasting and 22 to usual eating.
What they compared
The original groups two years after a twelve-week supervised fasting program.
What this means
Weight and diabetes control are different outcomes. Missing participants and self-chosen eating after the trial make long-term effects less certain.
Population & sample
46 originally randomized: 22 fasting, 24 control; 39 returned: 17 and 22
Arms & controls
Original randomized groups; continued fasting after twelve weeks was self-selected
Protocol & duration
INTERFAST-2: three days weekly at up to 500 kcal before noon, with insulin-adjustment protocol
Reported results
At two years, weight change was −4.2 versus −0.4 kg (P=.021). HbA1c, insulin dose and the composite outcome did not differ significantly. Ten of 17 fasting-group returnees continued at least occasionally.
What limits the inference
Attrition and unregulated post-trial behavior limit causal durability claims. This is follow-up of PMID 36508320, not a new trial; absence of reported severe hypoglycemia in this small follow-up does not establish general safety.
Source check & disclosures
Full article text checked, including methods and reported results; independent duplicate review and all supplements not completed.
Cohort relationships
Two-year follow-up of INTERFAST-2 (PMID 36508320), also reported in PMIDs 39267053 and 41527226; these are not independent trials.

Read original publication [101] DOI: 10.1111/dom.16158

Link to this study record
102Oh et al. · 2025Randomized three-arm fatty-liver trialTimed eating can be one workable strategy. This does not show it is uniquely effective or prevents long-term liver complications.Study design, results & limitations

Efficacy and safety of time-restricted eating in metabolic dysfunction-associated steatotic liver disease.

What happened
Reported relative liver-fat reductions were about 26% with timed eating and 25% with calorie restriction, versus little change with usual care.
Who took part
337 adults were randomized; 333 were analyzed: 113 usual care, 110 calorie restriction and 110 time-restricted eating.
What they compared
Sixteen weeks of a shorter eating window, daily calorie restriction or usual care.
What this means
Timed eating can be one workable strategy. This does not show it is uniquely effective or prevents long-term liver complications.
Population & sample
337 randomized; 333 analyzed: 113 usual care, 110 calorie restriction, 110 time restriction
Arms & controls
Usual care versus daily calorie restriction versus time restriction
Protocol & duration
16 weeks; NCT05579158
Reported results
Liver fat fell about 25.8% relatively with timing restriction versus 0.7% with usual care; calorie restriction produced a similar 24.7% decrease.
What limits the inference
Stronger sample for a surrogate endpoint; timing did not outperform calorie restriction. Relative fat reduction is not a 25.8-percentage-point absolute change or proven cirrhosis prevention.
Source check & disclosures
Original indexed abstract checked; full-text details not independently verified.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [102] DOI: 10.1016/j.jhep.2025.06.005

Link to this study record
103Bonham et al. · 2025Randomized three-arm superiority trialThe best schedule may depend on what a person can maintain. Secondary results should not replace the trial’s main comparison.Study design, results & limitations

Intermittent fasting for weight loss in night shift workers: a three-arm, superiority randomised clinical trial.

What happened
Fasting was not significantly better for the main weight and insulin-resistance outcomes. One schedule improved a cholesterol measure.
Who took part
250 night-shift workers: 85 assigned to daily calorie restriction, 83 to fasting on day shifts and 82 to fasting on night shifts; 170 completed.
What they compared
Three supported eating programs over 24 weeks.
What this means
The best schedule may depend on what a person can maintain. Secondary results should not replace the trial’s main comparison.
Population & sample
250 night-shift workers: 85 continuous restriction, 83 fasting on days off/day shifts, 82 fasting on night shifts; 170 completed
Arms & controls
Continuous restriction versus two modified 5:2 schedules
Protocol & duration
24 weeks; restricted days about 2,100 kJ
Reported results
Neither intermittent schedule outperformed continuous restriction for weight or insulin resistance; one schedule improved cholesterol more.
What limits the inference
32% attrition and wide weight-effect intervals; completer improvements do not establish superiority. Same parent trial as qualitative PMID 40437584.
Source check & disclosures
Original indexed abstract checked; full-text details not independently verified.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [103] DOI: 10.1016/j.ebiom.2025.105803

Link to this study record

Multiday fasting & safety

Complete fasting studies remain smaller and frequently lack a separate control group.

10Pietzner et al. · 2024Human uncontrolled studyA map of biological adaptation, not a timetable of benefits.Study design, results & limitations

Systemic proteome adaptions to 7-day complete caloric restriction in humans

What happened
Participants lost about 5.7 kg on average, and many blood proteins changed after roughly three days.
Who took part
Twelve volunteers: five women and seven men.
What they compared
Everyone completed a seven-day water-only fast. Researchers compared daily blood samples with each person’s earlier samples; there was no separate group eating normally.
What this means
Weight loss included more than body fat. Changes in proteins do not establish a health benefit, a safe fasting length, or the hour when cell recycling begins.
Population & sample
12 volunteers: 5 women and 7 men.
Arms & controls
One fasting cohort; zero separate control groups. Measurements before, daily during, and after fasting.
Protocol & duration
7-day water-only fast with approximately 3,000 blood proteins measured.
Reported results
Average weight loss was 5.7 kg. Broad protein changes became apparent after about 3 days. Genetic/protein analyses suggested potentially beneficial and adverse disease associations.
What limits the inference
No randomized comparator or observed disease reduction. Weight includes water and lean components. Protein changes at day 3 neither locate autophagy onset nor demonstrate that longer fasting improves health.
Source check & disclosures
Full text: study design, Results and Methods. Inferred health associations were distinguished from measured participant outcomes.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [10] DOI: 10.1038/s42255-024-01008-9

Link to this study record
11Scharf et al. · 2022Human uncontrolled studyFavorable and unfavorable markers changed together.Study design, results & limitations

The Effects of Prolonged Water-Only Fasting and Refeeding on Markers of Cardiometabolic Risk

What happened
Weight, blood pressure and one cholesterol measure fell. Other blood-fat and blood-sugar-control measures worsened after eating resumed.
Who took part
48 adults with overweight or obesity enrolled; 26 completed the study.
What they compared
Researchers measured the same people before a supervised water fast, at its end, and after eating resumed. The typical fast lasted about seventeen days. There was no nonfasting group.
What this means
The changes were mixed, and many enrollees did not finish. Without a comparison group, the study cannot establish an overall health benefit from fasting.
Population & sample
48 adults with overweight or obesity enrolled; 26 completed.
Arms & controls
One residential cohort; zero independent control groups. Baseline, end-of-fast and post-refeeding comparisons.
Protocol & duration
Median 17-day water-only fast and 8-day refeeding with whole plant foods, under supervision.
Reported results
Weight, systolic blood pressure and LDL were lower after refeeding. Triglycerides and HOMA-IR were higher; median HOMA-IR was 1.4 initially, 1.1 at fast end, and 3.1 after refeeding.
What limits the inference
Large attrition, short follow-up and a changed refeeding diet prevent causal attribution. HOMA-IR is a surrogate and fasting/refeeding alter its interpretation.
Source check & disclosures
Full text: Results and Table 1 take precedence over ambiguous timing in the abstract. Specialist fasting-center setting limits generalizability.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [11] DOI: 10.3390/nu14061183

Link to this study record
12Commissati et al. · 2025Human uncontrolled studyProlonged fasting was not uniformly anti-inflammatory.Study design, results & limitations

Prolonged fasting promotes systemic inflammation and platelet activation in humans: A medically supervised, water-only fasting and refeeding study

What happened
Weight fell, but blood signs of inflammation and blood-clotting activity increased. They generally moved back toward earlier levels when eating resumed.
Who took part
Twenty adults in a supervised fasting program; fifteen contributed to the detailed blood-protein analysis.
What they compared
Researchers compared measurements before, during and after about ten days of fasting. Six people switched to small amounts of juice or broth. There was no nonfasting group.
What this means
Long fasting did not reduce every sign of inflammation in this group. Blood changes alone do not show whether people will develop or avoid disease.
Population & sample
20 adults (11 women, 9 men); proteomic analyses used 15. Six participants switched to broth and/or juice after symptoms.
Arms & controls
One prospective cohort, zero randomized controls. Comparisons with other datasets were external, not assigned control arms.
Protocol & duration
Mean 9.8-day supervised fast and 5.3-day refeeding; intended water-only protocol had documented deviations.
Reported results
Inflammatory markers and platelet-related pathways rose; hsCRP increased 129%, generally returning toward baseline with refeeding. Weight also fell.
What limits the inference
Small, selected cohort; biomarkers do not prove heart attacks or other clinical harms. Protocol deviations and external datasets preclude clean regimen comparisons.
Source check & disclosures
Full text: participant/protocol Results, Table 1 and Figure 1. Clinic setting and mixed intake retained explicitly.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [12]

Link to this study record
13Finnell et al. · 2018Human retrospective safety studyMedical supervision did not eliminate significant adverse events.Study design, results & limitations

Is fasting safe? A chart review of adverse events during medically supervised, water-only fasting

What happened
Two serious problems required hospital care: dehydration and dangerously low blood sodium. Other significant symptoms were also recorded.
Who took part
Medical records from 652 people covering 768 visits to a supervised water-fasting center.
What they compared
Researchers reviewed symptoms and medical problems during fasting visits. They did not compare patients with a separate group eating normally.
What this means
Medical supervision did not eliminate risk. These records cannot establish that a particular fasting length is safe at home, or how its risks compare with other diets.
Population & sample
652 people contributed 768 eligible visits; 409 were women. Repeat visits are not independent participants.
Arms & controls
One clinic chart series; zero nonfasting control groups.
Protocol & duration
Residential water-only fasting: median 7 days, range 2–41, followed by supervised refeeding.
Reported results
Two serious events required hospitalization: dehydration and hyponatremia. The highest adverse-event grade was 3 in 212 visits and 4 in one visit.
What limits the inference
Retrospective records, selection and repeat visits limit incidence estimates. These observations cannot establish safety at home or causation versus a control.
Source check & disclosures
Full text: Methods, adverse-event Results and Table 2. Counts used because the reported serious-event percentage is inconsistent with its denominator. Authors include clinic-affiliated researchers.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [13] DOI: 10.1186/s12906-018-2136-6

Link to this study record
64Goldhamer et al. · 2001Human uncontrolled studyLarge blood-pressure changes came from an uncontrolled residential program.Study design, results & limitations

Medically supervised water-only fasting in the treatment of hypertension

What happened
Blood pressure fell substantially by the end of the full program. About nine in ten ended below the study’s high-blood-pressure threshold.
Who took part
174 adults with high blood pressure at one residential treatment center.
What they compared
Everyone followed a supervised program including roughly ten to eleven days of water-only fasting, food preparation beforehand and a low-salt plant-based diet afterward. There was no comparison group.
What this means
The result cannot separate fasting from reduced salt, rest, weight loss, medication changes or the later diet. It does not establish long-term benefit or home safety.
Population & sample
174 adults with hypertension in one supervised center.
Arms & controls
One cohort with before/after measurements; zero separate control groups.
Protocol & duration
About 10–11 days water-only fasting, preceded by food preparation and followed by 6–7 days low-fat, low-sodium plant-based refeeding.
Reported results
Average blood pressure fell approximately 37/13 mm Hg by the end of the combined program; about 90% were below 140/90.
What limits the inference
Cannot separate fasting, sodium reduction, weight loss, rest, medication changes and refeeding. No randomized comparison or long-term event outcome.
Source check & disclosures
Original abstract and author-hosted original Table 2; clinic-based study.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [64] DOI: 10.1067/mmt.2001.115263

Link to this study record
67Wilhelmi de Toledo et al. · 2019Human uncontrolled studyA large supervised cohort describes experience, not comparative effectiveness.Study design, results & limitations

Safety, health improvement and well-being during a 4 to 21-day fasting period in an observational study including 1422 subjects

What happened
Weight, blood pressure and some blood measurements improved. Symptoms and medical problems were recorded too.
Who took part
1,422 adults at a supervised fasting clinic, choosing different program lengths.
What they compared
People consumed small amounts of food, about 200–250 calories a day, for four to 21 days. Activity and residential care were also part of the program. There was no nonfasting group.
What this means
A large group is still not a controlled comparison. This was not water-only fasting, and the study cannot establish the best fasting length or safety outside the clinic.
Population & sample
1,422 adults; duration groups in Table 1: 659, 530, 196 and 37.
Arms & controls
Four self-selected duration groups; zero randomized or nonfasting controls.
Protocol & duration
Buchinger modified fasting, 200–250 kcal/day for 4–21 days, with activity and residential care.
Reported results
Weight, blood pressure and some metabolic markers improved; symptoms and adverse events were recorded.
What limits the inference
Selection, clinic setting, concurrent activities and follow-up limit causal and safety claims. Not water-only fasting; no demonstrated optimal duration.
Source check & disclosures
Full text. Table 1 sums to 1,422; a conflicting methods count was not used. Clinic affiliations disclosed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [67] DOI: 10.1371/journal.pone.0209353

Link to this study record
68Gabriel et al. · 2025Human uncontrolled secondary analysisEarly lean-mass loss and later partial recovery complicate weight-loss claims.Study design, results & limitations

Prolonged Water-Only Fasting Followed by a Whole-Plant-Food Diet Promotes Fat-Free Mass Recovery and Continued Fat Mass Loss in Adults With Overweight or Obesity

What happened
Most early weight loss was measured as lean mass rather than fat. Some of that lean mass returned after eating resumed; fat accounted for more of the lasting loss.
Who took part
Forty people enrolled in the larger program; 29 had complete body-composition scans for this analysis.
What they compared
Researchers measured the same people before and after a supervised water fast, typically fourteen days, then after food returned and during later plant-based eating. There was no comparison group.
What this means
Lean mass includes water as well as muscle, so scan changes are not a direct measure of muscle loss. The fast and the later diet cannot be judged separately here.
Population & sample
40 parent-cohort enrollees; 29 with complete DXA data analyzed.
Arms & controls
One repeated-measures cohort; zero independent controls.
Protocol & duration
Median 14 days water-only fasting, six days refeeding, then approximately 45 days follow-up with a plant-based diet.
Reported results
Fat-free mass accounted for 74% of initial weight loss; after partial recovery, fat accounted for 67% of total loss at follow-up.
What limits the inference
DXA fat-free mass includes fluid and glycogen-related changes, not only muscle. Attrition and subsequent diet prevent a fasting-only interpretation.
Source check & disclosures
Full text and participant flow. TrueNorth clinic ownership interests disclosed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [68] DOI: 10.1002/osp4.70086

Link to this study record
77Aweer et al. · 2025Human adverse-event case reportA case illustrates a potential severe harm, without estimating how commonly it occurs.Study design, results & limitations

Wernicke's Encephalopathy Following 40 Days of Water-Only Fasting: A Case Report

What happened
He had a serious brain disorder linked to vitamin B1 deficiency and improved after vitamin and electrolyte treatment.
Who took part
One previously reported 36-year-old man after a 40-day water-only fast.
What they compared
Doctors evaluated his symptoms and scans, then followed his response to treatment.
What this means
Longer fasting is not automatically better. This case shows a possible severe harm, rather than how often it happens.
Population & sample
One 36-year-old man; no comparison group.
Arms & controls
Clinical history, examination, imaging and response to treatment; no randomized or observational control cohort.
Protocol & duration
Presentation after a reported 40-day water-only fast; subsequent intravenous thiamine and electrolyte treatment.
Reported results
Wernicke’s encephalopathy was diagnosed from symptoms and MRI findings. Substantial improvement followed treatment.
What limits the inference
One case cannot quantify risk or determine a safe duration for others. It demonstrates why missing nutrients and refeeding require attention.
Source check & disclosures
Original 2025 indexed abstract checked; detailed full-text nutritional history not extracted.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [77] DOI: 10.1002/ccr3.71049

Link to this study record
109Kaufhold et al. · 2026Randomized waitlist-controlled trial with prediction analysesThis adds evidence beyond intermittent fasting. It was not water-only fasting and did not show superiority to a regular calorie-restricted diet.Study design, results & limitations

Machine learning identifies microbiome and clinical predictors of sustained weight loss following prolonged fasting.

What happened
Some weight and fat reduction remained at twelve weeks. Many metabolic changes returned toward baseline; 83% reported mild, temporary symptoms.
Who took part
38 healthy adults: 19 started immediately and 19 waited before receiving the same program.
What they compared
Five days with small amounts of vegetable juice and broth, compared with a waiting period.
What this means
This adds evidence beyond intermittent fasting. It was not water-only fasting and did not show superiority to a regular calorie-restricted diet.
Population & sample
38 healthy adults, initially 19 immediate fasting and 19 waitlist; all later underwent the intervention. External prediction datasets reuse previously published cohorts.
Arms & controls
Randomized waitlist control; no active calorie-restriction comparator.
Protocol & duration
Five days with 200–250 kcal/day from vegetable juice/broth, then refeeding; twelve-week follow-up. LEANER, NCT04452916.
Reported results
Resting energy expenditure fell about 62 kcal/day during fasting. Some weight/fat reduction remained at twelve weeks, while many metabolic changes reverted. Eighty-three percent reported mild transient adverse events.
What limits the inference
Small exploratory trial with no predetermined power calculation; post-fast diet was incompletely documented. It does not establish superiority to ordinary calorie restriction or validate a clinical prediction service.
Source check & disclosures
Original methods, results and Figure 1 checked; all supplements and independent review not completed.
Cohort relationships
LEANER is NCT04452916. Its external prediction analyses reuse Maifeld, Bahr, GENESIS and OralFast participants; they are not four new fasting trials.

Read original publication [109] DOI: 10.1186/s13073-026-01765-0

Link to this study record

Glycogen, glucose & ketone metabolism

Experiments that test the biological steps behind the cancer-starvation claim.

14Clarke et al. · 2025Human tissue + mouse experimentsA specific tumor dependency does not establish a fasting treatment.Study design, results & limitations

Glycogen drives tumour initiation and progression in lung adenocarcinoma

What happened
More stored sugar was linked with worse outcomes in the human samples. Blocking sugar storage strongly reduced tumor development in mice.
Who took part
Human lung-tumor samples, plus separate mouse experiments. One tissue comparison used 77 tumors of one lung-cancer type and 68 of another; a larger set included 276 of the first type.
What they compared
Researchers examined stored sugar in human tumors. In mice, they changed diets or disabled a gene that lets cells store sugar. They did not test fasting as a treatment in people.
What this means
Disabling a tumor’s sugar-storage gene is different from skipping meals. This study does not show that fasting empties tumors of sugar or cures lung cancer.
Population & sample
Human tissue comparison: 77 adenocarcinomas versus 68 squamous tumors; expanded adenocarcinoma cohort: 276. These cohorts must not simply be added.
Arms & controls
Mouse diet experiment: 2 arms, 6 animals each, sugar/fat gavage versus water vehicle. Genetic experiment: 4 cohorts: KP 6, glycogen-synthesis-deficient KPG 6, KPG vehicle 4, KPG diet 5.
Protocol & duration
Human observational tissue analysis; mice received 2 weeks of dietary exposure before tumor induction, with tumor assessment around day 100.
Reported results
More tumor glycogen correlated with worse human prognosis. Raising glycogen accelerated mouse tumors; deleting Gys1 markedly suppressed tumor development.
What limits the inference
Gene deletion at tumor initiation is not dietary fasting of established cancer. No human fasting, ketone-treatment or cure comparison.
Source check & disclosures
Full text: Figures 1–3, Extended Data 4 and Methods; counts are experiment-specific.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [14] DOI: 10.1038/s42255-025-01243-8

Link to this study record
15Owen et al. · 1969Human physiological studyKetosis does not mean that circulating glucose disappears.Study design, results & limitations

Liver and kidney metabolism during prolonged starvation

What happened
The liver and kidneys kept making blood sugar while ketones increased. Ketones are an alternative fuel made from fat.
Who took part
Eleven people with obesity; a smaller number had detailed liver and kidney measurements.
What they compared
Researchers followed changes during prolonged fasting, including measurements after several weeks. Everyone fasted; there was no separate group eating normally.
What this means
Not eating carbohydrate does not remove all sugar from the bloodstream. This small study did not test cancer treatment or establish a safe fasting duration.
Population & sample
11 people with obesity underwent serial metabolic measurements; invasive organ measurements involved a subset.
Arms & controls
One fasting cohort; zero randomized nonfasting controls. Within-person and organ-balance measurements.
Protocol & duration
Prolonged starvation, with liver/kidney measurements after several weeks.
Reported results
Blood glucose fell and stabilized while ketones increased. Organ measurements supported continued glucose production by liver and kidneys during prolonged fasting.
What limits the inference
Historical, small physiology study, not a safe-duration study or cancer-treatment trial. It undermines the premise that avoiding dietary carbohydrate removes all glucose from a tumor.
Source check & disclosures
Original indexed abstract; detailed individual organ-measurement denominators not extracted.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [15] DOI: 10.1172/JCI106016

Link to this study record
16Faubert et al. · 2017Human tracer + laboratory studyHuman tumors can use fuels beyond directly supplied glucose.Study design, results & limitations

Lactate Metabolism in Human Lung Tumors

What happened
Human lung tumors used lactate, a substance the body makes during normal energy use, as fuel. Tumors differed in which fuels they used.
Who took part
35 people with lung cancer, with related work in cells and mice. Five patients received the special tracing test for lactate.
What they compared
Researchers tracked where fuel molecules went in tumors and nearby lung tissue. This was a fuel-tracing experiment, not a comparison of fasting diets.
What this means
Cancer cells do not all depend on one fuel. The study did not show that changing a person’s diet would stop cancer growth or improve survival.
Population & sample
35 patients with non-small-cell lung cancer; 5 received labeled lactate. Eleven tissue fragments in that analysis came from those 5 people.
Arms & controls
Tracer study, not a randomized diet trial; matched tumor/adjacent lung and glucose/lactate tracing provided comparisons. Zero fasting-control arms.
Protocol & duration
Stable-isotope infusions around tumor resection, with accompanying cell and mouse experiments.
Reported results
Lactate carbon entered tumor energy metabolism. Tumors differed in how they used circulating nutrients.
What limits the inference
Tracer incorporation is a metabolic observation, not evidence that a particular diet changes survival. Tissue fragments are not independent patients.
Source check & disclosures
Full text: Human Subjects and Figure 2. The human tracer experiment is the comparison summarized here; animal cohorts are not pooled into it.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [16] DOI: 10.1016/j.cell.2017.09.019

Link to this study record
17De Feyter et al. · 2016Rat experimentSome brain-tumor models oxidized ketones instead of starving.Study design, results & limitations

A ketogenic diet increases transport and oxidation of ketone bodies in RG2 and 9L gliomas without affecting tumor growth

What happened
The tumors used ketones for energy. Rats on the ketogenic diet did not clearly live longer.
Who took part
Rats with two kinds of brain tumor. The survival comparisons included 21 rats with one tumor type and 19 with the other.
What they compared
For each tumor type, researchers compared a normal diet with a calorie-reduced ketogenic diet: a diet low in carbohydrate that raises fat-derived fuels called ketones.
What this means
This is a direct counterexample to the claim that cancer cells cannot use ketones. It does not predict how every human cancer will respond.
Population & sample
Rats bearing 2 glioma models. Survival cohorts: 9L, 11 standard-diet and 10 ketogenic; RG2, 9 standard and 10 ketogenic. Tracer panels used 7 per group.
Arms & controls
2 diet arms within each tumor model; one standard-diet control per model. No human participants.
Protocol & duration
Calorie-restricted ketogenic feeding began a week after implantation; serial imaging and labeled β-hydroxybutyrate traced metabolism.
Reported results
Tumors transported and oxidized ketones, with increased utilization on ketogenic diet. Neither model showed a significant survival benefit (P=.42 and .48).
What limits the inference
Two rat models cannot predict every human cancer. They do directly contradict the universal claim that cancer cells cannot use ketones.
Source check & disclosures
Full text: diet methods and Figures 2 and 6; survival and tracer denominators kept separate.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [17] DOI: 10.1093/neuonc/now088

Link to this study record
18Dmitrieva-Posocco et al. · 2022Mouse + organoid + human tissue studyA promising ketone-signaling mechanism, not proof of starvation.Study design, results & limitations

β-Hydroxybutyrate suppresses colorectal cancer

What happened
BHB slowed cell multiplication and reduced tumors in mice by changing signals inside cells.
Who took part
Several experiments in mice and lab-grown tissue, plus blood samples from 41 people and detailed tissue samples from nine patients.
What they compared
Researchers tested ketogenic diets or a ketone called BHB in bowel-cancer models. Human samples were studied for clues, not used in a randomized treatment trial.
What this means
This was not evidence that the tumors were starving. A helpful result in mice does not establish that a ketogenic diet treats bowel cancer in people.
Population & sample
Multiple mouse and organoid experiments; human blood cohort 41, with single-cell tissue analyses from 9 patients.
Arms & controls
Diet screens, BHB versus vehicle, and receptor/gene-deficient controls; no randomized human treatment arms. Mouse sample sizes vary by panel and are tabulated in Supplementary Table 3.
Protocol & duration
Ketogenic diets or administered BHB in colorectal-tumor models; organoids and observational human tissue analyses.
Reported results
BHB inhibited proliferation through HCAR2/HOPX signaling and reduced mouse tumor burden. Human samples showed related associations.
What limits the inference
Signaling-mediated growth suppression is different from inability to obtain energy. This was not a human cancer-cure trial. No study-wide animal total is supplied here.
Source check & disclosures
Full text and Extended Data 11; per-panel mouse denominators are in the linked supplement and were not completely extracted.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [18] DOI: 10.1038/s41586-022-04649-6

Link to this study record
19Xia et al. · 2017Mouse + cell experimentsA ketone can also promote growth in a specific tumor model.Study design, results & limitations

Prevention of Dietary-Fat-Fueled Ketogenesis Attenuates BRAF V600E Tumor Growth

What happened
One ketone helped tumors with a particular gene change grow faster.
Who took part
Mice carrying human skin-cancer cells, not people receiving cancer treatment. The highlighted experiments used seven or eight mice in each diet group.
What they compared
Researchers compared normal and high-fat diets, and separately tested ketones, using skin-cancer cells with different gene changes.
What this means
Ketones can have different effects in different cancers. This experiment does not show that ketogenic diets make all human cancers worse.
Population & sample
Human melanoma xenografts in mice; zero human treatment participants. Figure 1A: 8 mice per arm for A375 and 7 per arm for SK-MEL-2; serum assays used 3 per group.
Arms & controls
A375: 3 arms (standard chow, solid high-fat diet, paste high-fat diet). SK-MEL-2: 2 arms (standard, paste high-fat). One standard-diet control per model; separate ketone/vehicle experiments.
Protocol & duration
Diet began a week before tumor-cell injection and continued to the approximately 4-week endpoint.
Reported results
Acetoacetate enhanced BRAF V600E signaling and tumor growth; effects differed by mutation and ketone. The result was not a universal effect of all ketones.
What limits the inference
Preclinical, mutation-specific finding. It does not establish that ketogenic diets worsen human cancer generally. Dietary fat and administered ketones are distinct manipulations.
Source check & disclosures
Full text and original Figure 1 image: sample sizes checked in panel A, separately from serum assays in the legend. Use as a mechanistic counterexample, not a clinical effect estimate.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [19] DOI: 10.1016/j.cmet.2016.12.010

Link to this study record
73Song et al. · 2026Mouse experiments and human tissue observationsFasting reduced bowel-tumor growth in mice through nutrient-stress and immune pathways; the human component did not test fasting treatment.Study design, results & limitations

Intermittent Fasting Reduces Taurine Availability and Promotes Antitumor Immunity Through PERK Attenuation-Associated Paraptosis-Like Cell Death in Colorectal Cancer

What happened
Tumors grew less in fasting mice, with signs of stronger immune attack. Restoring taurine weakened some benefits. Combining fasting with the immune treatment also produced encouraging tumor-control results.
Who took part
Mice with bowel tumors, alongside separate human tissue studies. The main mouse comparison used 11–15 per group. Human analyses included an 85-case tissue set and another set of matched samples from 23 patients.
What they compared
Mice ate normally or fasted twice a week. Other groups also received an immune-based cancer treatment. Researchers added back a nutrient called taurine and altered immune cells to test what mattered.
What this means
The patient samples offered clues, but no patients were assigned fasting as a treatment. The result concerns nutrient stress and immunity; it does not prove that ketones starve cancer or that cell recycling removes tumors in people.
Population & sample
Apc-mutant and implanted colorectal-tumor mouse models plus human specimens. Main intestinal-tumor comparisons: 11–15 mice per group. Human tissue/survival cohort: 85 cases; separate matched tissue cohort: 23 patients. No randomized human fasting intervention.
Arms & controls
Principal mouse comparison: unrestricted feeding versus two weekly fasts. Four-group immunotherapy comparisons: ordinary feeding plus control antibody, ordinary feeding plus anti-PD-1, fasting plus control antibody, fasting plus anti-PD-1. MC38: 3–6 mice/group; CT26: 6/group. Further taurine-restoration, uptake-inhibition, CD8-depletion and gene-manipulation controls tested mechanism.
Protocol & duration
Two nonconsecutive 24-hour fasts weekly with water available. Protocols differed by mouse model; implanted-tumor interventions began after inoculation. Human work compared tissue features and survival associations, not assigned diets.
Reported results
Fasting reduced tumor burden and was associated with lower taurine and greater CD8 immune activity. Taurine replacement weakened the benefit; CD8 depletion impaired it. PERK reduction produced paraptosis-like stress, and combinations with anti-PD-1 improved tumor-control patterns in mice.
What limits the inference
Small, model-specific animal groups and observational human associations. Tumor-control findings do not establish human survival benefit. Taurine was not measured in the same human cohort used for transporter, immune-cell and survival analyses. This mechanism is neither glycogen starvation nor proof of canonical autophagic tumor clearance.
Source check & disclosures
Original full text published October 6, 2026; Figures 1, 2, 4 and 7, methods and limitations checked. Full supplement denominator and outcome extraction remains incomplete.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [73] DOI: 10.1002/advs.78107

Link to this study record
78Xie et al. · 2021Cells, mice and human tissueGlycogen dependence was absent in the tested kidney-cancer models.Study design, results & limitations

Glycogen metabolism is dispensable for tumour progression in clear cell renal cell carcinoma

What happened
The tumors kept growing despite the changes to glycogen, the stored form of sugar.
Who took part
Kidney-cancer cells, human tissue samples and mice. The two main mouse tests compared seven and ten pairs of tumors.
What they compared
Researchers disabled either sugar storage or release from storage and compared the altered tumors with control tumors in the same mice.
What this means
Some tumors can cope without glycogen. This directly challenges a claim that every cancer cell needs it to survive.
Population & sample
Human kidney-cancer cell lines; 20 matched human tumor/normal-tissue pairs; public data from 428 tumors and 66 normal samples. Mouse experiments used 7 tumor pairs for GYS1 deletion and 10 pairs for PYGB/PYGL deletion.
Arms & controls
Control and gene-deleted tumors grew on opposite flanks of each mouse. Cell experiments compared intact versus impaired glycogen synthesis/breakdown across glucose and oxygen conditions; drug and vehicle controls were also used.
Protocol & duration
Genetic disruption, metabolic tracing, culture-growth assays and mouse tumor follow-up. No human fasting intervention.
Reported results
Blocking glycogen synthesis or breakdown did not substantially impair growth in these models, despite abundant tumor glycogen.
What limits the inference
Established kidney-cancer models cannot represent every cancer or disease stage. Absence of dependence here does not invalidate positive findings elsewhere.
Source check & disclosures
Original full text, methods and Figure 4 checked; matched tumor pairs are not independent treatment participants.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [78] DOI: 10.1038/s42255-021-00367-x

Link to this study record
79Chen et al. · 2020Cells, mice and human tissueTargeting a glycogen-related protein showed promise without proving glycogen starvation.Study design, results & limitations

GYS1 induces glycogen accumulation and promotes tumor progression via the NF-κB pathway in Clear Cell Renal Carcinoma

What happened
Lowering the protein slowed growth in these experiments. Higher levels in patient tumors were associated with worse outcomes.
Who took part
Tissue records from 306 people with kidney cancer, plus cells and mice. Patients were not assigned fasting.
What they compared
Researchers raised or lowered a protein involved in storing sugar, then compared growth with control cells and tumors.
What this means
The protein also affects growth signals. This does not show that fasting empties tumors of glycogen or kills them.
Population & sample
306 surgical kidney-cancer cases; separate molecular analyses included 24 paired samples and 12 protein-analysis pairs. Mouse methods describe five animals in the experimental groups, without a fully reconciled total across experiments.
Arms & controls
High versus low tissue expression; GYS1 overexpression or silencing versus vector/scrambled controls in cells and mice. Additional sunitinib combination comparisons.
Protocol & duration
Observational tissue follow-up and gene manipulation; mouse tumors assessed over four weeks. No dietary treatment in patients.
Reported results
Higher GYS1 tracked worse prognosis. Manipulating it changed tumor growth and NF-κB signaling; silencing increased drug sensitivity in cells.
What limits the inference
The authors explicitly did not establish that tumor growth depended on glycogen as fuel. Different models and gene effects may explain disagreement with Xie’s study.
Source check & disclosures
Original full text, human cohort, mouse methods and stated limitations checked. Mouse allocation wording remains ambiguous.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [79] DOI: 10.7150/thno.46825

Link to this study record
80Zhuo et al. · 2026Mouse and cell experimentsA glycogen-related protein’s growth effect was separable from its sugar-storage function.Study design, results & limitations

Glycogen synthase 1 promotes breast cancer progression by promoting IκBα ubiquitination and degradation independent of its canonical enzyme function

What happened
Removing the protein slowed cancer growth. Lowering glycogen through the other methods did not have the same effect.
Who took part
Breast-cancer cells and mice, including five mice in each main tumor-growth group.
What they compared
Researchers removed a sugar-storage protein, then tested other ways to reduce stored sugar.
What this means
The protein’s separate signaling job mattered. A protein linked to glycogen is not, by itself, proof that cancer depends on glycogen for fuel.
Population & sample
Mouse and human breast-cancer cell lines. Tumor-growth comparison: 5 mice per group; lung-metastasis comparison: 4 per group. Cell-assay denominators vary by panel.
Arms & controls
GYS1 deletion versus control; separate deletion of GBE1/GYG1 and drug inhibition of glycogen synthesis; rescue experiments and signaling controls.
Protocol & duration
Gene manipulation and mouse tumor experiments; lung-metastasis assessment at 21 days. No patient fasting trial.
Reported results
GYS1 deletion reduced growth and migration. Other glycogen-lowering interventions did not reproduce this. Experiments implicated a scaffolding role in NF-κB signaling.
What limits the inference
Small, model-specific experiments. A result from deleting a protein cannot automatically be attributed to depletion of its metabolic product.
Source check & disclosures
Original 2026 full text, Figures 1–4 and methods checked. Laboratory subsamples are not human participants.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [80] DOI: 10.1186/s12964-026-02728-z

Link to this study record
81Kaluba et al. · 2025Mouse and cell experimentsKetones can support tumor metabolism, including through an alternative synthesis pathway.Study design, results & limitations

An alternative route for β-hydroxybutyrate metabolism supports cytosolic acetyl-CoA synthesis in cancer cells

What happened
Some tumors used ketones to make material for new cells. Blocking that processing hindered growth.
Who took part
Cancer cells and mice with pancreatic tumors or melanoma. One pancreatic experiment used five groups totaling 38 mice.
What they compared
Researchers tracked labeled ketones and disabled different ways tumors could process them.
What this means
Switching the body toward ketones does not guarantee that a tumor loses its fuel supply.
Population & sample
Cancer cell lines and male mice. Pancreatic tumors implanted in the pancreas: five genotype groups of 8, 7, 7, 8 and 8 mice. Tracing experiments used 3 mice per knockout comparison.
Arms & controls
Control versus disrupted ketone-processing genes; normal versus restricted calories in melanoma experiments. Cell tests varied glucose, fats and ketone availability.
Protocol & duration
Isotope tracing, gene deletion and tumor-growth experiments. Main culture work used 5 mM β-hydroxybutyrate; mouse tracing lasted 6.5 hours.
Reported results
Ketone carbon supplied cytosolic acetyl-CoA and lipids. Disrupting ketone-processing routes reduced growth in selected pancreatic and calorie-restricted melanoma models.
What limits the inference
A mechanistic demonstration, not a patient diet trial. It establishes capability in tested tumors, not that every ketogenic diet worsens cancer.
Source check & disclosures
Original full text, main Figures 1–4 and panel-specific mouse counts checked.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [81] DOI: 10.1038/s42255-025-01366-y

Link to this study record
82Jiang et al. · 2025Human observations and animal experimentsβ-hydroxybutyrate promoted metastasis in these pancreatic-cancer models.Study design, results & limitations

β-Hydroxybutyrate promotes cancer metastasis through β-hydroxybutyrylation-dependent stabilization of Snail

What happened
The ketone helped pancreatic-cancer cells spread in the tested models.
Who took part
People’s blood and tissue samples, cancer cells, and mice. Main mouse comparisons used five animals per group.
What they compared
Researchers added a ketone and tested ways to block its effects on a protein involved in cancer spread.
What this means
Ketones can have effects beyond providing energy. This finding warrants caution, but does not show that fasting causes cancer to spread in people.
Population & sample
Human serum analyses included 68 pancreatic-cancer cases; separate tissue analyses included 50. Main mouse metastasis panels used 5 animals per group; cell assays commonly used 3 biological samples.
Arms & controls
BHB versus vehicle, with genetic and drug controls affecting Snail and CBP. Patient analyses compared ketone levels and disease characteristics observationally.
Protocol & duration
Metabolic profiling, cell-migration assays and mouse metastasis experiments; no randomized human fasting or ketogenic diet.
Reported results
BHB stabilized the Snail protein and increased metastasis in tested models. Interfering with that pathway reduced the effect and improved gemcitabine combinations.
What limits the inference
Patient associations can reflect the disease itself. Animal supplementation and engineered tumors do not establish the effect of fasting on patient survival.
Source check & disclosures
Original full text, Figures 1, 2 and 6 and animal methods checked. Panel denominators should not be added as unique study enrollment.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [82] DOI: 10.1038/s41467-025-61541-3

Link to this study record
83Mashinchian et al. · 2025Laboratory cell experimentsKetone exposure increased resistance to fluorouracil in two cell models.Study design, results & limitations

β-Hydroxybutyrate promotes chemoresistance and proliferation in breast cancer cells

What happened
More cells survived treatment after BHB exposure in these experiments.
Who took part
One human and one dog cancer-cell line grown in the laboratory. No patients took part.
What they compared
Cells received chemotherapy with or without prior exposure to the ketone BHB.
What this means
This is another reason to test ketones by cancer and treatment type. It cannot predict what will happen to a person who fasts.
Population & sample
Human MCF7 and canine CF41.Mg cancer cell lines, grown in flat cultures and three-dimensional clusters. Main viability assays report three independent experiments; no people or live-animal treatment.
Arms & controls
Untreated cells, BHB exposure, fluorouracil alone, and BHB before or after fluorouracil, under different glucose conditions.
Protocol & duration
Cell viability, migration and cluster-size measurements, principally over 48–72 hours.
Reported results
BHB pretreatment increased viability relative to fluorouracil alone and enlarged three-dimensional clusters in both cell lines.
What limits the inference
Two cell lines and one drug cannot establish clinical resistance or the effect of a ketogenic diet. Replicate reporting differs by assay.
Source check & disclosures
Original full text and figure captions checked; culture repeats are not patient numbers.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [83] DOI: 10.1016/j.bbrep.2025.102217

Link to this study record
84Wang et al. · 2026Cells, mice and human tissueKetone use depended on both tumor genetics and the surrounding signals.Study design, results & limitations

Conditional dependency of oncogenic KRAS in driving ketone body catabolism and pancreatic cancer growth

What happened
Under particular conditions, cells used ketones to support energy production and growth.
Who took part
Pancreatic-cancer cells and mice; the main growth test used three groups of five mice.
What they compared
Researchers changed the machinery and signals that let cancer cells process ketones.
What this means
Cancer fuel use can adapt. A ketone-rich environment is not automatically a starvation environment for cancer.
Population & sample
Human pancreatic-cancer cell lines, mouse tumors and human tissue observations. Main mouse experiment: 3 engineered-cell groups, 5 animals each. Human cohort denominator not independently extracted.
Arms & controls
Control, restored normal OXCT1 and a modified OXCT1 protein; cell tests combined BHB with or without IGF1 and altered KRAS/SUCLA2 signaling.
Protocol & duration
Cell experiments included 2 mM BHB for 48 hours. Mouse tumors harvested at day 21.
Reported results
Cooperating signals increased ketone breakdown, acetyl-CoA, ATP and tumor growth. KRAS mutation alone was insufficient.
What limits the inference
Preclinical mechanism; human tissue trends require validation. No patient fasting intervention or survival comparison.
Source check & disclosures
Original 2026 full text, tumor-growth figure and animal methods checked. An antibody-supplier equity interest was disclosed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [84] DOI: 10.1016/j.jbc.2026.113332

Link to this study record
85Zeng et al. · 2026Mouse and cell experimentsKetogenic interventions improved tumor control in a specific bladder-cancer model.Study design, results & limitations

Ketogenic Diet and β-Hydroxybutyrate Suppress Bladder Cancer Growth and Enhance Anti-PD-L1 Immunotherapy Efficacy

What happened
Tumors grew less, and the treatment combinations produced encouraging results.
Who took part
Mice with bladder tumors and cancer cells in the lab. The first diet comparison used six mice per group.
What they compared
Researchers tested ordinary food, a ketogenic diet, added ketones, and combinations with an immune-based cancer treatment.
What this means
This is positive laboratory evidence worth following. It does not establish a cancer treatment for people or a universal ketone effect.
Population & sample
Mouse MB49 tumors and bladder-cancer cell lines. Initial diet comparison: 6 mice per group. Combination-experiment animal totals were not clearly specified in the checked Figure 7 caption.
Arms & controls
Normal versus ketogenic diet. Separate comparisons added BHB, anti-PD-L1 treatment, or their combinations, with normal-diet/control-treatment comparators.
Protocol & duration
Diet began four days after tumor implantation; initial growth measurements covered two weeks. Additional cell and immune experiments investigated mechanism.
Reported results
Diet and BHB reduced tumor growth, increased tumor-cell stress, and enhanced anti-PD-L1-associated control and immune-cell activity.
What limits the inference
Small animal experiments; no human treatment or survival trial. Benefits need not arise from simple sugar starvation.
Source check & disclosures
Original 2026 full text, Figures 1 and 7 and methods checked. Unresolved combination-group denominators are retained explicitly.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [85] DOI: 10.1111/cas.70447

Link to this study record
98Yang et al. · 2022Controlled mouse pancreatic-cancer experimentsA promising combination can work through mechanisms other than starving cancer of all usable fuel. Patient benefit still needs clinical testing.Study design, results & limitations

Ketogenic diet and chemotherapy combine to disrupt pancreatic cancer metabolism and growth.

What happened
Tumors used ketones as fuel. Diet alone did not help survival, but adding it to chemotherapy increased the survival gain in these models.
Who took part
Mice with pancreatic tumors: 28 in the diet-only survival comparison and 66 in the main chemotherapy comparison.
What they compared
Standard food and ketogenic food, tested with and without chemotherapy.
What this means
A promising combination can work through mechanisms other than starving cancer of all usable fuel. Patient benefit still needs clinical testing.
Population & sample
KPC pancreatic-cancer mouse models. Diet-only survival: 14 mice per arm. Pooled diet-plus-chemotherapy comparison: 31 standard-diet and 35 ketogenic-diet mice; separate spontaneous-tumor comparison: 5 versus 4.
Arms & controls
Standard versus ketogenic diet, with and without chemotherapy; separate immune-deficient experiments.
Protocol & duration
Sixteen-hour fast before diet allocation; gemcitabine, nab-paclitaxel and cisplatin on days 4 and 7. Metabolic tracer experiments used separate assay-specific samples.
Reported results
Tumors burned more ketones on ketogenic diet. Diet alone did not slow growth or extend survival; combined with chemotherapy it increased the average survival gain from 5.8 to 16 days in the pooled mouse experiments.
What limits the inference
Mouse results do not establish patient survival benefit. The proposed combination worked despite ketone utilization, not because all tumor cells were unable to use ketones; patent/commercial interests were disclosed.
Source check & disclosures
Original results, animal methods and main figure denominators checked; all supplements and independent review not completed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [98] DOI: 10.1016/j.medj.2021.12.008

Link to this study record
99Liu et al. · 2023Cell and animal colorectal-cancer experimentsAutophagy can be involved in a successful combination without fasting alone clearing the cancer. This was laboratory evidence, not a patient treatment trial.Study design, results & limitations

Fasting-mimicking diet synergizes with ferroptosis against quiescent, chemotherapy-resistant cells.

What happened
Growth slowed, but some cells entered a resting, drug-tolerant state. Adding drugs that trigger iron-related cell death increased killing, with recycling pathways involved.
Who took part
Bowel-cancer cells, patient-derived samples and mice. The first survival comparison used 15 mice; later groups generally used five to seven.
What they compared
Usual feeding, fasting-mimicking food, chemotherapy and combinations with experimental cell-killing drugs.
What this means
Autophagy can be involved in a successful combination without fasting alone clearing the cancer. This was laboratory evidence, not a patient treatment trial.
Population & sample
Colorectal-cancer cells, patient-derived cells/organoids and several mouse models. Main survival comparison: 7 usual-diet and 8 FMD mice; chemotherapy factorial experiment: 5 mice per arm. Further drug combinations used 5–7 mice per arm.
Arms & controls
Usual diet, FMD, chemotherapy and combination; separate ferroptosis-inducing drugs, autophagy/ferroptosis inhibitors and cell-state controls.
Protocol & duration
Two FMD cycles in the first model; each four-day cycle supplied 50% of normal intake on day one and 10% on days two to four. Cell and drug protocols differed.
Reported results
FMD slowed growth, but surviving colorectal-cancer cells could enter a resting, drug-tolerant state. Adding ferroptosis-inducing drugs increased killing, with autophagy implicated by reporter/inhibitor experiments.
What limits the inference
Preclinical, small model-specific groups and experimental drugs; lower proliferation is not complete cancer elimination. Cell-state and inhibitor results do not establish a human fasting duration or clinical treatment.
Source check & disclosures
Original full methods, results and Figures 1–8 checked; supplementary methods and independent review not completed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [99] DOI: 10.1016/j.ebiom.2023.104496

Link to this study record
100Lu et al. · 2023Cell experiments, mouse xenografts and human tissue observationsSome cancers can use ketones. This does not tell us that every ketogenic diet harms patients, but it contradicts a universal starvation claim.Study design, results & limitations

OXCT1 regulates NF-κB signaling pathway through β-hydroxybutyrate-mediated ketone body homeostasis in lung cancer.

What happened
Ketones restored growth under nutrient stress. Disabling the processing enzyme reduced cell and mouse-tumor growth.
Who took part
Lung-cancer cell lines, mice and 13 paired human tumor and nearby-tissue samples.
What they compared
Cells with and without added ketones, and cells with normal or reduced ability to process them.
What this means
Some cancers can use ketones. This does not tell us that every ketogenic diet harms patients, but it contradicts a universal starvation claim.
Population & sample
Lung-cancer lines including H1299; 13 paired human tumors/adjacent tissues; mouse group sizes not specified in main report
Arms & controls
Serum-starved cells with/without ketones; OXCT1 knockdown versus control in cells and grafts
Protocol & duration
Beta-hydroxybutyrate 1 or 5 mM in growth assays; pathway knockdown and rescue experiments
Reported results
Ketones rescued growth of serum-starved lung-cancer cells. Blocking ketone-use enzyme OXCT1 reduced growth and altered NF-kB signaling; high expression was associated with worse survival in human data.
What limits the inference
No patient dietary trial; main figure states three independent experiments, not three patients per treatment. Supplementary animal counts were not verified.
Source check & disclosures
Full article text checked, including methods and reported results; independent duplicate review and all supplements not completed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [100] DOI: 10.1016/j.gendis.2022.04.020

Link to this study record
104Muto-Fujita et al. · 2025Pediatric-cancer cell-line experimentThis supports glucose dependence in particular cells. Fasting in a person does not create a glucose-free bloodstream, and other cancers respond differently.Study design, results & limitations

Ketone Bodies Inhibit Growth of Hepatoblastoma and Rhabdoid Tumor of the Kidney Cells.

What happened
Ketones did not rescue growth when glucose was absent in these models.
Who took part
Three laboratory cell lines representing liver and kidney tumors found in children; no children received a diet intervention.
What they compared
Cells grown with different glucose and ketone conditions.
What this means
This supports glucose dependence in particular cells. Fasting in a person does not create a glucose-free bloodstream, and other cancers respond differently.
Population & sample
Huh6 liver-tumor cells, G401/WT-CLS1 kidney-tumor cells, fibroblast reference
Arms & controls
Normal versus glucose-free media, with/without ketones
Protocol & duration
Cell growth, ATP and ketone-enzyme assays
Reported results
Ketones did not rescue growth in glucose-free media; ATP response differed by tumor line.
What limits the inference
Positive support for poor ketone rescue in these specific cells, not clinical pediatric diet efficacy. Fasting does not make human blood glucose-free; title and conclusion exceed what was tested.
Source check & disclosures
Original indexed abstract checked; full-text details not independently verified.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [104] DOI: 10.21873/anticanres.17673

Link to this study record
107Wang et al. · 2026Thyroid-cancer cell and mouse experimentThis is evidence for a possible anticancer pathway. Slower growth is different from eliminating tumors, and no patient benefit was tested.Study design, results & limitations

Preliminary study on ketone body metabolism in anaplastic thyroid cancer.

What happened
Growth slowed and recycling markers changed. Blocking recycling weakened the cell effects, but the study did not show a clear increase in apoptosis.
Who took part
Two thyroid-cancer cell lines and mice; the reported diet comparison used five mice per group.
What they compared
Ketone treatment with and without a recycling-blocking drug; ketogenic versus standard food in mice.
What this means
This is evidence for a possible anticancer pathway. Slower growth is different from eliminating tumors, and no patient benefit was tested.
Population & sample
Two thyroid-cancer cell lines, generally three experimental replicates. Mouse methods report ten nude mice allocated five per diet; two tumor-line models are displayed, so the total across models is not unambiguously reconciled.
Arms & controls
Acetoacetate dose series with/without chloroquine; ketogenic versus standard diet in mice.
Protocol & duration
Cell treatments 2.5–12.5 mM acetoacetate, with 50 μM chloroquine in combination groups; mouse tumors followed four weeks.
Reported results
Acetoacetate slowed proliferation and arrested the cell cycle without a clear increase in apoptosis. Ketogenic diet reduced mouse tumor growth. Autophagy markers changed and chloroquine weakened the cell effects.
What limits the inference
Marker/inhibitor evidence supports a pathway hypothesis but is not proof of complete autophagic cancer-cell clearance. No human trial; drug off-target effects, doses and ambiguous total animal count limit inference.
Source check & disclosures
Original methods, results and figure captions checked; supplementary materials and independent review not completed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [107] DOI: 10.1530/etj-25-0305

Link to this study record

Cancer treatment studies

Dietary interventions alongside standard treatment; populations and statistical strength differ.

20de Groot et al. · DIRECT · 2020Human randomized phase II trialSome response signals, without improved complete pathological response.Study design, results & limitations

Fasting mimicking diet as an adjunct to neoadjuvant chemotherapy for breast cancer in the multicentre randomized phase 2 DIRECT trial

What happened
Scans suggested better tumor shrinkage in some comparisons. But the diet did not clearly improve the chance of finding no remaining cancer at surgery, or reduce treatment side effects.
Who took part
131 people with a particular type of breast cancer joined. The final comparison included 129: 65 on the special diet and 64 eating normally.
What they compared
Both groups received chemotherapy. One also followed a four-day, very-low-calorie diet designed to mimic fasting around treatment. The other continued its regular diet.
What this means
Only about one in five followed the diet through every treatment cycle, and supportive medicines differed between groups. The trial did not establish longer survival or show that fasting alone treats cancer.
Population & sample
131 randomized with HER2-negative stage II/III breast cancer; 129 analyzed after two exclusions: 65 FMD, 64 regular diet.
Arms & controls
2 arms, one regular-diet control; both received chemotherapy. Dexamethasone use differed between arms.
Protocol & duration
4-day FMD around chemotherapy cycles (3 days before and the treatment day). Only 20% adhered through all cycles.
Reported results
Pathological complete response: 10.8% versus 12.7% (P=.735); toxicity did not significantly differ. Radiological response favored FMD (adjusted OR 3.17, 95% CI 1.06–9.45). A favorable pathological-response analysis was per protocol.
What limits the inference
Adherence selection, multiple endpoints and supportive-medication differences weaken inference. No demonstrated survival benefit or standalone fasting effect.
Source check & disclosures
Full text: participant flow, compliance and tumor-response Results. FMD-related commercial interests disclosed. Registry NCT02126449; distinct from the metabolic DIRECT trial.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [20] DOI: 10.1038/s41467-020-16138-3

Link to this study record
21Voss et al. · ERGO2 · 2020Human randomized trialKetosis occurred, but the primary cancer outcome did not improve.Study design, results & limitations

ERGO2: A Prospective, Randomized Trial of Calorie-Restricted Ketogenic Diet and Fasting in Addition to Reirradiation for Malignant Glioma

What happened
Most people who finished the diet produced more ketones, the fuels made from fat. But cancer control and survival were not clearly better.
Who took part
Fifty people whose brain cancer had returned, assigned to two groups of 25.
What they compared
Both groups received more radiation treatment. One also followed nine days of a low-carbohydrate diet and fasting; the other ate its usual diet.
What this means
Producing ketones is not the same as stopping a tumor. This small trial did not demonstrate a treatment benefit from the diet-and-fasting plan.
Population & sample
50 patients with recurrent malignant glioma randomized 1:1; 4 left before treatment, and 20 completed the dietary intervention.
Arms & controls
2 arms of 25; one usual-diet control. Both arms received reirradiation.
Protocol & duration
9-day intervention: 3 days restricted-calorie ketogenic diet, 3 days fasting, then 3 ketogenic days.
Reported results
17/20 dietary completers developed ketosis. Six-month progression-free survival was 20% versus 16%, without a significant difference; overall survival also did not differ.
What limits the inference
Small and short intervention in heavily treated disease. A favorable exploratory subgroup based on glucose cannot replace the negative randomized result.
Source check & disclosures
Original trial abstract. Registry NCT01754350. Later diet-diary and spectroscopy reports concern the same trial, not independent replication.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [21] DOI: 10.1016/j.ijrobp.2020.06.021

Link to this study record
22Vernieri et al. · 2022Human uncontrolled trialMetabolic and immune changes support further trials, not efficacy claims.Study design, results & limitations

Fasting-Mimicking Diet Is Safe and Reshapes Metabolism and Antitumor Immunity in Patients with Cancer

What happened
Blood sugar, growth-related hormones and some immune-cell measurements changed in ways researchers considered worth studying further.
Who took part
101 selected patients with different cancers. Some detailed blood and tumor tests used smaller groups.
What they compared
Patients continued cancer treatment while repeating a five-day low-calorie diet designed to mimic fasting. There was no randomly assigned usual-diet group.
What this means
Changes in blood or immune cells do not prove that tumors shrink or people live longer. Results from carefully selected, monitored patients do not establish safety for everyone with cancer.
Population & sample
101 selected patients with different cancers in the principal clinical cohort; molecular analyses used smaller subsets.
Arms & controls
Single-arm dietary trial with standard anticancer therapies; zero randomized diet-control arms. Additional tissue comparisons were exploratory.
Protocol & duration
Repeated 5-day FMD cycles with nutritional monitoring and refeeding between cycles.
Reported results
Glucose and growth factors decreased; blood/tumor analyses showed changes consistent with potentially stronger antitumor immune activity.
What limits the inference
Safety and biomarker study, not a controlled survival trial. The title’s “safe” applies to selected monitored participants, not everyone with cancer. Standard treatment, selection and multiple analyses limit inference.
Source check & disclosures
Full text: clinical design and metabolic/immune Results. Commercial FMD relationships disclosed; registry NCT03340935. Online publication 2021, journal issue 2022.
Cohort relationships
Ligorio’s 14-patient survival comparison is a subanalysis of NCT03340935, the broader trial reported by Vernieri; these are overlapping publications.

Read original publication [22] DOI: 10.1158/2159-8290.CD-21-0030

Link to this study record
23Harvie et al. · 2025Human randomized screening trialAn encouraging signal under an exploratory statistical threshold.Study design, results & limitations

Randomised phase-2 screening trial of intermittent energy restriction plus resistance exercise versus resistance exercise alone during chemotherapy for advanced breast cancer

What happened
Cancer stayed controlled for longer in the diet group in this small trial, but the uncertainty was large enough that a real benefit is not yet established.
Who took part
68 women with advanced breast cancer: 35 assigned to the added diet program and 33 to the comparison group.
What they compared
Both groups had chemotherapy and strength exercise. One also received dietitian support, lower food intake on two days a week, and broader diet changes.
What this means
The study was designed to look for an early signal, using a less demanding statistical threshold than a confirmatory trial. It needs a larger follow-up study.
Population & sample
68 women with advanced breast cancer and BMI at least 24; 35 intervention, 33 control.
Arms & controls
2 arms; one active resistance-exercise control with written diet advice. Both received chemotherapy; additional dietitian support was in the intervention arm.
Protocol & duration
2 reduced-energy days weekly plus an energy-restricted Mediterranean-style diet and resistance exercise, continued until progression.
Reported results
Median progression-free survival: 42.0 versus 26.1 weeks; adjusted HR 0.729 (95% CI 0.391–1.361), one-sided P=.160. Some quality-of-life measures favored intervention.
What limits the inference
The trial used one-sided α=.20 to screen for a signal. The confidence interval includes no effect; this is not confirmatory survival evidence. Multiple intervention components.
Source check & disclosures
Full text: Methods and Results. Registry ISRCTN12841416. Food-containing restriction, not complete fasting.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [23] DOI: 10.1038/s41416-025-03129-8

Link to this study record
24Jameson et al. · 2026Human randomized screening trialPromising but imprecise results for ketogenic diet alongside chemotherapy.Study design, results & limitations

A randomized phase II trial of gemcitabine, nab-paclitaxel, cisplatin with or without a medically supervised ketogenic diet for patients with metastatic pancreatic cancer

What happened
The ketogenic group had longer cancer-control and survival times in the trial, but the results were too uncertain to establish a real advantage.
Who took part
36 people with pancreatic cancer that had spread joined; 32 were included in the main comparison, sixteen per group.
What they compared
Both groups received the same chemotherapy. One also followed a medically supervised ketogenic diet, which sharply limits carbohydrate. This was not a fasting study.
What this means
This was a small early trial that needs confirmation. It does not show that ketones starve pancreatic cancer or that diet can replace chemotherapy.
Population & sample
36 randomized; 32 evaluable, 16 per arm, with previously untreated metastatic pancreatic cancer.
Arms & controls
2 arms, one usual-diet control; both received the same triple chemotherapy. Ketogenic intervention included remote clinical support.
Protocol & duration
Ongoing supervised ketogenic diet; target BHB 0.5–3.0 mM. This was not a fasting trial.
Reported results
Median progression-free survival: 8.5 versus 6.2 months, HR 0.53 (95% CI 0.21–1.37), one-sided P=.096. Overall survival: 13.7 versus 10.2 months, HR 0.58 (0.25–1.37).
What limits the inference
Screening α=.20, small evaluable sample and post-randomization exclusions. Both intervals include no effect. Cannot establish ketone-mediated starvation or diet-only benefit.
Source check & disclosures
Full text: statistical analysis and participant Results. Virta Health employment/equity relationships disclosed. Published March 2026.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [24] DOI: 10.1002/cncr.70343

Link to this study record
25Padrão et al. · 2026Mouse experiments + human biomarkersA treatment-sensitizing pathway beyond simply depriving a tumor of fuel.Study design, results & limitations

Fasting boosts breast cancer therapy efficacy via glucocorticoid activation

What happened
Fasting strengthened the effects of hormone treatment in mice through changes in hormone signaling. Human samples showed clues consistent with that idea.
Who took part
Mouse experiments and two human groups of fifteen and 35. One highlighted mouse experiment reported six tumors per group, not necessarily six separate mice.
What they compared
Mice received ordinary food, fasting, hormone treatment, or both fasting and treatment. Human measurements were taken before and after low-calorie diet cycles, without a randomly assigned usual-diet group.
What this means
A possible explanation in mice is not proof of better survival in people. This was research into helping treatment work, not evidence that fasting alone kills cancer.
Population & sample
Mouse xenografts and human cohorts: 15 receiving endocrine therapy plus FMD; 35 in DigesT. Figure 1 reports 6 tumors per arm, not 6 independently verified mice.
Arms & controls
Key mouse experiment: 4 arms: ordinary feeding, fasting, tamoxifen, combination (three comparators to combination). Human comparisons were paired before/after, without randomized diet controls.
Protocol & duration
Mice: weekly 48-hour fasts over 4 weeks with/without tamoxifen. Humans: 5-day FMD cycles.
Reported results
Glucocorticoid-receptor signaling contributed to enhanced endocrine-treatment effects in mice; human hormone and tumor-marker changes were compatible with the mechanism.
What limits the inference
Mechanistic, not a confirmatory human survival trial. Human cohorts include previously studied material and must not be counted as independent efficacy replications.
Source check & disclosures
Full text: Figures 1–2 and clinical methods. Published online December 2025; January 2026 issue.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [25] DOI: 10.1038/s41586-025-09869-0

Link to this study record
59Bauersfeld et al. · 2018Human randomized crossoverA small chemotherapy pilot suggests symptom benefits, with substantial attrition.Study design, results & limitations

The effects of short-term fasting on quality of life and tolerance to chemotherapy in patients with breast and ovarian cancer: a randomized cross-over pilot study

What happened
Some comparisons suggested less fatigue and a smaller drop in quality of life during the low-food periods.
Who took part
Fifty people with breast or ovarian cancer were assigned to treatment schedules; 34 were included in the final analysis.
What they compared
Patients tried periods of very little food around chemotherapy and periods of ordinary eating, in different orders. The low-food periods lasted sixty hours and allowed juice or broth.
What this means
Many participants were not in the final analysis, and the order of treatments may matter. The study did not show that fasting kills tumors or extends life.
Population & sample
50 randomized (27 fasting first, 23 usual diet first); 34 analyzed (18 and 16).
Arms & controls
Two sequences; each patient received fasting and normocaloric periods during different chemotherapy cycles.
Protocol & duration
60 hours around chemotherapy: 36 before and 24 after. Juice/broth permitted up to 350 kcal/day.
Reported results
Quality-of-life and fatigue deterioration was smaller in some fasting comparisons, particularly the fasting-first sequence.
What limits the inference
16/50 not analyzed; unblinded symptoms, sequence effects and accumulated chemotherapy toxicity complicate inference. No tumor-eradication or survival test.
Source check & disclosures
Full text and original Figure 1. Abstract calls 34 randomized; flow diagram establishes 50 randomized and 34 analyzed. NCT01954836.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [59] DOI: 10.1186/s12885-018-4353-2

Link to this study record
60Dorff et al. · 2016Human nonrandomized pilotA feasibility study cannot establish anticancer effectiveness.Study design, results & limitations

Safety and feasibility of fasting in combination with platinum-based chemotherapy

What happened
Reported problems attributed to fasting were generally mild or moderate. Differences in blood-cell damage and low white-cell counts were not clear enough to establish protection.
Who took part
Twenty selected cancer patients: six assigned to a 24-hour fast, seven to 48 hours and seven to 72 hours.
What they compared
Researchers tried different fasting lengths around chemotherapy, allowing small rescue amounts of food when needed. There was no group receiving chemotherapy without fasting.
What this means
The small study mainly asked whether this could be done in selected, monitored patients. It could not establish general safety or better cancer outcomes.
Population & sample
20 selected cancer patients: 6 at 24 hours, 7 at 48 hours, 7 at 72 hours.
Arms & controls
3 sequential duration cohorts; zero nonfasting control groups and no randomization.
Protocol & duration
Around two cycles of platinum chemotherapy; 72-hour schedule included 48 hours before and 24 after. Rescue intake up to 200 kcal/day permitted.
Reported results
Fasting-attributed toxicities were grade 2 or lower. Differences in neutropenia (P=.17) and DNA damage (P=.08) were nonsignificant.
What limits the inference
Excluded vulnerable patients, including those with low BMI or substantial weight loss. Small uncontrolled comparisons do not establish efficacy or general safety.
Source check & disclosures
Full text; NCT00936364. Includes the clinical program reported preliminarily by Cheng.
Cohort relationships
Cheng’s preliminary human blood findings and Dorff’s later pilot concern NCT00936364; they are not independent clinical replications.

Read original publication [60] DOI: 10.1186/s12885-016-2370-6

Link to this study record
61de Groot et al. · 2015Human randomized trialAn early breast-cancer pilot measured blood toxicity, not cancer starvation.Study design, results & limitations

The effects of short-term fasting on tolerance to (neo) adjuvant chemotherapy in HER2-negative breast cancer patients: a randomized pilot study

What happened
Some blood-cell measurements looked better with fasting. Other treatment side effects did not clearly improve.
Who took part
Thirteen women with a particular type of breast cancer: seven in the fasting group and six in the comparison group.
What they compared
Both groups received chemotherapy. One also fasted for 48 hours around treatment; the other followed normal healthy-eating guidance.
What this means
Thirteen people are too few for strong treatment claims. Changes in blood tests do not show that fasting starves cancer or helps patients live longer.
Population & sample
13 women with HER2-negative stage II/III breast cancer; 7 fasting, 6 control.
Arms & controls
2 arms: fasting versus normal healthy-diet guidance, both with chemotherapy.
Protocol & duration
48-hour fast, from 24 hours before until 24 hours after chemotherapy.
Reported results
Some red-cell and platelet measures favored fasting; nonhematological toxicity did not differ. Blood-cell DNA damage was exploratory.
What limits the inference
Very small pilot with multiple laboratory outcomes. Does not establish survival benefit or efficacy of fasting alone.
Source check & disclosures
Full text; NCT01304251. Distinct from the later DIRECT trial.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [61] DOI: 10.1186/s12885-015-1663-5

Link to this study record
86Bahrami et al. · 2024Human randomized trialA small trial reported favorable response and toxicity results for a combined diet-and-supplement program.Study design, results & limitations

Fasting mimicking diet during neo-adjuvant chemotherapy in breast cancer patients: a randomized controlled trial study

What happened
More people in the diet group had a strong tumor response, and some severe side effects were less frequent.
Who took part
44 people receiving chemotherapy for breast cancer, randomly divided into two groups of 22.
What they compared
One group added a low-calorie diet around treatment and omega-3 supplements. The other continued its regular diet.
What this means
Encouraging, but small and not a survival study. The supplements and reporting inconsistencies make the result harder to interpret.
Population & sample
44 people with HER2-negative breast cancer, 22 per arm; diabetes excluded.
Arms & controls
2 arms: chemotherapy plus modified FMD and 2 g omega-3, versus chemotherapy with regular diet. One usual-diet control; no omega-3-only arm.
Protocol & duration
Eight three-week chemotherapy cycles. Full methods describe four dietary days per cycle: about 1,100 kcal on day 1 and 500 on days 2–4.
Reported results
Table 5: high pathological response in 16/22 versus 9/22; imaging response in 15/22 versus 8/22. Severe vomiting: 3/22 versus 12/22; severe neutropenia: 5/22 versus 10/22. No survival endpoint established.
What limits the inference
Small single-center trial, multiple outcomes, and omega-3 cointervention. Table 5 gives P=.03 for response comparisons while prose gives .01; neither should be silently substituted.
Source check & disclosures
Original full text, randomization, intervention and Tables 2 and 5 checked. Registry IRCT20171227038099N1; reporting discrepancies retained.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [86] DOI: 10.3389/fnut.2024.1483707

Link to this study record
87Ligorio et al. · 2024Human nonrandomized comparisonA favorable survival association requires confirmation in a randomized trial.Study design, results & limitations

Adding fasting-mimicking diet to first-line carboplatin-based chemotherapy is associated with better overall survival in advanced triple-negative breast cancer patients: A subanalysis of the NCT03340935 trial

What happened
Survival was longer in the diet group, including after statistical adjustments.
Who took part
14 people receiving a fasting-style diet with chemotherapy, compared with 76 other people receiving chemotherapy.
What they compared
The groups were selected through different care pathways, rather than assigned randomly.
What this means
This is a meaningful positive signal, but differences between the patients could contribute. A randomized survival trial is needed.
Population & sample
14 advanced triple-negative breast-cancer patients from NCT03340935; 76 consecutive patients treated outside that diet group.
Arms & controls
2 nonrandomized cohorts: cyclic FMD plus carboplatin-gemcitabine versus carboplatin-based chemotherapy alone. Statistical adjustment and matching do not recreate randomization.
Protocol & duration
Trial-cohort subanalysis with overall-survival follow-up. Companion to the broader Vernieri cohort, not an independent replication.
Reported results
Median survival 30.3 versus 17.2 months; adjusted hazard ratio 0.40 (95% CI 0.19–0.86).
What limits the inference
Only 14 diet participants, treatment/selection differences and residual confounding. Association does not establish that fasting caused longer survival.
Source check & disclosures
Original indexed abstract checked; full-text allocation history and detailed baseline balance not extracted.
Cohort relationships
Ligorio’s 14-patient survival comparison is a subanalysis of NCT03340935, the broader trial reported by Vernieri; these are overlapping publications.

Read original publication [87] DOI: 10.1002/ijc.34701

Link to this study record
88Ligorio et al. · 2025Human phase II studyBREAKFAST provides a clinical and mechanistic signal, with comparator limits.Study design, results & limitations

Early downmodulation of tumor glycolysis predicts response to fasting-mimicking diet in triple-negative breast cancer patients

What happened
They reported encouraging responses. Tumors that responded showed early changes in how they processed sugar.
Who took part
30 people with early triple-negative breast cancer receiving chemotherapy before surgery.
What they compared
The registered groups both received chemotherapy and the fasting-style diet; one also received metformin. There was no usual-diet group.
What this means
The available summary cannot tell us how much benefit the diet added. Chemotherapy was also being given, and complete study details still need checking.
Population & sample
30 early-stage triple-negative breast-cancer patients in the BREAKFAST trial, NCT04248998.
Arms & controls
Registry specifies 2 randomized experimental arms: chemotherapy plus FMD, with or without metformin. Both received FMD; no chemotherapy-only diet control. Published arm-specific numbers are not supplied in the checked abstract.
Protocol & duration
Five-day severely calorie-restricted cycles every three weeks. Tumor response was primary; longer-term outcomes and tumor molecular analyses were also reported.
Reported results
Authors report favorable pathological complete response and longer-term outcomes. Responding tumors showed early reductions in glycolysis-related activity. Exact response numerator not extracted from the checked source.
What limits the inference
Small cohort; no estimated benefit over a matched standard-diet arm established here. A molecular predictor does not prove glycogen depletion or a ketone-starvation mechanism.
Source check & disclosures
Original indexed abstract and NCT04248998 registry checked. Full article, response counts and historical registry amendments remain incompletely extracted.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [88] DOI: 10.1016/j.cmet.2024.11.004

Link to this study record
89Gabriel et al. · 2024Human case seriesPositive case observations support investigation, not a demonstrated lymphoma treatment.Study design, results & limitations

Prolonged water-only fasting in the management of low-grade follicular lymphoma: a case series

What happened
Their scans improved. One person developed low potassium and needed broth during the fast.
Who took part
Three people with a type of lymphoma that can grow slowly and sometimes change during observation.
What they compared
Doctors compared scans before and after prolonged fasting and a plant-food diet. There was no comparison group.
What this means
These cases deserve study, but cannot tell us whether fasting caused the improvement or how safe it would be for others.
Population & sample
3 adults aged 49, 50 and 56 with low-to-intermediate-grade follicular lymphoma. No control group.
Arms & controls
Before-and-after clinical observations. Fasting was combined with refeeding and a whole-plant-food diet; prior disease and treatment histories differed.
Protocol & duration
Supervised prolonged fasts, including 18- and 21-day periods, with some repeated interventions and subsequent oncology follow-up.
Reported results
Lymph-node size and imaging activity decreased. One patient developed potassium of 2.9 mmol/L and received vegetable broth during the final fasting days.
What limits the inference
Follicular lymphoma can fluctuate or regress during observation. Three selected cases, combined interventions and no control cannot establish efficacy, safety rates or a cancer-clearing mechanism.
Source check & disclosures
Original full text and case histories checked. The first case’s broth means the entire reported fast was not strictly water-only.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [89] DOI: 10.1186/s13256-024-04609-w

Link to this study record
90Herbert et al. · 2026Human randomized feasibility trialA 2026 feasibility report supports further testing, not an efficacy conclusion.Study design, results & limitations

A feasibility randomised controlled trial of water-only fasting prior to CAPOX chemotherapy for stage 2/3 colorectal cancer

What happened
Most fasting participants could follow the plan, but recruiting enough people was difficult.
Who took part
12 people receiving bowel-cancer chemotherapy, randomly split into groups of six.
What they compared
A 36-hour fast before treatment was compared with usual dietary advice.
What this means
The study tested whether a larger trial could work. It was too small to show that fasting improves cancer outcomes.
Population & sample
79 screened, 44 eligible, 12 randomized: 6 fasting and 6 control. Five fasting participants remained for cycles 2 and 3.
Arms & controls
2 arms: 36-hour water-only fasting before CAPOX versus standard dietary advice. Concealed online randomization; laboratory analyses masked.
Protocol & duration
Fasting before the first three chemotherapy cycles. Primary outcomes concerned recruitment, retention, adherence and acceptability.
Reported results
All six followed the first-cycle fast; five continued and adhered in later cycles. Recruitment was difficult, and some viewed fasting as an additional burden.
What limits the inference
Twelve participants cannot establish tumor control, survival benefit or reliable toxicity differences. Pandemic disruption affected recruitment.
Source check & disclosures
Original 2026 full text, randomization and Table 4 checked. ISRCTN17994717.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [90] DOI: 10.1186/s40814-026-01848-0

Link to this study record
91Briata et al. · 2026Human trial interim safety reportThis publication reports early safety, not proof of cancer prevention or treatment.Study design, results & limitations

Time-Restricted Eating and Metformin in Invasive Breast Cancer or DCIS: A Randomized, Phase IIb, Presurgical Trial. Preliminary Safety Analysis

What happened
No side effect met the study’s definition of a dose-limiting problem, although milder symptoms occurred.
Who took part
The first 14 patients in the treatment group of a larger breast-cancer study.
What they compared
They tried nightly fasting with metformin and monitoring before surgery. This report focused on their side effects.
What this means
This supports continuing the trial. It does not yet show that the combination slows cancer.
Population & sample
First 14 participants completing the experimental arm of an ongoing randomized presurgical breast-cancer/DCIS trial.
Arms & controls
Parent trial has two arms: at least 16-hour nightly fasting, nutrition counseling and metformin with glucose monitoring, versus monitoring alone. This interim analysis reports the treated group only.
Protocol & duration
Four to six weeks before surgery, with gradual metformin introduction and monitoring for dose-limiting toxicity.
Reported results
No dose-limiting toxicity or toxicity-related discontinuation in these 14. Eight experienced at least one adverse event; tumor-proliferation efficacy results were not supplied.
What limits the inference
Combination intervention, small selected sample and interim analysis. Absence of severe events here does not establish broad safety or efficacy.
Source check & disclosures
Original full text and interim safety methods checked; indexed journal year 2026, first online publication 2025.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [91] DOI: 10.1158/1940-6207.capr-25-0104

Link to this study record
92Yeary et al. · 2026Human randomized pilotWeight-loss feasibility in cancer patients is different from an anticancer effect.Study design, results & limitations

Intermittent fasting versus continuous calorie energy reduction in prostate cancer patients on active surveillance

What happened
Both groups lost weight and all participants stayed in the study for six months.
Who took part
20 men whose prostate cancer was being monitored, with ten randomly assigned to each diet.
What they compared
Eating much less on two days a week was compared with reducing calories every day.
What this means
The trial shows these programs can be practical. It does not show that either diet slowed the cancer.
Population & sample
20 men with prostate cancer on active surveillance; 10 per arm, all retained through six months. 95% White; mean age about 66.
Arms & controls
2 active arms: 5:2 intermittent restriction versus continuous calorie restriction. No usual-diet control.
Protocol & duration
Two 600-kcal days weekly in the intermittent arm; supported weight-loss programs with three- and six-month assessments.
Reported results
Six-month weight change from baseline: −7.50% with intermittent restriction and −5.79% with continuous restriction. These within-arm results do not establish between-arm superiority.
What limits the inference
Small feasibility study. Exploratory metabolic and immune changes do not establish reduced prostate-cancer progression.
Source check & disclosures
Original 2026 indexed abstract checked. NCT05764330; full-text comparative effect estimates not extracted.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [92] DOI: 10.1093/tbm/ibag069

Link to this study record
96Luo et al. · 2025Human observational cohortBlood ketone associations are not equivalent to the effects of changing diet.Study design, results & limitations

Ketone bodies and cancer risk in the general population: The prevention of renal and vascular end-stage disease (PREVEND) study

What happened
Higher ketones were initially linked with more cancer, but the relationship became uncertain after accounting for other factors.
Who took part
More than 6,000 adults followed for roughly 13 years; 828 developed cancer.
What they compared
Researchers compared cancer rates at different blood-ketone levels. They did not assign diets.
What this means
This cannot tell us whether deliberately raising ketones changes cancer risk.
Population & sample
6,079 PREVEND participants; 828 developed cancer over a median 13.4 years.
Arms & controls
Observed ketone concentrations and subsequent cancer incidence; no assigned intervention or randomized control group.
Protocol & duration
Baseline circulating ketones measured by magnetic resonance spectroscopy; statistical adjustment for demographic, metabolic and inflammatory factors.
Reported results
Higher ketones initially tracked greater risk. The association weakened after adjustment and was no longer statistically significant after including inflammation: HR 1.09 per doubling (95% CI 0.97–1.23).
What limits the inference
Confounding, underlying illness and a baseline measurement limit causal interpretation. It proves neither benefit nor harm from ketogenic diets or fasting.
Source check & disclosures
Original full text and adjusted models checked; the indexed record did not contain an abstract.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [96] DOI: 10.1111/eci.70100

Link to this study record
105Xue et al. · 2025Single-arm breast-cancer feasibility pilotCycles are not people, and feasibility is not proof of safety for everyone or better cancer survival.Study design, results & limitations

The feasibility and safety of fasting-mimicking diet in breast cancer patients with chemotherapy in China.

What happened
Five severe symptom events were reported across the 84 cycles. Some blood and body measurements changed.
Who took part
30 women receiving breast-cancer chemotherapy; 27 completed the study, across 84 diet cycles.
What they compared
Three five-day diet cycles, without a concurrent comparison group.
What this means
Cycles are not people, and feasibility is not proof of safety for everyone or better cancer survival.
Population & sample
95 screened, 30 enrolled, 27 completed; 84 FMD cycles. Later body-composition samples fell to 19.
Arms & controls
Single arm; historical adverse-event assumptions for sample-size calculation, not a concurrent control.
Protocol & duration
Three five-day recipe-based FMD cycles before/on chemotherapy, with 21-day final washout.
Reported results
Five grade III-or-higher self-reported events were recorded over 84 cycles, including fatigue, dizziness, constipation, abdominal pain and leg weakness. Some metabolic/body-composition measures changed; no cancer efficacy was established.
What limits the inference
Events per cycle are not the proportion of patients harmed. No concurrent control and concurrent chemotherapy limit safety attribution; repeated cycles are not independent participants.
Source check & disclosures
Original methods, results, safety denominator and body-composition sample counts checked; independent review not completed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [105] DOI: 10.1007/s10549-025-07756-w

Link to this study record
106Klement et al. · 2026Long-term post-hoc KETOCOMP cohort analysisThe study cannot rule out a modest effect, but it does not establish longer survival from ketogenic eating.Study design, results & limitations

Survival outcomes of rectal and head and neck cancer patients receiving radio(chemo)therapy with a ketogenic diet. A post-hoc analysis from the KETOCOMP trial.

What happened
None of the reported survival comparisons showed a statistically clear advantage, including exploratory matched comparisons.
Who took part
81 people with rectal or head-and-neck cancer: 35 followed ketogenic food and 46 a standard diet.
What they compared
Two nonrandomized diet groups receiving radiation, sometimes with chemotherapy.
What this means
The study cannot rule out a modest effect, but it does not establish longer survival from ketogenic eating.
Population & sample
81 patients: 35 ketogenic, 46 standard diet; rectal/head-neck cohorts
Arms & controls
Nonrandomized diet groups during radio/chemotherapy
Protocol & duration
Median follow-up about 71 months rectal and 77 months head/neck
Reported results
No significant overall, progression-free or locoregional recurrence-free survival benefit in either cancer cohort, including exploratory matched analyses.
What limits the inference
Important longer follow-up of existing KETOCOMP participants, not a new randomized trial. Borderline recurrence/skin-toxicity trends cannot establish efficacy or broad safety.
Source check & disclosures
Original indexed abstract checked; full-text details not independently verified.
Cohort relationships
Long-term KETOCOMP follow-up; overlaps earlier rectal/head-and-neck reports including PMIDs 34233255 and 35499696.

Read original publication [106] DOI: 10.1007/s00066-025-02499-5

Link to this study record
108Dantas et al. · 2026Randomized feasibility trialThis was a feasibility study. It did not establish cancer shrinkage or longer survival, and the extra tissue controls weaken causal interpretation.Study design, results & limitations

Very low-carbohydrate ketogenic diet in treatment-naïve women with endometrial cancer and overweight: a randomized feasibility study.

What happened
Glucose and insulin fell, LDL cholesterol rose, and selected tumor immune-cell measurements changed. No severe diet-related events were reported.
Who took part
19 women were randomized; 15 completed: nine ketogenic and six standard diet. Extra nonrandomized tissue samples were used for one analysis.
What they compared
Three to four weeks of provided ketogenic meals or usual food, both with weekly nutrition counseling.
What this means
This was a feasibility study. It did not establish cancer shrinkage or longer survival, and the extra tissue controls weaken causal interpretation.
Population & sample
775 screened, 19 randomized 2:1; 15 completed: 9 ketogenic and 6 standard diet. RNA analysis used 5 versus 3; immune-cell histology used 7 versus 17, including additional nonrandomized controls.
Arms & controls
Two randomized diets with weekly counseling; twelve extra standard-diet tissue cases were added for histology.
Protocol & duration
Three to four weeks before surgery; provided ketogenic meals intended to meet energy needs, standard-diet arm not provided meals. NCT03285152.
Reported results
Diet adherence was high and no grade 3–4 adverse events were reported. Weight, glucose and insulin fell, LDL rose, and exploratory tumor immune-cell markers changed.
What limits the inference
Feasibility, not tumor-response or survival trial. Small sample, attrition, short duration, tissue-sampling differences and added nonrandomized histology controls limit immune findings.
Source check & disclosures
Original full text, participant flow in prose, diet methods and Figures 1–3 checked; supplement-level allocation counts not fully reconciled.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [108] DOI: 10.1038/s41467-026-73519-w

Link to this study record

Autophagy, regeneration & longevity

Direct human measurements alongside the animal work that informs, but cannot settle, human questions.

26Jamshed et al. · 2019Human randomized crossoverAutophagy-related gene expression is not the same as measured recycling.Study design, results & limitations

Early Time-Restricted Feeding Improves 24-Hour Glucose Levels and Affects Markers of the Circadian Clock, Aging, and Autophagy in Humans

What happened
Earlier eating slightly lowered average blood sugar and changed the activity of a gene linked with cell recycling.
Who took part
Eleven adults with overweight; some blood tests included fewer samples.
What they compared
Each person tried two eating schedules for four days each: 8 a.m. to 2 p.m. and 8 a.m. to 8 p.m. Food was matched between schedules.
What this means
A gene becoming more active does not prove cells recycled more material. The experiment cannot tell us when recycling starts during a fast or whether it improves health.
Population & sample
11 adults with overweight participated; assay-specific sample sizes were smaller for some endpoints.
Arms & controls
2 randomized crossover conditions; each person provided a control period. No independent third group.
Protocol & duration
4 days per condition: 08:00–14:00 eating versus 08:00–20:00, with matched food.
Reported results
Early eating increased morning LC3A gene expression and ketones; mean 24-hour glucose was about 4 mg/dL lower.
What limits the inference
Measured blood-cell gene expression, not autophagic flux. Morning/evening samples cannot identify onset, peak or duration across organs. No cancer or longevity endpoint.
Source check & disclosures
Full text: Methods and gene-expression Results. The 18-hour daily fast was a protocol, not a proven activation threshold.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [26] DOI: 10.3390/nu11061234

Link to this study record
27Bensalem et al. · 2025Human randomized substudyDirect human flux evidence is promising but exploratory.Study design, results & limitations

Intermittent time-restricted eating may increase autophagic flux in humans: an exploratory analysis

What happened
An extra analysis found a difference between intermittent fasting and the booklet group at six months, but not at two months. A decline in the booklet group may explain part of it.
Who took part
121 adults joined this part of a larger trial; 114 had the starting blood measurement. They came from three groups: 47, 44 and 23 people.
What they compared
For six months, people either ate very little before noon on three days a week, cut calories every day, or received a healthy-eating booklet. Researchers measured recycling activity in blood cells.
What this means
This is an early clue, not proof that fasting increases recycling throughout the body. It gives no reliable start hour, best fasting length, cancer benefit or longer-life result.
Population & sample
121 enrolled in the metabolic DIRECT substudy. Baseline flux available for 114: 47 intermittent restriction, 44 daily restriction, 23 standard care; later assay denominators vary.
Arms & controls
3 randomized arms; two comparators: active daily calorie restriction and standard-care booklet. Only dietary intervention arms received individual diet counseling.
Protocol & duration
30% of energy needs before noon then a 20-hour fast on 3 days/week, versus 30% daily restriction, for 6 months. Blood sampled after a standardized overnight fast.
Reported results
Blood-cell flux change favored intermittent restriction versus standard care at 6 months (post hoc P=.04), not 2 months. No significant within-fasting-group increase; control decline may partly explain the difference.
What limits the inference
Blood immune cells do not represent every organ. No clinical cancer, lifespan or hour-by-hour outcome.
Source check & disclosures
Full publisher PDF: Table 1, Methods and Results. Assay-related patents disclosed. NCT03689608, unrelated to breast-cancer DIRECT.
Cohort relationships
Bensalem is an autophagy substudy of the Teong trial (NCT03689608).

Read original publication [27] DOI: 10.1113/JP287938

Link to this study record
28Espinoza et al. · 2025 onlineHuman randomized pilotA small pilot suggests formulation-specific changes in blood cells.Study design, results & limitations

Effect of fasting-mimicking diet on markers of autophagy and metabolic health in human subjects

What happened
One diet showed a favorable change in blood-cell recycling measurements. The result was not consistent across both diets or all sample times.
Who took part
Thirty adults: eleven on one low-calorie diet, ten on another, and nine on a comparison diet.
What they compared
The two diets were designed to mimic parts of fasting while still providing food. Researchers took blood samples during six days of dieting and again after eating resumed.
What this means
Thirty people are too few for a firm conclusion. Blood-cell measurements do not establish whole-body benefits or a precise fasting timetable.
Population & sample
30 adults: ProLon 11, FMD2 10, control 9; 83% women. Baseline BMI differed across arms.
Arms & controls
3 randomized arms: two FMD formulations and one food-containing control diet.
Protocol & duration
6-day dietary intervention; sampling at baseline, days 4 and 6, and day 8 after 48 hours of refeeding.
Reported results
The study reported between-group differences in autophagic-flux markers by day 6; the favorable signal was associated with ProLon and was not consistent across formulations/timepoints.
What limits the inference
Small, unblinded pilot; LC3 protein-ratio assay in blood cells, not whole-body clearance. Does not determine the first hour of activation or a clinically beneficial duration.
Source check & disclosures
Full text: Table 1, autophagy Results and Discussion. Sponsored by L-Nutra; commercial formulations and relationships disclosed. NCT06115551.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [28] DOI: 10.1007/s11357-025-02035-4

Link to this study record
29Vendelbo et al. · 2014Human crossover physiologyA 72-hour muscle sample does not establish peak autophagy.Study design, results & limitations

Fasting Increases Human Skeletal Muscle Net Phenylalanine Release and This Is Associated with Decreased mTOR Signaling

What happened
More of a building block of protein left the muscles during the longer fast. The recycling measurements did not give a clear answer about how much recycling increased.
Who took part
Eight healthy men, each studied after two different fasting periods.
What they compared
Researchers compared an overnight fast with a 72-hour fast. They sampled muscle and measured signs of protein breakdown and cell recycling.
What this means
A muscle sample after three days cannot tell us when recycling peaks. The study also does not support a promise that fasting always protects muscle.
Population & sample
8 healthy men studied under two fasting conditions.
Arms & controls
2 conditions in randomized order: overnight versus 72-hour fasting; each participant served as his own control. Insulin stimulation provided an additional physiological comparison.
Protocol & duration
Muscle biopsies, amino-acid balance and insulin-clamp measurements.
Reported results
Net muscle phenylalanine release increased and mTOR signaling fell. LC3B-II and p62 findings did not allow a clear determination of increased autophagic flux.
What limits the inference
Static protein levels can reflect formation or impaired clearance. Small male-only experiment; increased muscle protein release is not evidence that all muscle is protected during fasting.
Source check & disclosures
Full text: subjects, experimental design and Discussion of LC3B/p62 interpretation.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [29] DOI: 10.1371/journal.pone.0102031

Link to this study record
30Mizushima et al. · 2004Mouse experimentAutophagy varies by tissue and does not simply rise forever.Study design, results & limitations

In vivo analysis of autophagy in response to nutrient starvation using transgenic mice expressing a fluorescent autophagosome marker

What happened
Organs did not respond alike. Some recycling-related measurements rose and then fell; some recycling was already happening before fasting.
Who took part
Mice with a special marker that made parts of the recycling process visible. A key set of measurements used at least five mice per condition.
What they compared
Researchers compared mice eating normally with mice without food for 24 or 48 hours. Water remained available.
What this means
There is no single pattern for the whole body. A day without food also affects a mouse differently from a person, so these times are not human fasting targets.
Population & sample
Reporter mice; Figure 7 quantitative values represent at least 5 mice per condition. No human participants.
Arms & controls
3 feeding states: fed control, 24-hour food deprivation, 48-hour deprivation; one fed comparator for the two fasting durations.
Protocol & duration
Food withheld with water available; fluorescent LC3 structures assessed across organs.
Reported results
Some tissues showed substantial baseline activity. Starvation responses differed: liver markers rose at 24 hours and returned near baseline at 48 hours, while other tissues behaved differently.
What limits the inference
Autophagosome counts are not a universal flux meter. A 24-hour fast in a mouse is not physiologically interchangeable with 24 hours in a person.
Source check & disclosures
Full text: Methods, organ-specific Results and Figure 7. Counts describe measured panels, not one pooled study-wide animal total.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [30] DOI: 10.1091/mbc.e03-09-0704

Link to this study record
31Takamura et al. · 2011Genetically modified mouse studyNormal cellular recycling can restrain tumor initiation.Study design, results & limitations

Autophagy-deficient mice develop multiple liver tumors

What happened
All seventeen mice in the main affected group developed noncancerous liver growths; neither comparison group did.
Who took part
In one main comparison, seventeen mice lacked a working cell-recycling gene. Two comparison groups contained fourteen and four mice.
What they compared
Researchers altered genes and followed liver changes over time. The comparison mice did not have the same complete loss of the recycling gene. No fasting diet was tested.
What this means
Normal cell recycling can help prevent damage. But disabling a gene does not show that increasing recycling through fasting prevents cancer in people.
Population & sample
Key Atg5 comparison after 9 months: 17 deficient mice, 14 heterozygous controls and 4 floxed controls; separate Atg7/p62 experiments also reported.
Arms & controls
3 genotypes in the highlighted experiment; two genetic control groups. No fasting intervention and no human participants.
Protocol & duration
Long-term loss of autophagy genes, with liver pathology and cellular-damage measurements.
Reported results
All 17 deficient mice in the highlighted cohort developed liver tumors, versus neither control group. Tumors were benign adenomas with abnormal mitochondria and damage signals.
What limits the inference
Removing essential autophagy is not the inverse of prescribing a fast. Preventing abnormal cells from arising and treating established malignant disease are different questions.
Source check & disclosures
Full text: Results and Figure 1. Denominators refer to the specified genetic comparison.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [31] DOI: 10.1101/gad.2016211

Link to this study record
32Guo et al. · 2013Genetically modified mouse studyEstablished tumor biology can depend on autophagy for survival.Study design, results & limitations

Autophagy suppresses progression of K-ras-induced lung tumors to oncocytomas and maintains lipid homeostasis

What happened
Removing the recycling gene reduced cancerous tumor growth and changed the kinds of tumors that developed.
Who took part
Mice engineered to develop lung tumors. Key scans used three mice per gene group and timepoint; lung-weight measurements used five.
What they compared
Researchers compared tumors with a working recycling gene against tumors missing that gene, with normal lung tissue as another reference.
What this means
Cell recycling can help an existing tumor survive as well as help healthy cells. This does not prove that fasting worsens human cancer, but it challenges the idea that more recycling is always protective.
Population & sample
Kras-driven mouse lung-tumor models; Figure 1 histology/CT uses 3 mice per genotype/timepoint and lung-weight panels 5.
Arms & controls
Key comparison: autophagy-intact versus Atg7-deficient tumors; one tumor-bearing genetic control, plus normal-lung reference. Additional p53 comparisons are separate experiments.
Protocol & duration
Kras activation with or without Atg7 deletion; tumor progression and mitochondrial/lipid metabolism followed.
Reported results
Atg7 loss reduced malignant tumor burden and shifted lesions toward oncocytomas with defective mitochondria. Autophagy supported tumor metabolic function.
What limits the inference
Genetic models do not show that fasting worsens all cancers or that inhibiting autophagy is universally beneficial. They contradict a universal “more autophagy kills cancer” rule.
Source check & disclosures
Full text: Results and Figure 1. No human participants or fasting-control arms.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [32] DOI: 10.1101/gad.219642.113

Link to this study record
33Imada et al. · 2024Mouse + organoid studyRegeneration and tumor growth can share biological pathways.Study design, results & limitations

Short-term post-fast refeeding enhances intestinal stemness via polyamines

What happened
Eating again boosted repair and cell growth in the gut. With the cancer-protection gene switched off, it also increased tumors.
Who took part
Mice, with seven to thirteen per group in the highlighted tumor experiments. No people received a fasting intervention.
What they compared
Researchers compared ordinary feeding, a 24-hour fast, and eating again after fasting. Some mice also had a cancer-protection gene deliberately switched off.
What this means
Repair and tumor growth can share some of the same machinery. This special mouse model does not establish that ordinary fasting or refeeding causes cancer in people.
Population & sample
Mouse intestinal models; key Figure 5 tumor comparisons used 7–13 mice per group across 3 experiments. No human intervention.
Arms & controls
Fed, 24-hour-fasted, and post-fast-refed states, including different refeeding durations; fed and fasted groups served as comparators for refeeding.
Protocol & duration
Fasting/refeeding with experimentally induced Apc tumor-suppressor loss; lesions assessed at 1 and 3 weeks.
Reported results
Refeeding enhanced stem-cell proliferation and regeneration through mTORC1/polyamine pathways; Apc loss during this state produced more intestinal tumors.
What limits the inference
Requires a specific engineered tumor model. Does not demonstrate that ordinary fasting causes human cancer. Regeneration is not automatically selective for healthy cells.
Source check & disclosures
Full text: Figure 5 and Results; samples and controls vary by experiment.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [33] DOI: 10.1038/s41586-024-07840-z

Link to this study record
34Di Francesco et al. · 2024Randomized mouse lifespan studyStrong animal longevity evidence is still not human longevity evidence.Study design, results & limitations

Dietary restriction impacts health and lifespan of genetically diverse mice

What happened
Both fasting and daily calorie cutting extended lifespan in these mice. Genetic differences influenced lifespan more than diet did.
Who took part
960 female mice were assigned to diets; 937 reached the point when the diets began.
What they compared
Five groups ate freely, fasted one day a week, fasted two days a week, or received one of two daily calorie reductions. Researchers followed their lifespans.
What this means
This is meaningful evidence in female mice, not proof that a fasting schedule adds years to human life. Better health-test results did not always predict longer survival.
Population & sample
960 genetically diverse female mice assigned; 937 survived to intervention at 6 months. Survival groups: 188, 188, 190, 189 and 182.
Arms & controls
5 arms: unrestricted food (188), 1 fasting day/week (188), 2 consecutive fasting days/week (190), 20% calorie restriction (189), 40% restriction (182). One unrestricted control plus active dietary comparisons.
Protocol & duration
Assigned diets maintained through natural lifespan; repeated health assessments.
Reported results
Both intermittent fasting and calorie restriction extended lifespan; genetics had a larger influence on variation in lifespan than diet. Health-marker improvements did not consistently track longer survival.
What limits the inference
Female mice, not people. Does not establish a human fasting duration, an autophagy-mediated explanation, or that maximum restriction is desirable.
Source check & disclosures
Full text: Figure 1, design and Results; original 960 and survival-analysis 937 distinguished.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [34] DOI: 10.1038/s41586-024-08026-3

Link to this study record
35Singh et al. · 2026Human randomized crossoverA plausible nutrient-sensing theory did not predict the human result.Study design, results & limitations

Reduced dietary protein intake does not alter autophagy in human blood: A randomized crossover study in healthy adults

What happened
Eating less protein did not clearly increase the measured recycling activity in blood cells.
Who took part
74 healthy adults started; 63 completed both diets.
What they compared
Each person spent four weeks on a lower-protein diet and four weeks on a higher-protein diet, with a four-week gap between them. Calories were intended to stay similar.
What this means
This was not a fasting experiment. It shows why a theory about nutrient intake needs direct testing before it becomes a promise about cell recycling.
Population & sample
74 healthy adults randomized; 63 completed both periods.
Arms & controls
2 diet conditions; each participant was his/her own control: 10% versus 20% of energy from protein, with intended energy balance.
Protocol & duration
Two 4-week periods, separated by a 4-week washout. Whole-blood immune-cell flux assay.
Reported results
Adjusted flux difference: −8.46 ng LC3B-II/mg protein/hour (95% CI −24.06 to 7.14; P=.28), without a significant benefit from lower protein.
What limits the inference
This is not a fasting trial. It cautions against assuming that manipulating one nutrient-sensing pathway reliably increases human autophagy.
Source check & disclosures
Original journal abstract and PubMed record; published October 2026. Related assay patents disclosed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [35] DOI: 10.1016/j.clnu.2026.106778

Link to this study record
62Cheng et al. · 2014Animal study with human pilotStem-cell regeneration in mice is not a demonstrated human immune-system reset.Study design, results & limitations

Prolonged fasting reduces IGF-1/PKA to promote hematopoietic-stem-cell-based regeneration and reverse immunosuppression

What happened
Fasting-related changes helped protect or rebuild blood-forming cells in mice. The human observations were preliminary.
Who took part
Several mouse experiments, plus very small human blood-sample comparisons during chemotherapy. Human measurements varied from one to six people per fasting group and test.
What they compared
Researchers compared fasting and feeding in mice and altered growth-related signals. The human observations compared fasting durations, without a separate nonfasting group.
What this means
This does not establish a three-day immune-system reset in people. Blood-cell changes are different from proving fewer infections, better cancer outcomes or a younger immune system.
Population & sample
Mice: Figure 1B legend reports n=20, 1E n=12. Human Figure 1F has 1–5 observations per timepoint and 6 per fasting-duration group for the lymphoid/myeloid ratio.
Arms & controls
Mouse fasting versus fed chemotherapy conditions and IGF-1/PKA perturbations; human fasting-duration cohorts without a nonfasting control.
Protocol & duration
Repeated 48-hour fasting cycles in mice; human preliminary 24- versus 72-hour fasting around chemotherapy.
Reported results
Mouse findings connect IGF-1/PKA signaling with hematopoietic stem-cell protection and regeneration; human findings were preliminary lymphocyte observations.
What limits the inference
Species, regimen and endpoint differ from ordinary human fasting. Stem-cell markers do not establish fewer infections or rejuvenation in people.
Source check & disclosures
Original author-hosted PDF, Figure 1 and 2016 correction checked. Figures 2D/3G labels were corrected from MPP to HSC+MPP. Human cohort overlaps Dorff.
Cohort relationships
Cheng’s preliminary human blood findings and Dorff’s later pilot concern NCT00936364; they are not independent clinical replications.

Read original publication [62] DOI: 10.1016/j.stem.2014.04.014

Link to this study record
69Chaudhary et al. · 2022Human and animal experimentLiver and muscle markers did not respond alike.Study design, results & limitations

Intermittent fasting activates markers of autophagy in mouse liver, but not muscle from mouse or humans

What happened
Some recycling-related measurements rose in mouse liver, but muscle measurements in mice and people did not show the same pattern.
Who took part
Fifty women in two fasting programs, plus separate mouse experiments with eight to sixteen mice per group.
What they compared
Women followed repeated 24-hour fasts with different total food amounts for eight weeks. There was no nonfasting human group. Mouse experiments compared fasting with ordinary feeding.
What this means
A result from liver does not tell us what happens in muscle or the whole body. These gene and protein tests also do not measure the complete recycling process or locate its start time.
Population & sample
50 women in two randomized fasting protocols; mice, 8–16 per diet/fasting group.
Arms & controls
Human fasting at 70% versus 100% energy requirements; no nonfasting human arm. Mouse chow/high-fat diets crossed with fasting/ad-libitum conditions.
Protocol & duration
Eight weeks; 24-hour fasts on three nonconsecutive days. Human muscle biopsies after 12 and 24 hours.
Reported results
Some mouse liver markers increased; muscle markers did not. Human muscle gene changes were mixed or reduced with weight loss.
What limits the inference
Static gene/protein measures do not quantify flux or establish onset. Mouse liver results cannot be assigned to human muscle or the whole body.
Source check & disclosures
Original indexed abstract; arm-specific human counts and all tissue denominators not extracted.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [69] DOI: 10.1016/j.nut.2022.111662

Link to this study record
70Pietrocola et al. · 2016Mouse and cell experimentsIn these mice, functioning tumor autophagy was needed for fasting to improve chemotherapy response.Study design, results & limitations

Caloric Restriction Mimetics Enhance Anticancer Immunosurveillance

What happened
Fasting helped chemotherapy slow tumor growth. That extra benefit disappeared when tumor-cell recycling or the relevant immune cells could not work properly.
Who took part
Mice with tumors and cells grown in the lab. The main mouse comparisons used at least five or seven mice per group, depending on the experiment. No people received fasting treatment.
What they compared
Researchers compared usual feeding, fasting, chemotherapy, and fasting with chemotherapy. They repeated the tests after weakening either the mice’s immune response or the tumors’ ability to recycle material inside cells.
What this means
This supports a role for cell recycling in helping treatment work in some settings. It does not show that fasting alone clears cancer, or that the same effect occurs in people.
Population & sample
Mouse tumors and laboratory cells; no human fasting trial. Figure 1: at least 7 mice per group in immunocompetent comparisons, at least 5 per group in immune-deficient and Atg5-knockdown comparisons; one representative experiment of two. These are panel-specific denominators, not a total study enrollment.
Arms & controls
Four conditions within the principal tumor experiments: fed plus vehicle, fasted plus vehicle, fed plus chemotherapy, fasted plus chemotherapy. Comparisons repeated with functional versus impaired immunity and control versus Atg5-knockdown tumor cells. Separate experiments tested calorie-restriction-mimicking compounds.
Protocol & duration
48-hour fasting followed by mitoxantrone or oxaliplatin in mice with palpable tumors. Tumor-growth follow-up; Figure 1 comparisons include days 27 and 31. This mouse fast caused approximately 20% reversible weight loss.
Reported results
Fasting improved chemotherapy-associated tumor control when tumor autophagy and T lymphocytes were functional. Atg5 knockdown or immune deficiency removed the additional benefit. Related compound experiments implicated extracellular ATP signaling and reduced regulatory T-cell suppression.
What limits the inference
Strong mechanistic controls in selected mouse models do not establish human efficacy, a safe fasting duration, or direct digestion of whole tumors by autophagy. A pathway required for immune-assisted treatment response is not proof that increasing it always helps.
Source check & disclosures
Original full text, Figure 1 caption and methods checked. Authors report randomized, blinded mouse experiments and prospective sample-size calculations. Supplementary panels have additional, distinct denominators.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [70] DOI: 10.1016/j.ccell.2016.05.016

Link to this study record
71Buono et al. · 2023Mouse and cell experimentsA fasting-style diet helped chemotherapy in mice with leukemia; reduced tumor autophagy was part of the proposed explanation.Study design, results & limitations

Fasting-Mimicking Diet Inhibits Autophagy and Synergizes with Chemotherapy to Promote T-Cell-Dependent Leukemia-Free Survival

What happened
More than half the normal-weight mice given both treatments survived beyond 110 days without leukemia symptoms. The combination worked better than chemotherapy alone. Reduced recycling inside cancer cells helped explain the benefit.
Who took part
Two main survival experiments involved 50 normal-weight mice and 20 mice with obesity, all with leukemia. Researchers also studied mouse and human cancer cells in the lab.
What they compared
Four groups received ordinary food, a fasting-style diet, chemotherapy, or both diet and chemotherapy. In the 50-mouse experiment, the groups contained 10, 12, 14 and 14 mice.
What this means
This is encouraging evidence for a treatment combination in mice. It shows why “more cell recycling is always better” is too simple. No patients were treated, and researchers knew which treatment each mouse received.
Population & sample
BCR-ABL B-ALL mouse models and mouse/human leukemia cell lines; no human treatment trial. Normal-weight survival experiment: 50 mice. Obese survival experiment: 20 mice. Additional drug, immune-depletion and tissue experiments used separate groups.
Arms & controls
Normal-weight experiment: 4 groups: usual diet plus vehicle (10), fasting-mimicking diet plus vehicle (12), usual diet plus vincristine (14), diet plus vincristine (14). Obese experiment: the same four treatment conditions, 5 mice each, on a high-fat background diet.
Protocol & duration
Four weekly cycles of a 3-day mouse fasting-mimicking diet and weekly vincristine, beginning one week after leukemia-cell injection. Survival followed beyond 110 days. Cell experiments used reduced glucose and serum for 24–48 hours.
Reported results
Over half the normal-weight combination group survived beyond 110 days without leukemia symptoms; survival exceeded vincristine alone (P<.05). Diet alone did not significantly improve median survival (25 versus 23 days), although overall survival differed. Autophagy inhibition reproduced or strengthened treatment effects; CD8-cell depletion reversed leukemia-free survival.
What limits the inference
Preclinical disease and treatment models; no evidence of patient cure. Small groups, no blinding and no sample-size estimation. Results support inhibition of a tumor survival pathway in this setting, not a universal rule that fasting increases autophagy and therefore kills cancer.
Source check & disclosures
Original full text, Figures 1–3 and methods checked. Random group assignment reported. Valter Longo disclosed equity in L-Nutra; other authors reported no competing interests.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [71] DOI: 10.3390/cancers15245870

Link to this study record
72Gates et al. · 2026Laboratory preprint · not peer reviewedNutrient deprivation made lab-grown cell layers expel damaged cells through a process distinct from ordinary autophagic digestion.Study design, results & limitations

Starvation improves epithelial fitness by selectively extruding DNA damaged cells

What happened
The deprived cell layers pushed out more damaged cells. This happened in a short wave. The expelled cells used some recycling-related machinery, but were not simply digested by the usual recycling process.
Who took part
Layers of dog kidney cells and human airway cells grown in the lab. The main time-course test used 12 separate experimental repeats, photographed in 36 areas. No people or live animals fasted.
What they compared
Researchers compared cells in their usual nutrient mixture with cells deprived of nutrients. They also added nutrients back and switched off different cell processes to investigate the cause.
What this means
This early report has not yet been reviewed by other scientists for journal publication. It does not show that fasting removes cancer in people. Its one-to-four-hour lab observations cannot tell someone how long to fast.
Population & sample
Cultured canine kidney and human airway epithelial cell lines, with additional altered-cell experiments; no human participants or whole-animal fasting intervention. Figure 1B–C: 12 biological replicates, 36 imaging fields. Nutrient-addition/removal tests: 3 replicates, 9 fields. KRas-altered cell experiment: 5 replicates, 15 fields.
Arms & controls
Main two-condition comparison: complete growth medium versus nutrient-poor buffer. Additional controls separately removed or restored glucose, glutamine or serum, and altered p53, ATG5, autophagy initiation or lysosomal digestion. Imaging fields are subsamples, not independent people or animals.
Protocol & duration
Microscopy during nutrient deprivation; main time course 4 hours, with nutrient-specific tests at 2 hours. Altered-cell experiments followed 12 hours. The checked manuscript is the September 24, 2026 preprint version.
Reported results
Cell extrusion increased within an hour, peaked around 2 hours and returned to baseline by 4 hours in the cultured model. DNA-damaged and KRas-altered cells were preferentially expelled. ATG5/LC3 machinery contributed, but canonical lysosomal digestion was dispensable; non-expelled cells showed ordinary autophagic recycling.
What limits the inference
Not peer reviewed. These are culture conditions, not a human fasting clock. The authors explicitly leave fasting-induced extrusion in mice and humans unresolved and note that the body maintains glucose production. Removing altered cells from a cultured layer is not demonstrated eradication of established human cancer.
Source check & disclosures
Original preprint full text, Figures 1–3 and discussion checked. Indexed as a preprint, not a completed peer-reviewed clinical study. Authors declared no competing interests.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [72] DOI: 10.64898/2026.09.01.748489

Link to this study record
74Hofer et al. · 2024Human observations and cross-species experimentsA plausible fasting–autophagy pathway, without a proven human longevity benefit or universal timing rule.Study design, results & limitations

Spermidine is essential for fasting-mediated autophagy and longevity

What happened
Spermidine rose in human samples. In experimental organisms, blocking the pathway reduced some recycling and health effects of fasting.
Who took part
Human blood samples from four groups, alongside experiments in cells, insects, worms and mice.
What they compared
Researchers measured a natural molecule called spermidine during fasting, then blocked its production in laboratory experiments.
What this means
This helps explain a possible mechanism. It does not tell a person when beneficial recycling starts or prove that fasting extends human life.
Population & sample
Four human sample cohorts: 109 enrolled; 63 with available samples; 23 from a prior alternate-day-fasting group; and a new small cohort with up to 18 serum samples per time point. Separate yeast, worm, fly, mouse and human-cell experiments.
Arms & controls
Human before/after or 12-versus-36-hour comparisons; no single randomized human longevity trial. Mechanistic experiments used fed/fasted, gene-disruption, inhibitor and rescue controls.
Protocol & duration
Human protocols ranged from 12/36-hour sampling to multiday modified fasting with about 200–350 kcal/day. Animal and cellular protocols differed.
Reported results
Fasting increased spermidine in human samples. Blocking its synthesis or downstream processing impaired fasting-associated autophagy and selected health/lifespan effects in experimental organisms.
What limits the inference
Human metabolite measurements are not demonstrations of whole-body autophagy timing, longer life, or cancer clearance. Cohorts reuse earlier studies; assay-specific sample counts vary.
Source check & disclosures
Original full text, human cohort methods and Figure 1 checked. Multiple commercial interests in autophagy, supplements and fasting services disclosed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [74] DOI: 10.1038/s41556-024-01468-x

Link to this study record
95Ozlu Karahan et al. · 2025Human randomized trialA marker change is weaker evidence than direct measurement of autophagic flux.Study design, results & limitations

Effects of Intermittent Fasting on Liver Steatosis and Fibrosis, Serum FGF-21 and Autophagy Markers in Metabolic Dysfunction-Associated Fatty Liver Disease: A Randomized Controlled Trial

What happened
A blood protein associated with cell recycling increased in the timing group after eight weeks.
Who took part
48 adults with fatty liver disease, divided into groups of 26 and 22.
What they compared
Both groups reduced calories; one also ate within eight hours each day.
What this means
Measuring that protein is not the same as watching recycling happen. This cannot establish a fasting countdown or show cancer removal.
Population & sample
48 adults with metabolic fatty liver disease: 26 energy-plus-time restriction, 22 energy restriction alone.
Arms & controls
2 active dietary arms with similar prescribed calorie content; one also used an eight-hour eating window. Sealed-envelope assignment; laboratory/imaging assessors masked.
Protocol & duration
Eight weeks, with blood markers and liver imaging before and after. Serum ATG-5 and Beclin-1 measured by immunoassay.
Reported results
ATG-5 increased within the time-restricted group; Beclin-1 did not change significantly. Liver-fat measures improved in both arms, with a larger change in the timing arm.
What limits the inference
No direct flux assay, tissue measurement or hourly sampling. A within-arm significant result is not itself evidence of a significant between-arm difference; no cancer outcome.
Source check & disclosures
Original full text, assignment, group counts and marker results checked. Short duration and multiple outcomes limit interpretation.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [95] DOI: 10.3390/life15050696

Link to this study record
97Pietrocola et al. · 2017Human/mouse fasting physiology and blood-cell assay studyDifferent cells and tests give different answers. This does not establish when whole-body autophagy starts, peaks or stops.Study design, results & limitations

Metabolic effects of fasting on human and mouse blood in vivo.

What happened
The special test detected increased recycling after a 24-hour fast. A direct blood-cell measurement at 48 hours did not clearly show it.
Who took part
Healthy volunteers and mice. The key human tests used four volunteers per condition.
What they compared
Researchers compared blood cells before and after fasting, including a test that temporarily blocked the recycling process outside the body.
What this means
Different cells and tests give different answers. This does not establish when whole-body autophagy starts, peaks or stops.
Population & sample
Healthy volunteers and mice; Figure 5 shows four volunteers per condition for paired 48-hour blood comparison and four for 24-hour ex vivo flux assays; mouse panels use three or five animals per condition.
Arms & controls
Fed versus fasting; isolated cells tested with and without a degradation inhibitor.
Protocol & duration
Mice fasted 48 hours; humans up to four days.
Reported results
Direct blood-cell measurements at 48 hours did not clearly detect increased human autophagy. After 24-hour fasting, an ex vivo assay blocking degradation detected increased flux; responses differed by cell type and species.
What limits the inference
Small assay-specific samples and processing outside the body; no universal start, peak or stop time follows. Human four-day fasting and mouse 48-hour starvation are not physiologically equivalent.
Source check & disclosures
Original indexed abstract and original Figure 5–6 captions checked; full methods unavailable through checked repository routes.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [97] DOI: 10.1080/15548627.2016.1271513

Link to this study record

Muscle, cognition & different populations

Trials and observational findings in groups often underrepresented in general weight-loss claims.

51Moro et al. · 2016Human randomized trialLean mass was maintained in a small resistance-training trial.Study design, results & limitations

Effects of eight weeks of time-restricted feeding (16/8) on basal metabolism, maximal strength, body composition, inflammation, and cardiovascular risk factors in resistance-trained males

What happened
The shorter-window group lost more fat. Both groups maintained strength and lean mass, the part of body weight that is not fat.
Who took part
34 men who already trained with weights, seventeen in each group.
What they compared
For eight weeks, both groups lifted weights and reported similar calories and nutrients. One ate its meals in eight hours; the other spread them over a longer day.
What this means
This small study of trained men does not tell us what happens to muscle during multiday fasting or in older, frail or untrained people.
Population & sample
34 resistance-trained men, 17 per arm.
Arms & controls
2 meal-timing arms; both performed resistance training with matched reported energy and macronutrients.
Protocol & duration
Eight weeks; meals at 13:00/16:00/20:00 versus 08:00/13:00/20:00.
Reported results
Fat mass declined more with restricted timing (P=.0448); strength and fat-free mass were maintained in both arms. Testosterone and IGF-1 decreased in the restricted arm.
What limits the inference
Short, small, selected male sample. Reported intake is not fully controlled feeding; findings cannot establish muscle preservation during multiday fasting.
Source check & disclosures
Full text, methods and results; eight-week outcome only.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [51] DOI: 10.1186/s12967-016-1044-0

Link to this study record
52Haganes et al. · 2022Human randomized trialCombining exercise and time restriction helped secondary outcomes, not the primary test.Study design, results & limitations

Time-restricted eating and exercise training improve HbA1c and body composition in women with overweight/obesity: A randomized controlled trial

What happened
The combined plan improved some body-fat and blood-sugar measurements. But the main test of how the body handled a sugary drink found no clear advantage.
Who took part
131 women with overweight or obesity, divided among four groups.
What they compared
For seven weeks, the groups tried a ten-hour eating window, intense exercise three times a week, both together, or neither.
What this means
A few positive measurements do not turn the main negative result into a proven benefit. Longer studies are needed to understand lasting effects.
Population & sample
131 women with overweight or obesity; original arm counts not extracted.
Arms & controls
4 arms: 10-hour eating window, high-intensity intervals, both, or no intervention. Three comparators for the combined program.
Protocol & duration
Seven weeks; exercise three times weekly. Primary outcome: oral-glucose-tolerance area under the curve.
Reported results
Primary comparisons with control were nonsignificant (P=.36, .057 and .14). Combined treatment favored some secondary fat and HbA1c measures.
What limits the inference
Short duration and multiple outcomes; secondary findings should not overwrite a null primary outcome.
Source check & disclosures
Original indexed abstract; NCT04019860.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [52] DOI: 10.1016/j.cmet.2022.09.003

Link to this study record
53Manoogian et al. · 2022Human randomized trialAn eating window was feasible for selected firefighters working 24-hour shifts.Study design, results & limitations

Feasibility of time-restricted eating and impacts on cardiometabolic health in 24-h shift workers: The Healthy Heroes randomized control trial

What happened
People in the eating-window group shortened their eating day to about eleven hours on average. Some health measurements improved, especially in people starting with higher risk.
Who took part
150 firefighters working 24-hour shifts joined; 137 completed the study.
What they compared
Both groups received Mediterranean-diet advice. One also aimed to eat within ten hours each day for twelve weeks; the other did not add that target.
What this means
This suggests the plan can fit some shift schedules. It does not establish the same benefit for every shift worker, and most did not achieve the exact ten-hour target.
Population & sample
150 randomized; 137 completed, 70 time restriction and 67 standard care.
Arms & controls
2 arms; both received Mediterranean-diet advice, one also targeted a 10-hour window.
Protocol & duration
12 weeks; intervention participants shortened eating from about 14.1 to 11.1 hours on average.
Reported results
Reported benefits included VLDL particle size and selected cardiometabolic measures in participants with elevated baseline risk.
What limits the inference
Subgroup and secondary results do not establish a universal benefit for shift workers. Achieved windows differed from the target.
Source check & disclosures
Full text and flow; NCT03533023.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [53] DOI: 10.1016/j.cmet.2022.08.018

Link to this study record
54Lister et al. · 2024Human randomized trialIntermittent restriction did not outperform continuous restriction in adolescents.Study design, results & limitations

Intermittent Energy Restriction for Adolescents With Obesity: The Fast Track to Health Randomized Clinical Trial

What happened
Both groups had similar improvements in body size relative to age and height. The intermittent schedule did not clearly work better.
Who took part
141 teenagers with obesity: 71 in the intermittent group and seventy in the daily-diet group. Ninety-seven completed a year.
What they compared
Both groups received intensive care from a clinical team and began with the same very-low-calorie phase. They then followed different calorie-cutting schedules.
What this means
This was closely supported adolescent care, not evidence for teenagers fasting on their own. Many participants did not finish the year.
Population & sample
141 adolescents: 71 intermittent and 70 continuous restriction; 97 completed (43 and 54).
Arms & controls
2 intensive multidisciplinary programs; one active dietary comparator.
Protocol & duration
52 weeks; both began with four weeks of very-low-energy intake before differing restriction schedules.
Reported results
BMI z score decreased by about 0.28 in each group, without a between-group advantage.
What limits the inference
Substantial attrition and intensive clinical support. Does not justify unsupervised adolescent fasting; psychological monitoring was part of care.
Source check & disclosures
Full text and flow; ACTRN12617001630303. Related eating-behavior reports reuse this cohort.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [54] DOI: 10.1001/jamapediatrics.2024.2869

Link to this study record
55Houston et al. · 2026Human randomized trialAn older-adult pilot addresses feasibility, not frailty prevention.Study design, results & limitations

Caloric restriction and time-restricted eating in older adults with overweight or obesity: The Health, Aging, and Later-Life Outcomes Pilot Study

What happened
Weight and some movement tests changed across the groups. The calorie-cutting groups also lost some lean tissue, which includes muscle and water.
Who took part
Ninety adults aged sixty or older with overweight or obesity. About 92% stayed in the nine-month trial.
What they compared
Three groups tried calorie cutting with in-person support, calorie cutting with remote support, or an eight-hour eating window.
What this means
This was a small trial designed mainly to see whether the plans were practical. It could not reliably rank them or establish that they prevent frailty.
Population & sample
90 adults aged at least 60 with overweight or obesity; nine-month retention 92%. Exact allocation per arm not extracted.
Arms & controls
3 arms: prescribed 20% calorie restriction in person, remotely, or an 8-hour eating window.
Protocol & duration
Nine months; actual calorie reduction in the calorie-restriction groups was much smaller than prescribed.
Reported results
Within-group weight changes: −4.4, −6.7 and −1.0 kg, respectively. Some functional measures improved; lean mass fell in calorie-restriction groups.
What limits the inference
Pilot was not powered to rank effectiveness. Within-group changes do not establish a between-group benefit or safety in frail adults.
Source check & disclosures
Full text; HALLO feasibility trial. Outcomes should be interpreted as exploratory.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [55] DOI: 10.1093/gerona/glag061

Link to this study record
56Wang et al. · 2025Human randomized trialA small pilot found a cognitive-test signal that needs replication.Study design, results & limitations

Effects of 15:9 time-restricted eating on cognitive function and anthropometric parameters among the elderly with mild cognitive impairment: an assessor-blinded, pilot randomized controlled trial

What happened
The eating-window group improved on a general thinking test and some memory measures.
Who took part
46 older adults with mild memory or thinking problems, 23 in each group.
What they compared
One group ate within nine hours each day for twelve weeks. The other continued its usual eating pattern. Researchers then compared memory and thinking tests.
What this means
This is an encouraging small pilot, not proof that fasting prevents dementia. The study did not establish lasting improvements in everyday independence.
Population & sample
46 older adults with mild cognitive impairment, 23 per arm.
Arms & controls
2 arms: 15:9 time restriction versus usual eating; outcome assessor blinded.
Protocol & duration
12 weeks; cognitive and memory tests, with anthropometric measures.
Reported results
MoCA between-group difference 1.70 points (95% CI 0.46 to 2.94; P=.007); some memory measures improved.
What limits the inference
Small, short study with multiple tests. A test-score change does not establish prevention of dementia or lasting everyday functional improvement.
Source check & disclosures
Full text; ChiCTR2100052766.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [56] DOI: 10.1186/s12877-025-06788-7

Link to this study record
57McLeod et al. · 2026Human randomized trialRepeated cognitive tests improved in both groups, without a fasting advantage.Study design, results & limitations

Impact of a 4-week time-restricted eating (16:8) intervention on working memory and attentional control in adult women: A randomized controlled pilot

What happened
Both groups improved with repeated testing. The eating window did not add a clear memory or attention benefit.
Who took part
43 healthy adult women: 23 in the eating-window group and twenty in the comparison group.
What they compared
For four weeks, one group used an eight-hour eating window and the other kept its usual habits. Both repeated memory and attention tests weekly.
What this means
Practice can improve test scores even when a diet has no effect. This brief study does not settle what happens in older people or those with memory disorders.
Population & sample
43 healthy women: 23 time restriction, 20 control.
Arms & controls
2 arms: 16:8 eating versus usual behavior; one control.
Protocol & duration
Four weeks with weekly cognitive testing.
Reported results
Cognitive performance improved over repeated testing in both groups, without an intervention-specific effect; BMI declined with time restriction.
What limits the inference
Brief, small study in healthy women; practice effects and population differences limit comparison with cognitive-impairment trials.
Source check & disclosures
Original indexed abstract; full-text methods and attrition not extracted.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [57] DOI: 10.1016/j.actpsy.2026.107510

Link to this study record
58Chen et al. · 2025Human observational cohortShort reported eating duration was associated with cardiovascular mortality; causality is unresolved.Study design, results & limitations

Association of eating duration less than 8 h with all-cause, cardiovascular, and cancer mortality

What happened
People reporting very short eating windows had more heart-related deaths in the analysis.
Who took part
19,831 US adults from an existing health survey, followed through records for about eight years on average.
What they compared
Researchers estimated daily eating windows from two food recalls and compared later deaths across six eating-duration groups. They did not assign anyone to fast.
What this means
A link is not proof that fasting caused the deaths. Illness, other habits and inaccurate food recall may explain it, and two reported days do not establish a long-term fasting routine.
Population & sample
19,831 US adults from NHANES; median follow-up 8.1 years.
Arms & controls
Six observed eating-duration categories, with 12–14 hours as reference; zero randomized intervention groups.
Protocol & duration
Eating duration inferred from two dietary recalls, linked to later mortality records.
Reported results
Under eight hours was associated with cardiovascular mortality: adjusted HR 2.35 (95% CI 1.39–3.98). Other mortality associations were not consistently robust.
What limits the inference
Two recalls do not identify a sustained intentional fasting regimen. Illness, lifestyle, measurement error and reverse causation can explain associations.
Source check & disclosures
Original peer-reviewed abstract; not merely the earlier conference headline. Full covariate and sensitivity tables not extracted.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [58] DOI: 10.1016/j.dsx.2025.103278

Link to this study record
63Choi et al. · 2016Animal study with human randomized pilotPromising autoimmune-disease biology, with only preliminary human evidence.Study design, results & limitations

A Diet Mimicking Fasting Promotes Regeneration and Reduces Autoimmunity and Multiple Sclerosis Symptoms

What happened
Some quality-of-life measures improved in the human diet groups. The mouse work also suggested changes in immune activity and nerve repair.
Who took part
Sixty people with multiple sclerosis, a condition that damages nerve coverings, divided into three groups of twenty. Separate experiments used mice.
What they compared
People tried usual eating, six months of a ketogenic diet, or one seven-day low-calorie diet followed by Mediterranean eating for six months.
What this means
This small human pilot did not establish a cure or reliable prevention of relapses. The later diet changes make it impossible to assign the results to the fasting-style week alone.
Population & sample
60 people with relapsing-remitting multiple sclerosis, 20 per arm. Separate mouse experiments; Figure 3 used 5–6 mice/group.
Arms & controls
Human arms: control diet, ketogenic diet, one FMD cycle followed by Mediterranean diet. Two comparators; separate mouse diet/transfer controls.
Protocol & duration
Human FMD lasted seven days, followed by six months Mediterranean eating; ketogenic arm lasted six months.
Reported results
Human quality-of-life signals accompanied mouse immune and myelin findings. Eight human relapses occurred: four control, one ketogenic, three FMD.
What limits the inference
Pilot was not a confirmatory relapse or disability trial. Subsequent diet confounds fasting attribution; mouse regeneration is not human cure.
Source check & disclosures
Original full-text human-results section and Figure 3; NCT01538355. FMD commercial interests disclosed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [63] DOI: 10.1016/j.celrep.2016.05.009

Link to this study record
65Salti et al. · 2004Human observational surveyRamadan observations show why diabetes medication and fluid context matter.Study design, results & limitations

A population-based study of diabetes and its characteristics during the fasting month of Ramadan in 13 countries: results of the epidemiology of diabetes and Ramadan 1422/2001 (EPIDIAR) study

What happened
Severe low-blood-sugar episodes were reported more often during Ramadan than during other months.
Who took part
12,243 people with diabetes across thirteen countries were included in the analysis: 1,070 with type 1 and 11,173 with type 2 diabetes.
What they compared
Researchers asked about eating, medicines and health during Ramadan and compared reported problems with other months. No one was randomly assigned to a diet.
What this means
Ramadan changes food, drink, sleep and medication timing together. These findings in diabetes cannot be applied directly to healthy adults using a water-permitted eating window.
Population & sample
12,914 recruited; 12,243 analyzed across 13 countries: 1,070 type 1 and 11,173 type 2 diabetes.
Arms & controls
Observed Ramadan versus other-month histories; zero assigned intervention groups.
Protocol & duration
Retrospective EPIDIAR survey; daytime religious fasting includes fluid abstention and changes in meals, sleep and medication.
Reported results
Severe hypoglycemia episodes/month were reported as 0.14 versus 0.03 in type 1, and 0.03 versus 0.004 in type 2 diabetes.
What limits the inference
Recall and confounding limit causal estimates. Religious daytime fasting is not equivalent to water-permitted multiday fasting or healthy-adult TRE.
Source check & disclosures
Original indexed abstract and recruitment/analyzed distinction; full individual-level data not reviewed.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [65] DOI: 10.2337/diacare.27.10.2306

Link to this study record
75Bahr et al. · 2025Human randomized trialThe primary MRI outcome did not establish an advantage for fasting or ketogenic eating.Study design, results & limitations

Fasting, ketogenic, and anti-inflammatory diets in multiple sclerosis: a randomized controlled trial with 18-month follow-up

What happened
Neither showed a clear advantage for the main measure: new brain lesions. Some other measurements were encouraging.
Who took part
105 people with multiple sclerosis, assigned to three diets; 81 completed 18 months.
What they compared
A fasting program and a ketogenic diet were each compared with a standard healthy diet.
What this means
Secondary benefits deserve further study, but should not replace the trial’s main result or be called an MS cure.
Population & sample
105 people with relapsing-remitting multiple sclerosis: 34 standard healthy diet, 35 fasting, 36 ketogenic diet. 81 completed.
Arms & controls
3 randomized arms; an active healthy-diet comparator. Existing disease-modifying treatment was permitted. MRI and disability assessors were masked.
Protocol & duration
18 months. Fasting arm combined seven-day fasts every six months with intermittent fasting six days weekly; ketogenic arm used 20–40 g carbohydrate/day.
Reported results
No between-group difference in new T2 brain lesions at 18 months. Some exploratory cognition, mood and biomarker outcomes favored particular diets.
What limits the inference
Attrition, background treatment, low new-lesion activity and multiple exploratory outcomes limit conclusions. No proof of disease reversal.
Source check & disclosures
Original full text, Table 1 and primary-outcome analysis checked. NCT03508414, registered retrospectively.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [75] DOI: 10.1186/s40795-025-01156-5

Link to this study record
76James et al. · 2024Human randomized pilotSleep improved relative to control; a specific cognitive advantage was not established.Study design, results & limitations

Health impacts of a remotely delivered prolonged nightly fasting intervention in stressed adults with memory decline and obesity: A nationwide randomized controlled pilot trial

What happened
Sleep improved more with fasting. The study did not establish an extra benefit for memory or stress.
Who took part
58 adults who reported memory concerns and stress and were living with obesity.
What they compared
28 tried longer overnight fasting; 30 received health education with similar staff contact.
What this means
A sleep benefit is worth investigating, but is not evidence that fasting prevents dementia.
Population & sample
58 adults with obesity, stress and perceived memory decline: 28 fasting, 30 health education. 49 completed, 22 and 27 respectively.
Arms & controls
2 randomized groups with comparable weekly staff contact; health education was the attention control.
Protocol & duration
Eight weeks: a 14-hour overnight fast on six nights weekly, with no calories after 20:00.
Reported results
Sleep-quality score improved relative to control by 2.52 points (95% CI 0.74–4.30). Stress and everyday cognition improved over time without a significant group difference.
What limits the inference
Pilot without a prospective power calculation; subjective outcomes, selective population and nine withdrawals. No dementia-prevention endpoint.
Source check & disclosures
Original full text and participant counts checked. The reported retention percentage is inconsistent with 49/58; raw counts are preserved here.
Cohort relationships
No linked companion report identified in this selected library; this is not an exhaustive cohort-overlap audit.

Read original publication [76] DOI: 10.1017/cts.2024.651

Link to this study record

THE COMPLETE ASSESSMENT PASS

Look beyond the featured studies

All 928 previously retained candidates now have a documented outcome: 840 evidence assessments, 36 exclusions after source checks, and 52 records with inaccessible results. An assessment can be based on an abstract; it does not mean every full paper was verified.

Each record preserves the study design, sample and controls when reported, the finding, its limits and the source actually checked. Missing numbers stay marked as missing. Unavailable papers are not counted as evidence for or against fasting.

Download all 928 assessments and original citations · JSON

CRITICAL APPRAISAL

What could distort the result?

These outcome-specific notes make the reasoning inspectable. A randomized label alone does not remove bias. Remaining allocation, protocol-history and missing-data questions are shown explicitly; full validated signaling forms and independent review remain pending.

Bahrami et al. · Pathological response and chemotherapy toxicity
Randomization / allocation
Block randomization stratified by receptor status and stage; 22 participants per arm. Concealment and independent assessor masking are not fully established here.
Deviations from assignment
The diet arm also received omega-3. The comparison therefore estimates a treatment package, not fasting alone.
Missing outcomes
Outcome tables use 22 per arm; full missing-data and adherence history still needs independent verification.
Outcome measurement
Pathological and imaging response differ from survival. Toxicity grading and multiple endpoints require careful interpretation.
Selective reporting
Table 5 reports P=.03 for response comparisons, while the prose reports .01. Raw event counts are retained; no silent reconciliation.

Interpretation: Encouraging small-trial signal, with cointervention and reporting concerns. Not confirmatory evidence of fasting-specific cancer efficacy. [86]

Open study record
van den Burg et al. · Medication use and HbA1c at 12 months
Randomization / allocation
Randomized 51 versus 49; outcome assessors masked.
Deviations from assignment
Usual-care comparator does not match diet-product provision or intervention intensity. Medication changes were part of the clinical program.
Missing outcomes
Eight intervention and ten control participants lost to follow-up. Mixed-model intention-to-treat analysis depends on missing-data assumptions.
Outcome measurement
HbA1c is objective; medication-effect scores and combined management categories require clinical interpretation.
Selective reporting
Two primary outcomes and multiple secondary measures; historical registry changes not fully reconciled. Commercial co-funding disclosed.

Interpretation: Useful evidence for this supervised program in diet/metformin-treated diabetes, with limits on fasting-specific attribution and broader applicability. [93]

Open study record
Liu et al. · 12-month weight change
Randomization / allocation
Randomized active-control design; concealment details require second extraction.
Deviations from assignment
Both groups prescribed calories; open-label behavior and achieved intake can differ.
Missing outcomes
118/139 completed; missing-data assumptions remain relevant.
Outcome measurement
Measured weight is less vulnerable to expectation bias than subjective symptoms.
Selective reporting
Primary weight comparison reported with confidence interval; full registry history not independently reconciled.

Interpretation: Some unresolved concerns; useful evidence against a large extra benefit in this protocol. [2]

Open study record
Catenacci et al. · 12-month weight change
Randomization / allocation
165 randomized, 84 versus 81; source access here is primarily abstract/protocol.
Deviations from assignment
Intensive behavioral program in both arms; achieved restriction and attendance can contribute.
Missing outcomes
125/165 completed; intention-to-treat analysis reported.
Outcome measurement
Objective weight endpoint; assessor procedures not fully extracted.
Selective reporting
Primary result and 95% CI reported; all protocol amendments not audited.

Interpretation: Incomplete formal assessment; modest benefit with an interval near no difference. [1]

Open study record
Lowe et al. · 12-week weight and subgroup lean mass
Randomization / allocation
141 randomized; analysis subset should not be called the randomized sample.
Deviations from assignment
Open-label eating patterns; self-reported adherence cannot establish equal calories.
Missing outcomes
25 supplied no weight data; substantial concern about missing outcomes.
Outcome measurement
Weight and DXA are objective; lean mass is not a direct muscle-protein measure.
Selective reporting
Lean-mass findings are from a small subgroup; corrected Figure 2C is noted.

Interpretation: Missingness and subgroup interpretation warrant caution; no primary weight advantage. [37]

Open study record
Maruthur et al. · 12-week weight under isocaloric feeding
Randomization / allocation
Randomized, small single-site trial; baseline weight imbalance reported.
Deviations from assignment
Provided isocaloric meals help isolate timing; full adherence details need extraction.
Missing outcomes
All 41 randomized participants completed.
Outcome measurement
Objective weight; glucose endpoints secondary.
Selective reporting
Primary weight outcome reported; historical protocol reconciliation incomplete.

Interpretation: Useful mechanistic human comparison; imprecision and incomplete full-text appraisal remain. [42]

Open study record
Carter et al. · 12-month HbA1c equivalence
Randomization / allocation
Two randomized active dietary arms; unequal group sizes are not evidence of failed randomization.
Deviations from assignment
Medication adjustment was part of supervised care and can affect glycemic outcomes.
Missing outcomes
97/137 completed; attrition matters especially in equivalence analyses.
Outcome measurement
Laboratory HbA1c is objective; equivalence margin concerns clinical interpretation.
Selective reporting
Prespecified 90% interval/margin used; weight equivalence should not be inferred.

Interpretation: Some concerns, especially missingness; follow-up no longer demonstrated equivalence. [44]

Open study record
de Groot et al. · DIRECT · Chemotherapy toxicity and pathological complete response
Randomization / allocation
131 randomized; 129 analyzed. Small phase II multicenter trial.
Deviations from assignment
Only about one fifth adhered through every cycle; dexamethasone differed by arm.
Missing outcomes
Exclusions and poor adherence make per-protocol estimates vulnerable to selection.
Outcome measurement
Pathological response more objective than symptoms; assessor blinding not fully reconciled.
Selective reporting
Primary toxicity null; favorable imaging/per-protocol findings are secondary, not a positive primary trial.

Interpretation: Some concerns for assignment effects; high selection concern for adherence-based claims. [20]

Open study record
Voss et al. · ERGO2 · Six-month progression-free survival
Randomization / allocation
50 randomized; 25 per arm; small recurrent-brain-tumor population.
Deviations from assignment
Four withdrew before diet; 20 completed the dietary protocol and 17 achieved ketosis.
Missing outcomes
Attrition and small numbers constrain estimates; all survival censoring not independently audited.
Outcome measurement
Radiological progression is clinically relevant but requires consistent assessment.
Selective reporting
Primary endpoint was null; ketosis or responder subgroups should not replace it.

Interpretation: Imprecise trial with adherence concerns; no demonstrated primary benefit. [21]

Open study record
Jameson et al. · Progression-free and overall survival
Randomization / allocation
36 randomized, 32 evaluable; very small groups.
Deviations from assignment
Open-label ketogenic support alongside identical chemotherapy backbone; nutritional adherence varied.
Missing outcomes
Four randomized participants excluded from evaluable population; potential selection concern.
Outcome measurement
Survival relevant; radiological progression can depend on assessment schedules.
Selective reporting
One-sided alpha .20 screening design; 95% hazard-ratio intervals include 1.

Interpretation: Exploratory evidence with substantial imprecision; not a confirmatory efficacy conclusion. [24]

Open study record
Harvie et al. · Progression-free survival
Randomization / allocation
68 randomized, 35 versus 33; small advanced-breast-cancer sample.
Deviations from assignment
Dietary package and professional support alongside exercise; cannot isolate fasting biology.
Missing outcomes
Progression/censoring and all sensitivity analyses need independent full-text audit.
Outcome measurement
Disease progression relevant; open-label behavior remains a limitation.
Selective reporting
One-sided .20 screening threshold; HR 95% CI crosses 1.

Interpretation: Promising screening result, low certainty pending larger trials and independent appraisal. [23]

Open study record
Bensalem et al. · Blood-cell autophagic flux at six months
Randomization / allocation
Substudy of a randomized parent trial; 121 selected, 114 with baseline samples.
Deviations from assignment
Intermittent early eating combines timing and energy restriction; not a single continuous fast.
Missing outcomes
Substudy availability and time-specific samples can introduce selection.
Outcome measurement
Flux assay is closer to the process than static markers, but blood is not whole-body tissue.
Selective reporting
Six-month difference was post hoc; two-month comparison and within-arm change did not establish an increase.

Interpretation: Exploratory mechanistic evidence; insufficient for universal timing or clinical-benefit claims. [27]

Open study record
Wilkinson et al. · 12-week weight
Randomization / allocation
61 randomized; original sequence/concealment details not independently re-extracted.
Deviations from assignment
Behavioral intervention with self-monitoring; equal achieved intake not established.
Missing outcomes
52 completed; missingness may influence results.
Outcome measurement
Weight objective; not the originally registered primary mechanistic endpoint.
Selective reporting
Primary outcome changed after final participant visit, reportedly before analysis; original history needs reconciliation.

Interpretation: Reporting concern lowers confidence; retain the finding with its amendment history visible. [66]

Open study record
Pietzner et al. · Blood-protein changes during seven-day fasting
Randomization / allocation
No randomization or independent control group.
Deviations from assignment
Supervised water-only exposure; behavior/time effects cannot be separated.
Missing outcomes
Only 12 volunteers; no broad clinical outcome population.
Outcome measurement
Proteomic measurements are molecular outcomes, not measured disease prevention.
Selective reporting
Thousands of proteins and genetically inferred disease associations; not a clinical benefit trial.

Interpretation: Very low certainty for causal health benefit; useful descriptive physiology. [10]

Open study record

TRANSPARENCY & REUSE

A traceable, revisable review

Version 2026.10.24.3 · Search cutoff: 24 October 2026. Independent clinical and methodological review: pending.

1,970unique records retrieved
109publication records extracted
928candidates assessed or access-limited

1,970 retrieved records: 77 are featured in the library, 827 have additional source-specific assessments, 1014 were excluded with recorded reasons, and 52 have inaccessible original results. These categories sum to the complete retrieval. The previously retained group of 928 is now documented individually: 840 evidence assessments, 36 exclusions and 52 access limitations. Most assessments use original abstracts; full-text checks are identified individually. The 32 supplementary publications bring the featured library to 109. Publication counts are not independent-trial counts; no independent duplicate review is claimed.

Inspect the ten searches and their yields

Database: Europe PMC, restricted to PubMed records (SRC:MED). Executed 2026-10-24T13:50:58.736Z. Counts overlap across searches; the unique total is deduplicated by indexed identifier.

Weight and body composition · 290 retrieved

((TITLE_ABS:"intermittent fasting" OR TITLE_ABS:"time restricted" OR TITLE_ABS:"time-restricted" OR TITLE_ABS:"alternate day fasting" OR TITLE_ABS:"fasting mimicking" OR TITLE_ABS:"fasting-mimicking") AND (TITLE_ABS:randomized OR TITLE_ABS:randomised) AND (TITLE:weight OR TITLE:obesity OR TITLE:obese OR TITLE:fat)) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Glucose and diabetes · 254 retrieved

((TITLE_ABS:"intermittent fasting" OR TITLE_ABS:"time restricted" OR TITLE_ABS:"time-restricted" OR TITLE_ABS:"alternate day fasting" OR TITLE_ABS:"fasting mimicking" OR TITLE_ABS:"fasting-mimicking") AND (TITLE_ABS:randomized OR TITLE_ABS:randomised) AND (TITLE:diabetes OR TITLE:glycemic OR TITLE:glycaemic OR TITLE:glucose OR TITLE:metabolic)) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Multiday and water-only fasting · 184 retrieved

((TITLE:fasting AND (TITLE:prolonged OR TITLE:"water-only" OR TITLE:"water only" OR TITLE:complete OR TITLE:periodic)) AND (TITLE_ABS:human OR TITLE_ABS:patients OR TITLE_ABS:participants OR TITLE_ABS:adults)) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Cancer dietary trials · 394 retrieved

((TITLE_ABS:fasting OR TITLE_ABS:ketogenic) AND (TITLE:cancer OR TITLE:glioma OR TITLE:chemotherapy OR TITLE:neoplasm) AND (TITLE_ABS:trial OR TITLE_ABS:randomized OR TITLE_ABS:randomised)) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Human autophagy measurements · 80 retrieved

((TITLE_ABS:fasting OR TITLE_ABS:"dietary restriction" OR TITLE_ABS:"protein restriction") AND TITLE_ABS:autophagy AND (TITLE:human OR TITLE:humans OR TITLE_ABS:"autophagic flux")) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Muscle, physical function and exercise · 101 retrieved

((TITLE_ABS:"intermittent fasting" OR TITLE_ABS:"time restricted" OR TITLE_ABS:"time-restricted" OR TITLE_ABS:"alternate day fasting" OR TITLE_ABS:"fasting mimicking" OR TITLE_ABS:"fasting-mimicking") AND (TITLE_ABS:randomized OR TITLE_ABS:randomised) AND (TITLE:muscle OR TITLE:resistance OR TITLE:exercise OR TITLE:physical OR TITLE:older)) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Cognition and neurological disease · 38 retrieved

((TITLE_ABS:"intermittent fasting" OR TITLE_ABS:"time restricted" OR TITLE_ABS:"time-restricted" OR TITLE_ABS:"alternate day fasting" OR TITLE_ABS:"fasting mimicking" OR TITLE_ABS:"fasting-mimicking") AND (TITLE:cognitive OR TITLE:cognition OR TITLE:memory OR TITLE:sclerosis OR TITLE:neurodegenerative) AND (TITLE_ABS:trial OR TITLE_ABS:randomized OR TITLE_ABS:randomised)) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Underrepresented populations and harms · 77 retrieved

((TITLE_ABS:"intermittent fasting" OR TITLE_ABS:"time restricted" OR TITLE_ABS:"time-restricted" OR TITLE_ABS:"alternate day fasting" OR TITLE_ABS:"fasting mimicking" OR TITLE_ABS:"fasting-mimicking") AND (TITLE:adolescents OR TITLE:reproductive OR TITLE:hormones OR TITLE:shift OR TITLE:firefighters OR TITLE:"eating disorder" OR TITLE:adverse)) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Religious fasting and clinical outcomes · 222 retrieved

((TITLE:Ramadan OR TITLE:"religious fasting") AND (TITLE_ABS:randomized OR TITLE_ABS:randomised OR TITLE_ABS:prospective) AND (TITLE_ABS:glucose OR TITLE_ABS:weight OR TITLE_ABS:adverse OR TITLE_ABS:safety)) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Glycogen and ketones in tumors · 509 retrieved

((TITLE:glycogen OR TITLE:ketone OR TITLE:hydroxybutyrate OR TITLE:acetoacetate) AND (TITLE:cancer OR TITLE:tumor OR TITLE:tumour OR TITLE:glioma OR TITLE:melanoma)) AND FIRST_PDATE:[1800-01-01 TO 2026-10-24] AND SRC:MED

Repeat this search

Download and cite the evidence

Use the original papers for scientific claims. Exports contain publication metadata, not copies of copyrighted articles. The study table preserves control groups, denominators and source-access limitations.

Study table · CSVStudy data · JSONCitations · RISCitations · BibTeXExact searches & retrievalsScreening ledgerSource & correction auditIndependent review packetAll 928 assessments & citations

Suggested page citation: Fasting Clock. Fasting - the evidence. Version 2026.10.24.3. Add the page URL and your access date. Study-record links remain stable across updates; the dated exports identify this edition.

Companion publications are not new trials

Maifeld reports a 71-person substudy of the 145-person Cramer program (NCT02099968). [7] [8]

Bensalem is an autophagy substudy of the Teong trial (NCT03689608). [46] [27]

Carter’s 12- and 24-month reports follow the same randomized cohort (ACTRN12615000383561). [44] [45]

Cheng’s preliminary human blood findings and Dorff’s later pilot concern NCT00936364; they are not independent clinical replications. [62] [60]

Ligorio’s 14-patient survival comparison is a subanalysis of NCT03340935, the broader trial reported by Vernieri; these are overlapping publications. [22] [87]

Corrections and source discrepancies

This log records issues identified during extraction. Indexed notices are also preserved in the source audit; absence from the log is not proof that a paper has no correction.

Xie et al. · Coverage correction · 24 October 2026. Added a direct counterexample to universal tumor glycogen dependence, alongside an earlier positive kidney-cancer study and the 2026 GYS1 signaling study. The subsequent assessment pass documents all 928 retained candidates, including explicit source-access limitations. [78]

Bahrami et al. · Source discrepancy and cointervention. The intervention included omega-3 and four dietary days per cycle. Table 5 response counts are 16/22 versus 9/22 and 15/22 versus 8/22, with P=.03; prose gives .01. The discrepancy is retained rather than resolved by assumption. [86]

James et al. · Source discrepancy. The paper reports 49 completers from 58 randomized, alongside an inconsistent retention percentage. The record uses the raw overall and arm-specific counts. [76]

Ligorio et al. · Comparator clarification. The current primary registry specifies FMD plus chemotherapy in both experimental arms, with metformin added in one. A randomized label does not create a usual-diet comparator. Full published arm-specific outcomes and historical amendments remain to be reconciled. [88] Original correction

Pietrocola et al. · Coverage update · 24 October 2026. Added previously omitted evidence that tumor autophagy can be required for fasting-assisted chemotherapy response in mice. The explanation now explicitly distinguishes this from universal tumor clearance and from models in which autophagy supports cancer survival. [70]

Gates et al. · New evidence · 24 October 2026. Added the September 2026 cell-extrusion preprint, explicitly labeled as not peer reviewed. Its laboratory time course is not presented as a human fasting or autophagy timeline. Song 2026 and Buono 2023 were also added; four additional original-source leads remain partly assessed. [72]

Bauersfeld et al. · Source discrepancy. Used original Figure 1: 50 randomized and 34 analyzed. The abstract describes 34 as randomized. [59]

Wilhelmi de Toledo et al. · Source discrepancy. Used Table 1 duration groups (659/530/196/37), which sum to 1,422; the conflicting methods count is not propagated. [67]

Yang et al. · Source discrepancy. Conflicting long-term HbA1c values in abstract and body are omitted; remission denominators are retained. [5]

Lowe et al. · Published correction. Figure 2C was corrected in 2020; a 2021 notice clarifies messages sent to participants. Both notices are linked in the source audit; the complete message history is not independently reconstructed. [37]

Cheng et al. · Published correction. The 2016 correction changes flow-cytometry labels in Figures 2D and 3G from MPP to HSC+MPP. Original Figure 1 denominators were inspected separately. [62] Original correction

Wilkinson et al. · Outcome amendment. The primary outcome changed after the final participant visit; the paper says this occurred before analysis. Historical registration remains a review task. [66]

Finnell et al. · Source discrepancy. Visits and unique people are distinguished. Event counts are preferred to an internally inconsistent serious-event percentage. [13]

Independent review: work defined, approval pending

No expert endorsement is claimed. The review packet supplies outcome-specific appraisals, unresolved checks and a response template requiring qualifications, conflicts and dated findings.

  • Oncology / clinical nutrition: Check cancer-trial interpretation, nutritional vulnerability, ketosis claims and treatment boundaries.
  • Autophagy / metabolism research: Verify flux terminology, tissue differences, figure denominators and the absence of validated human timing milestones.
  • Trial methodology / statistics: Complete outcome-specific bias signaling forms, registry-history comparisons and evidence-certainty judgments.
  • Dietitian / patient communication: Assess readability, medication context, eating-disorder considerations and inappropriate extrapolation to vulnerable groups.

Revision and correction policy

Revision 2026.10.24.3: 109 original-publication records. This revision documents all 928 retained candidates and adds 13 featured summaries. It incorporates conflicting cancer-metabolism findings, autophagy measurements, prolonged fasting, active-control trials, long-term follow-up and adverse events. Access-limited records and unresolved full-text details remain explicit; no inaccessible result is inferred from a title. The original retrieval log is preserved. Corrections should identify the study, the disputed statement, the original source and a proposed replacement. A maintainer must verify the source, amend the shared record and record a dated change before publication; no automatic ingestion or automatic expert approval is used.

Future update priority: obtain unavailable original reports, extend abstract-based assessments with full methods and supplements, reconcile historical registries, obtain independent review, and assess new confirmatory trials. A documented maintenance workflow accompanies this page. No recurring search automation has been scheduled.

HOW TO READ THIS REVIEW

Methods, limits & interpretation

Review question and eligibility

Among humans or experimental models, what do specified fasting or ketogenic protocols do to clinical outcomes, harms or directly measured mechanisms compared with usual intake, active diets, another fasting condition or baseline? The library prioritizes original randomized comparisons, longer follow-up, direct cancer-metabolism experiments and autophagy assays; uncontrolled prolonged-fasting cohorts are retained with lower causal weight. Protocol-only reports and reviews are not counted as original outcome studies.

Reproducible search, with explicit limits

Ten dated queries were run in Europe PMC’s PubMed record collection (SRC:MED), covering weight, glucose, prolonged fasting, cancer, autophagy, muscle, cognition, populations, religious fasting and tumor fuels. Exact strings, timestamps and retrieved identifiers are downloadable. This was supplemented by targeted PubMed, publisher, reference and primary educational-material searches. This is one structured database search, not an exhaustive multi-database systematic review. No language restriction was added to the queries; extraction relied on accessible English material.

Initial screening is not full-text review

Every record in the fixed retrieval has an initial eligibility disposition. The audit distinguishes included extractions, exclusions with reasons, and potentially relevant candidates still lacking detailed extraction. Decisions identify whether titles, abstracts or full text were checked; no candidate is automatically treated as verified evidence. Supplementary records found outside the ten queries are counted separately. The review protocol was documented during this work, not prospectively registered. No duplicate independent screening or exhaustive reference-chain screening has been completed.

Extraction rules

Each record preserves population, species, randomized/enrolled versus analyzed counts, arms and comparator type, duration, outcomes, uncertainty, limitations and source access. For animal studies, counts refer to named experiments or figures, not an invented combined animal total. Missing denominators are labeled as gaps. Null primary outcomes remain visible beside positive secondary findings.

Preprints and the focused cancer update

The focused recheck added four reports; subsequent screening added 23 more original-publication summaries. Seven separately identified reports still have incomplete extraction. A preprint is research shared before completed journal peer review; its presence in PubMed or PMC does not change that status. Preprints are labeled on timeline cards, selected summaries, records and citation exports. Additional partly assessed sources are identified separately and do not increase the extracted-study count.

Critical appraisal and certainty

Outcome-specific appraisal notes examine allocation, deviations, missingness, measurement and reporting. Confidence profiles consider design, consistency, directness and precision. These are provisional editorial judgments, not completed validated RoB 2 forms, independent GRADE ratings or a meta-analysis. Unresolved signaling questions and registry history are explicit expert-review tasks. Observational studies are assessed for selection, confounding and reverse causation.

Source identity, corrections and conflicts

An indexed bibliographic audit resolves all library records and identifies linked correction notices. Original figures and corrections were inspected where accessible. The audit is not a guarantee that every supplement, notice or historical registration has been reconciled. Commercial clinic, diet-product and assay interests are recorded when verified; silence is not a declaration of no conflicts.

Independent review and maintenance

External clinical and statistical review is pending. The downloadable packet defines specialties, questions, unresolved checks and a reviewer response template with expertise and conflict declarations. The maintenance runbook specifies repeat searches, screening, correction checks and dated revisions. No automatic update schedule or completed expert approval is implied.

A few terms worth knowing

Randomized / crossover
Random assignment helps balance groups. In a crossover, the same people receive both conditions in a specified order, usually with a washout.
Primary / exploratory outcome
The primary outcome is the main planned test. Secondary, subgroup and post hoc findings generally need more cautious interpretation.
CI / HR / OR
A confidence interval (CI) shows uncertainty. Hazard ratios (HR) compare event rates over time; odds ratios (OR) compare odds, not absolute probabilities. A ratio of 1 represents no difference.
Progression-free survival / overall survival
Time without progression or death is different from time until death from any cause. Neither is interchangeable with a laboratory marker.
Autophagic flux / PBMC
Flux describes material moving through cellular recycling. PBMCs are a subset of blood immune cells, not every tissue in the body.

ORIGINAL SOURCES

Follow the evidence

Numbered links throughout the review open the original publication or its indexed record. Study details are above; educational material and clinical guidance are identified separately below.

  1. The Effect of 4:3 Intermittent Fasting on Weight Loss at 12 Months: A Randomized Clinical TrialCatenacci et al. · 2025Study design & limitations
  2. Calorie Restriction with or without Time-Restricted Eating in Weight LossLiu et al. · 2022Study design & limitations
  3. Effect of Alternate-Day Fasting on Weight Loss, Weight Maintenance, and Cardioprotection Among Metabolically Healthy Obese Adults: A Randomized Clinical TrialTrepanowski et al. · 2017Study design & limitations
  4. Early Time-Restricted Feeding Improves Insulin Sensitivity, Blood Pressure, and Oxidative Stress Even Without Weight Loss in Men with PrediabetesSutton et al. · 2018Study design & limitations
  5. Effect of an Intermittent Calorie-restricted Diet on Type 2 Diabetes Remission: A Randomized Controlled TrialYang et al. · 2023Study design & limitations
  6. Fasting-mimicking diet and markers/risk factors for aging, diabetes, cancer, and cardiovascular diseaseWei et al. · 2017Study design & limitations
  7. Effects of Fasting and Lifestyle Modification in Patients with Metabolic Syndrome: A Randomized Controlled TrialCramer et al. · 2022Study design & limitations
  8. Fasting alters the gut microbiome reducing blood pressure and body weight in metabolic syndrome patientsMaifeld et al. · 2021Study design & limitations
  9. Controlled trial of fasting and one-year vegetarian diet in rheumatoid arthritisKjeldsen-Kragh et al. · 1991Study design & limitations
  10. Systemic proteome adaptions to 7-day complete caloric restriction in humansPietzner et al. · 2024Study design & limitations
  11. The Effects of Prolonged Water-Only Fasting and Refeeding on Markers of Cardiometabolic RiskScharf et al. · 2022Study design & limitations
  12. Prolonged fasting promotes systemic inflammation and platelet activation in humans: A medically supervised, water-only fasting and refeeding studyCommissati et al. · 2025Study design & limitations
  13. Is fasting safe? A chart review of adverse events during medically supervised, water-only fastingFinnell et al. · 2018Study design & limitations
  14. Glycogen drives tumour initiation and progression in lung adenocarcinomaClarke et al. · 2025Study design & limitations
  15. Liver and kidney metabolism during prolonged starvationOwen et al. · 1969Study design & limitations
  16. Lactate Metabolism in Human Lung TumorsFaubert et al. · 2017Study design & limitations
  17. A ketogenic diet increases transport and oxidation of ketone bodies in RG2 and 9L gliomas without affecting tumor growthDe Feyter et al. · 2016Study design & limitations
  18. β-Hydroxybutyrate suppresses colorectal cancerDmitrieva-Posocco et al. · 2022Study design & limitations
  19. Prevention of Dietary-Fat-Fueled Ketogenesis Attenuates BRAF V600E Tumor GrowthXia et al. · 2017Study design & limitations
  20. Fasting mimicking diet as an adjunct to neoadjuvant chemotherapy for breast cancer in the multicentre randomized phase 2 DIRECT trialde Groot et al. · DIRECT · 2020Study design & limitations
  21. ERGO2: A Prospective, Randomized Trial of Calorie-Restricted Ketogenic Diet and Fasting in Addition to Reirradiation for Malignant GliomaVoss et al. · ERGO2 · 2020Study design & limitations
  22. Fasting-Mimicking Diet Is Safe and Reshapes Metabolism and Antitumor Immunity in Patients with CancerVernieri et al. · 2022Study design & limitations
  23. Randomised phase-2 screening trial of intermittent energy restriction plus resistance exercise versus resistance exercise alone during chemotherapy for advanced breast cancerHarvie et al. · 2025Study design & limitations
  24. A randomized phase II trial of gemcitabine, nab-paclitaxel, cisplatin with or without a medically supervised ketogenic diet for patients with metastatic pancreatic cancerJameson et al. · 2026Study design & limitations
  25. Fasting boosts breast cancer therapy efficacy via glucocorticoid activationPadrão et al. · 2026Study design & limitations
  26. Early Time-Restricted Feeding Improves 24-Hour Glucose Levels and Affects Markers of the Circadian Clock, Aging, and Autophagy in HumansJamshed et al. · 2019Study design & limitations
  27. Intermittent time-restricted eating may increase autophagic flux in humans: an exploratory analysisBensalem et al. · 2025Study design & limitations
  28. Effect of fasting-mimicking diet on markers of autophagy and metabolic health in human subjectsEspinoza et al. · 2025 onlineStudy design & limitations
  29. Fasting Increases Human Skeletal Muscle Net Phenylalanine Release and This Is Associated with Decreased mTOR SignalingVendelbo et al. · 2014Study design & limitations
  30. In vivo analysis of autophagy in response to nutrient starvation using transgenic mice expressing a fluorescent autophagosome markerMizushima et al. · 2004Study design & limitations
  31. Autophagy-deficient mice develop multiple liver tumorsTakamura et al. · 2011Study design & limitations
  32. Autophagy suppresses progression of K-ras-induced lung tumors to oncocytomas and maintains lipid homeostasisGuo et al. · 2013Study design & limitations
  33. Short-term post-fast refeeding enhances intestinal stemness via polyaminesImada et al. · 2024Study design & limitations
  34. Dietary restriction impacts health and lifespan of genetically diverse miceDi Francesco et al. · 2024Study design & limitations
  35. Reduced dietary protein intake does not alter autophagy in human blood: A randomized crossover study in healthy adultsSingh et al. · 2026Study design & limitations
  36. Time-Restricted Eating Without Calorie Counting for Weight Loss in a Racially Diverse Population : A Randomized Controlled TrialLin et al. · 2023Study design & limitations
  37. Effects of Time-Restricted Eating on Weight Loss and Other Metabolic Parameters in Women and Men With Overweight and Obesity: The TREAT Randomized Clinical TrialLowe et al. · 2020Study design & limitations
  38. Effects of intermittent and continuous calorie restriction on body weight and metabolism over 50 wk: a randomized controlled trialSchübel et al. · 2018Study design & limitations
  39. Effectiveness of Early Time-Restricted Eating for Weight Loss, Fat Loss, and Cardiometabolic Health in Adults With Obesity: A Randomized Clinical TrialJamshed et al. · 2022Study design & limitations
  40. Effects of early, late and self-selected time-restricted eating on visceral adipose tissue and cardiometabolic health in participants with overweight or obesity: a randomized controlled trialDote-Montero et al. · 2025Study design & limitations
  41. A randomised controlled trial of the 5:2 dietHajek et al. · 2021Study design & limitations
  42. Effect of Isocaloric, Time-Restricted Eating on Body Weight in Adults With Obesity : A Randomized Controlled TrialMaruthur et al. · 2024Study design & limitations
  43. Effect of Time-Restricted Eating on Weight Loss in Adults With Type 2 Diabetes: A Randomized Clinical TrialPavlou et al. · 2023Study design & limitations
  44. Effect of Intermittent Compared With Continuous Energy Restricted Diet on Glycemic Control in Patients With Type 2 Diabetes: A Randomized Noninferiority TrialCarter et al. · 2018Study design & limitations
  45. The effect of intermittent compared with continuous energy restriction on glycaemic control in patients with type 2 diabetes: 24-month follow-up of a randomised noninferiority trialCarter et al. · 2019Study design & limitations
  46. Intermittent fasting plus early time-restricted eating versus calorie restriction and standard care in adults at risk of type 2 diabetes: a randomized controlled trialTeong et al. · 2023Study design & limitations
  47. Time-Restricted Eating in Adults With Metabolic Syndrome : A Randomized Controlled TrialManoogian et al. · 2024Study design & limitations
  48. A 5:2 Intermittent Fasting Meal Replacement Diet and Glycemic Control for Adults With Diabetes: The EARLY Randomized Clinical TrialGuo et al. · 2024Study design & limitations
  49. Time-restricted eating versus dietetic guidance on glycaemic outcomes in adults at risk of type 2 diabetes: a non-inferiority randomised clinical trialParr et al. · 2026Study design & limitations
  50. Effects of Time-Restricted Eating on Nonalcoholic Fatty Liver Disease: The TREATY-FLD Randomized Clinical TrialWei et al. · 2023Study design & limitations
  51. Effects of eight weeks of time-restricted feeding (16/8) on basal metabolism, maximal strength, body composition, inflammation, and cardiovascular risk factors in resistance-trained malesMoro et al. · 2016Study design & limitations
  52. Time-restricted eating and exercise training improve HbA1c and body composition in women with overweight/obesity: A randomized controlled trialHaganes et al. · 2022Study design & limitations
  53. Feasibility of time-restricted eating and impacts on cardiometabolic health in 24-h shift workers: The Healthy Heroes randomized control trialManoogian et al. · 2022Study design & limitations
  54. Intermittent Energy Restriction for Adolescents With Obesity: The Fast Track to Health Randomized Clinical TrialLister et al. · 2024Study design & limitations
  55. Caloric restriction and time-restricted eating in older adults with overweight or obesity: The Health, Aging, and Later-Life Outcomes Pilot StudyHouston et al. · 2026Study design & limitations
  56. Effects of 15:9 time-restricted eating on cognitive function and anthropometric parameters among the elderly with mild cognitive impairment: an assessor-blinded, pilot randomized controlled trialWang et al. · 2025Study design & limitations
  57. Impact of a 4-week time-restricted eating (16:8) intervention on working memory and attentional control in adult women: A randomized controlled pilotMcLeod et al. · 2026Study design & limitations
  58. Association of eating duration less than 8 h with all-cause, cardiovascular, and cancer mortalityChen et al. · 2025Study design & limitations
  59. The effects of short-term fasting on quality of life and tolerance to chemotherapy in patients with breast and ovarian cancer: a randomized cross-over pilot studyBauersfeld et al. · 2018Study design & limitations
  60. Safety and feasibility of fasting in combination with platinum-based chemotherapyDorff et al. · 2016Study design & limitations
  61. The effects of short-term fasting on tolerance to (neo) adjuvant chemotherapy in HER2-negative breast cancer patients: a randomized pilot studyde Groot et al. · 2015Study design & limitations
  62. Prolonged fasting reduces IGF-1/PKA to promote hematopoietic-stem-cell-based regeneration and reverse immunosuppressionCheng et al. · 2014Study design & limitations
  63. A Diet Mimicking Fasting Promotes Regeneration and Reduces Autoimmunity and Multiple Sclerosis SymptomsChoi et al. · 2016Study design & limitations
  64. Medically supervised water-only fasting in the treatment of hypertensionGoldhamer et al. · 2001Study design & limitations
  65. A population-based study of diabetes and its characteristics during the fasting month of Ramadan in 13 countries: results of the epidemiology of diabetes and Ramadan 1422/2001 (EPIDIAR) studySalti et al. · 2004Study design & limitations
  66. Time-Restricted Eating Promotes Weight Loss and Favorable Changes in Adipose in Obesity: The TREAD Randomized Control TrialWilkinson et al. · 2026Study design & limitations
  67. Safety, health improvement and well-being during a 4 to 21-day fasting period in an observational study including 1422 subjectsWilhelmi de Toledo et al. · 2019Study design & limitations
  68. Prolonged Water-Only Fasting Followed by a Whole-Plant-Food Diet Promotes Fat-Free Mass Recovery and Continued Fat Mass Loss in Adults With Overweight or ObesityGabriel et al. · 2025Study design & limitations
  69. Intermittent fasting activates markers of autophagy in mouse liver, but not muscle from mouse or humansChaudhary et al. · 2022Study design & limitations
  70. Caloric Restriction Mimetics Enhance Anticancer ImmunosurveillancePietrocola et al. · 2016Study design & limitations
  71. Fasting-Mimicking Diet Inhibits Autophagy and Synergizes with Chemotherapy to Promote T-Cell-Dependent Leukemia-Free SurvivalBuono et al. · 2023Study design & limitations
  72. Starvation improves epithelial fitness by selectively extruding DNA damaged cellsGates et al. · 2026Study design & limitations
  73. Intermittent Fasting Reduces Taurine Availability and Promotes Antitumor Immunity Through PERK Attenuation-Associated Paraptosis-Like Cell Death in Colorectal CancerSong et al. · 2026Study design & limitations
  74. Spermidine is essential for fasting-mediated autophagy and longevityHofer et al. · 2024Study design & limitations
  75. Fasting, ketogenic, and anti-inflammatory diets in multiple sclerosis: a randomized controlled trial with 18-month follow-upBahr et al. · 2025Study design & limitations
  76. Health impacts of a remotely delivered prolonged nightly fasting intervention in stressed adults with memory decline and obesity: A nationwide randomized controlled pilot trialJames et al. · 2024Study design & limitations
  77. Wernicke's Encephalopathy Following 40 Days of Water-Only Fasting: A Case ReportAweer et al. · 2025Study design & limitations
  78. Glycogen metabolism is dispensable for tumour progression in clear cell renal cell carcinomaXie et al. · 2021Study design & limitations
  79. GYS1 induces glycogen accumulation and promotes tumor progression via the NF-κB pathway in Clear Cell Renal CarcinomaChen et al. · 2020Study design & limitations
  80. Glycogen synthase 1 promotes breast cancer progression by promoting IκBα ubiquitination and degradation independent of its canonical enzyme functionZhuo et al. · 2026Study design & limitations
  81. An alternative route for β-hydroxybutyrate metabolism supports cytosolic acetyl-CoA synthesis in cancer cellsKaluba et al. · 2025Study design & limitations
  82. β-Hydroxybutyrate promotes cancer metastasis through β-hydroxybutyrylation-dependent stabilization of SnailJiang et al. · 2025Study design & limitations
  83. β-Hydroxybutyrate promotes chemoresistance and proliferation in breast cancer cellsMashinchian et al. · 2025Study design & limitations
  84. Conditional dependency of oncogenic KRAS in driving ketone body catabolism and pancreatic cancer growthWang et al. · 2026Study design & limitations
  85. Ketogenic Diet and β-Hydroxybutyrate Suppress Bladder Cancer Growth and Enhance Anti-PD-L1 Immunotherapy EfficacyZeng et al. · 2026Study design & limitations
  86. Fasting mimicking diet during neo-adjuvant chemotherapy in breast cancer patients: a randomized controlled trial studyBahrami et al. · 2024Study design & limitations
  87. Adding fasting-mimicking diet to first-line carboplatin-based chemotherapy is associated with better overall survival in advanced triple-negative breast cancer patients: A subanalysis of the NCT03340935 trialLigorio et al. · 2024Study design & limitations
  88. Early downmodulation of tumor glycolysis predicts response to fasting-mimicking diet in triple-negative breast cancer patientsLigorio et al. · 2025Study design & limitations
  89. Prolonged water-only fasting in the management of low-grade follicular lymphoma: a case seriesGabriel et al. · 2024Study design & limitations
  90. A feasibility randomised controlled trial of water-only fasting prior to CAPOX chemotherapy for stage 2/3 colorectal cancerHerbert et al. · 2026Study design & limitations
  91. Time-Restricted Eating and Metformin in Invasive Breast Cancer or DCIS: A Randomized, Phase IIb, Presurgical Trial. Preliminary Safety AnalysisBriata et al. · 2026Study design & limitations
  92. Intermittent fasting versus continuous calorie energy reduction in prostate cancer patients on active surveillanceYeary et al. · 2026Study design & limitations
  93. Integration of a fasting-mimicking diet programme in primary care for type 2 diabetes reduces the need for medication and improves glycaemic control: a 12-month randomised controlled trialvan den Burg et al. · 2024Study design & limitations
  94. Effects of 3 months of 10-h per-day time-restricted eating and 3 months of follow-up on bodyweight and cardiometabolic health in Danish individuals at high risk of type 2 diabetes: the RESET single-centre, parallel, superiority, open-label, randomised controlled trialQuist et al. · 2024Study design & limitations
  95. Effects of Intermittent Fasting on Liver Steatosis and Fibrosis, Serum FGF-21 and Autophagy Markers in Metabolic Dysfunction-Associated Fatty Liver Disease: A Randomized Controlled TrialOzlu Karahan et al. · 2025Study design & limitations
  96. Ketone bodies and cancer risk in the general population: The prevention of renal and vascular end-stage disease (PREVEND) studyLuo et al. · 2025Study design & limitations
  97. Metabolic effects of fasting on human and mouse blood in vivo.Pietrocola et al. · 2017Study design & limitations
  98. Ketogenic diet and chemotherapy combine to disrupt pancreatic cancer metabolism and growth.Yang et al. · 2022Study design & limitations
  99. Fasting-mimicking diet synergizes with ferroptosis against quiescent, chemotherapy-resistant cells.Liu et al. · 2023Study design & limitations
  100. OXCT1 regulates NF-κB signaling pathway through β-hydroxybutyrate-mediated ketone body homeostasis in lung cancer.Lu et al. · 2023Study design & limitations
  101. Sustained weight reduction following 12 weeks of intermittent fasting intervention in people with insulin-treated type 2 diabetes-Two-year follow-up of the randomised controlled InterFast-2 trial.Azhar et al. · 2025Study design & limitations
  102. Efficacy and safety of time-restricted eating in metabolic dysfunction-associated steatotic liver disease.Oh et al. · 2025Study design & limitations
  103. Intermittent fasting for weight loss in night shift workers: a three-arm, superiority randomised clinical trial.Bonham et al. · 2025Study design & limitations
  104. Ketone Bodies Inhibit Growth of Hepatoblastoma and Rhabdoid Tumor of the Kidney Cells.Muto-Fujita et al. · 2025Study design & limitations
  105. The feasibility and safety of fasting-mimicking diet in breast cancer patients with chemotherapy in China.Xue et al. · 2025Study design & limitations
  106. Survival outcomes of rectal and head and neck cancer patients receiving radio(chemo)therapy with a ketogenic diet. A post-hoc analysis from the KETOCOMP trial.Klement et al. · 2026Study design & limitations
  107. Preliminary study on ketone body metabolism in anaplastic thyroid cancer.Wang et al. · 2026Study design & limitations
  108. Very low-carbohydrate ketogenic diet in treatment-naïve women with endometrial cancer and overweight: a randomized feasibility study.Dantas et al. · 2026Study design & limitations
  109. Machine learning identifies microbiome and clinical predictors of sustained weight loss following prolonged fasting.Kaufhold et al. · 2026Study design & limitations
  110. β-hydroxybutyrate enhances the metabolic fitness of CAR T cells in cancerLiu et al. · 2026

    Original abstract checked. BHB improved engineered immune-cell function and tumor control in preclinical models; a healthy-volunteer component measured immune-cell energetics, not patient cancer response. Sample sizes and full controls remain unverified here.

  111. PAK5-Mediated Suppression of β-Hydroxybutyrate Production Promotes Breast Cancer Metastasis and Can Be Overcome with Ketogenic DietXing et al. · 2026

    Original abstract checked. BHB or ketogenic diet reduced progression in breast-cancer models involving PAK5/HMGCS2 signaling; human tissue associations were observational. Group sizes and complete experimental controls remain unverified here.

  112. Induction of a metabolic switch from glucose to ketone metabolism programs ketogenic diet-induced therapeutic vulnerability in lung cancerWu et al. · 2025

    Original abstract checked. Ketones supported lung tumor-initiating cells under glucose limitation while creating sensitivity to particular metabolic inhibitors. This is not evidence for diet alone as treatment. Full model denominators and controls remain unverified here.

  113. PROFASTA: protocol for fasting, exercise and human autophagic-flux measurementMasedunskas et al. · 2024

    Original study protocol, not completed results. A planned three-day fast and proposed blood-cell measurements do not establish an observed autophagy threshold.

  114. NCT04248998 - registered BREAKFAST study designBREAKFAST investigators / ClinicalTrials.gov · checked 2026

    Primary registry specifies chemotherapy plus FMD in both arms, with metformin added in one. Current registry checked; complete historical amendments and published arm-specific outcomes are not reconciled.

  115. Systemic hypoxia drives glycogen-fueled progression of lung adenocarcinomaClarke et al. · 2026 · preprint

    Original preprint, posted September 3, 2026; not peer reviewed. Targeted recheck examined indexed primary-text excerpts. Full methods, all denominators and risk-of-bias extraction remain pending; not counted as an extracted library record.

  116. 16-h fasting optimizes cancer immunotherapy in mice and humansChen et al. · 2026

    Original study, published online February 19, 2026. Indexed abstract checked; publisher full text could not be accessed. Human sample size, allocation, controls and clinical effect sizes remain unverified here; not counted as a fully extracted library record.

  117. Intermittent fasting alters tumor burden, autophagy, and metabolites in chronic lymphocytic leukemiaStringer et al. · 2024 · preprint

    Original December 16, 2024 preprint; not peer reviewed in the version checked. Primary abstract examined. Complete methods, comparator and assay extraction remain pending; not counted as an extracted library record. Do not count its conference abstract as another trial.

  118. Fasting enhances the efficacy of Sorafenib in breast cancer via mitophagy mediated ROS-driven p53 pathwayLi et al. · 2025

    Original preclinical study. Publisher highlights and abstract checked; complete group sizes, control details and full-text appraisal remain pending. The reported combination inhibited mitophagy; the title alone does not establish increased recycling.

  119. Processed Food Addiction - original lecture slides, pages 46–49Pradip Jamnadas, MD / Cardiovascular Interventions · 2022

    Primary educational material, visually checked. These slides express claims; they are not controlled outcome evidence.

  120. Fasting for Survival - original lecture linked by the speaker’s practiceDr. Pradip Jamnadas, MD / Cardiovascular Interventions · accessed 2026

    Educational perspective, not an experimental study. The primary practice page and lecture link were verified; a complete primary transcript was not available for line-by-line quotation.

  121. Exercise, Diet, and Weight Management During Cancer Treatment: ASCO GuidelineLigibel et al. / ASCO · 2022

    Professional guidance, not an original fasting trial. Its evidence assessment predates the 2025–2026 screening trials included above.

  122. Nutrition support for adults - recommendations 1.3–1.4NICE · CG32

    Clinical nutrition and refeeding guidance; not an endorsement of elective fasting or a home refeeding protocol.

Back to top ↑