evipedia.ai
← Back

Fasting-Mimicking Diet for Health & Longevity

Evidence Review created on 10/25/2026 using AI4L / Opus 5.5

Also known as: Fasting Mimicking Diet, FMD, Periodic Fasting-Mimicking Diet, Prolon, ProLon, Intermittent Reduced Calorie Diet, IRCD

Motivation

The fasting-mimicking diet is a five-day eating plan that is low in calories, protein and sugar and relatively rich in plant fats. It is designed to trigger many of the body’s responses to water-only fasting while still allowing small meals, appealing to people seeking possible benefits of prolonged fasting without its full hardship. It is repeated periodically, from monthly to a few times a year, with normal eating in between, and is sold mainly as a packaged meal kit called Prolon.

The approach grew out of laboratory research on how short spells without food switch cells into a protective, repair-oriented state, an idea first explored as a way to shield healthy tissue during cancer treatment. Interest widened as animal studies of repeated cycles and early human trials examined blood sugar regulation and markers linked to biological aging, and as the diet became commercially available.

This review examines the evidence for and against periodic fasting-mimicking cycles in health-focused adults: expected benefits, risks, who should avoid them, practical protocols, and how much of the research comes from parties with a financial stake in the product.

Benefits - Risks - Protocol - Conclusion

This section lists in-depth podcasts, articles and papers that give a high-level overview of the fasting-mimicking diet.

Only four items are listed because no further freely accessible overview qualified. Peter Attia’s freely available “Ask Me Anything” episode #2 covers fasting-mimicking diet protocols only in a brief segment of a broader water-fasting episode, and his longer discussion in episode #72 is member-only, so neither is listed. No directly relevant content was found from Andrew Huberman (his site covers time-restricted eating (eating within a daily window) and other intermittent fasting, not this diet). Life Extension’s only fasting-mimicking diet content located is a 2020 news brief on a mouse study combining the diet with vitamin C, which gives no high-level overview of the diet.

Grokipedia

  • Fasting-mimicking diet

    AI-generated encyclopedia overview of the five-day protocol, proposed mechanisms, clinical trials, safety cautions and commercial criticisms; useful for orientation but not a primary source.

Examine

No dedicated Examine article on the fasting-mimicking diet exists; Examine covers it only through short research-feed summaries of individual trials.

ConsumerLab

No ConsumerLab article on the fasting-mimicking diet or Prolon exists; ConsumerLab’s search returns only unrelated supplement content.

Systematic Reviews

The following systematic reviews and meta-analyses cover the diet’s cardiometabolic, diabetes, cancer and preclinical evidence.

No systematic review addresses the diet’s transient adverse effects (fatigue, headache, dizziness) or low blood sugar, so the principal risk is unrepresented; Mohammadzadeh et al. pooled fat-free mass (no significant loss), and Liu et al. covers chemotherapy-related adverse effects only.

Mechanism of Action

The FMD supplies about 4,600 kJ (≈1,100 kcal) on day 1 and about 3,000 kJ (≈720 kcal) on days 2–5, with 9–11% of energy from protein, little sugar and mostly unsaturated plant fat (Brandhorst et al., 2024). Most mechanistic work comes from Valter Longo’s University of Southern California laboratory; Longo holds equity in L-Nutra, which sells the Prolon FMD, and the university licenses patents to it (a publisher correction disclosing competing interests). Proposed mechanisms:

  • Growth-signal reduction: Low protein and sugar lower IGF-1 (insulin-like growth factor 1, a liver-made growth signal) and the activity of mTOR (mechanistic target of rapamycin, a cell-growth switch) and PKA (protein kinase A, a sugar-responsive growth signal), pathways whose suppression extends animal lifespan (Brandhorst et al., 2015).
  • Fuel switch: Depleted glycogen (stored glucose) shifts metabolism toward fat burning and ketone production (fat-derived fuels such as β-hydroxybutyrate).
  • Breakdown, then rebuilding: Autophagy (cellular self-recycling) and shrinkage of some tissues and immune-cell pools during the diet precede stem-cell-driven regeneration on refeeding in mice (Brandhorst et al., 2015).
  • Differential stress resistance: Healthy cells enter a protected low-growth state, whereas cancer cells with growth-pathway mutations cannot, so short-term starvation protected mice but not injected cancer cells against high-dose chemotherapy (Raffaghello et al., 2008).

A competing view attributes most human effects to the calorie deficit and weight loss; one trial found weight loss not significantly different from daily calorie restriction (Sadeghian et al., 2021). Refeeding-driven stem-cell bursts may also promote tumors in mouse intestine (Imada et al., 2024).

Historical Context & Evolution

The FMD was originally intended as an oncology tool. Longo’s laboratory found that short-term starvation protected yeast and mice, but not cancer cells, against high-dose chemotherapy (Raffaghello et al., 2008). In a case series of ten patients who fasted around chemotherapy, the six treated both with and without fasting reported less fatigue, weakness and gastrointestinal upset while fasting (Safdie et al., 2009). Because water-only fasting is hard and risky for frail patients, the group designed a plant-based meal kit that keeps fasting signals.

Its move toward health optimization came in 2015, when bimonthly cycles extended lifespan in mice and a small human pilot reported lower aging-related risk markers (Brandhorst et al., 2015). A 100-person randomized trial in healthy adults followed (Wei et al., 2017), and L-Nutra commercialized the kit as Prolon.

Critics noted the early human evidence rested on two small, short trials run by the developer’s group (Sofi, 2020). Groups in Leiden, Stanford, Heidelberg and Ahvaz ran their own randomized trials, and a Milan group ran a single-arm cancer trial (Vernieri et al., 2022). Findings cut both ways: improved diabetes control (van den Burg et al., 2024) and activity of Crohn’s disease (a chronic inflammatory bowel disease) (Kulkarni et al., 2026) on one side; no blood-pressure benefit in medicated diabetes (van den Burg et al., 2024) and no gain in complete tumor response (de Groot et al., 2020) on the other. Many of these trials still used L-Nutra product or had Longo as co-author, so independent replication remains limited.

Expected Benefits

High 🟩 🟩 🟩

Weight and Body-Fat Loss ⚠️ Conflicted

Five-day cycles produce modest weight loss, mostly fat, including visceral (around-the-organs) fat (Schoonakker et al., 2025). Shown by Longo’s group (Wei et al., 2017), Pijl’s Leiden group (van den Burg et al., 2024), Tang and Lin in Wuhan (Tang & Lin, 2020) and a type 2 diabetes network meta-analysis (Zeng et al., 2024). An eleven-trial meta-analysis found no significant pooled weight effect, but a 6.6 kg loss in its type 2 diabetes subgroup (Mohammadzadeh et al., 2025). Net reading: loss is consistent in type 2 diabetes, unconfirmed across all populations.

Magnitude: 3.6 kg more weight loss than usual care at 12 months (95% CI, the confidence interval within which the true effect likely lies, −5.2 to −2.1 kg) (van den Burg et al., 2024).

Glycemic Control in Type 2 Diabetes

In type 2 diabetes treated with diet or metformin, monthly cycles lowered HbA1c (glycated hemoglobin, a three-month blood-sugar average) while glucose-lowering medication was reduced. Shown by van den Burg and Pijl’s Leiden group (van den Burg et al., 2024) and Tang and Lin in Wuhan (Tang & Lin, 2020), and pooled in a meta-analysis (Mohammadzadeh et al., 2025). The Leiden effect lost significance after adjustment for weight loss, and gains after stopping periodic fasting were generally not sustained (van den Burg et al., 2023).

Magnitude: HbA1c 3.2 mmol/mol (0.3 percentage points) lower than usual care at 12 months (95% CI −6.2 to −0.2 mmol/mol) despite less medication in a trial co-funded by L-Nutra (van den Burg et al., 2024); pooled across trials, 8.6 mmol/mol lower than control (95% CI −12.4 to −4.8) (Mohammadzadeh et al., 2025).

Medium 🟩 🟩

Blood Pressure ⚠️ Conflicted

In Wei’s randomized comparison, three cycles lowered systolic pressure versus a normal diet, with larger drops in people starting above 120 mmHg (Wei et al., 2017), and a meta-analysis of eleven trials found a modest pooled reduction (Mohammadzadeh et al., 2025). The 12-month Leiden trial in type 2 diabetes, where most participants had treated hypertension, found no effect (van den Burg et al., 2024). Net reading: a small reduction is likely in people with elevated pressure but is not established in medicated type 2 diabetes.

Magnitude: Pooled systolic pressure 4.1 mmHg lower than control (95% CI −7.6 to −0.7) and diastolic 2.3 mmHg lower (95% CI −4.2 to −0.4) (Mohammadzadeh et al., 2025); no effect in the Leiden trial (systolic −0.4 mmHg; 95% CI −6.0 to +5.1) (van den Burg et al., 2024).

Liver Fat Reduction

Monthly cycles reduced liver fat, measured as the MRI (magnetic resonance imaging) proton density fat fraction used to diagnose steatotic (fatty) liver disease, and an MRI marker of liver inflammation and fibrosis (scarring) in van den Burg’s Leiden trial (van den Burg et al., 2025). Longo’s group reported a similar fall in a secondary analysis of healthy adults (Brandhorst et al., 2024). Imaging change has not yet been linked to fewer liver events in FMD trials.

Magnitude: Liver fat fraction 2.8 percentage points lower than usual care at 12 months (95% CI −4.7 to −0.8) (van den Burg et al., 2025).

Inflammatory Bowel Disease Activity ⚠️ Conflicted

In mild-to-moderate Crohn’s disease, a Stanford open-label (unblinded) randomized trial of three monthly cycles improved clinical response, remission and fecal calprotectin (a stool marker of gut inflammation) (Kulkarni et al., 2026). A Miami pilot in ulcerative colitis (chronic inflammation of the colon lining), stopped early at 32 patients, missed its primary response endpoint but improved symptom scores (Damas et al., 2025). Net reading: benefit is supported in Crohn’s disease, while ulcerative colitis remains untested at adequate size.

Magnitude: Clinical response in 69.2% with the diet vs 43.8% on the usual diet after three cycles (Kulkarni et al., 2026).

Low 🟩

LDL Cholesterol and Triglycerides in At-Risk Adults ⚠️ Conflicted

A post hoc (after-the-fact) analysis of Wei’s trial found that LDL (low-density lipoprotein) cholesterol and triglycerides fell in participants with elevated baseline values (Wei et al., 2017). The randomized comparison and a meta-analysis found no lipid effect (Mohammadzadeh et al., 2025). Net reading: no reliable lipid benefit is shown.

Magnitude: LDL fell 14.9 mg/dL in diet participants with total cholesterol above 199 mg/dL (post hoc, no control group) (Wei et al., 2017); pooled trials showed no significant lipid change (Mohammadzadeh et al., 2025).

C-Reactive Protein in At-Risk Adults ⚠️ Conflicted

CRP (C-reactive protein, a blood marker of inflammation) fell in Wei’s participants who started above 1 mg/L, but the randomized comparison against a normal diet showed no difference (Wei et al., 2017). Net reading: an inflammation benefit is unconfirmed by controlled data.

Magnitude: CRP fell 1.6 mg/L in participants starting above 1 mg/L (post hoc, no control group); the randomized between-group difference was not significant (Wei et al., 2017).

Tumor Response to Cancer Therapy ⚠️ Conflicted

In rodents, FMD enhanced chemotherapy, hormone and immunotherapy (Pereira et al., 2026). In DIRECT (131 women, HER2-negative breast cancer: tumors lacking the HER2 receptor), FMD improved radiological response, but pathological complete response (no remaining invasive cancer) was unchanged (de Groot et al., 2020). Net reading: tumor shrinkage, not eradication, improved.

Magnitude: Radiological complete or partial response more likely with the diet (OR 3.17, odds ratio: how many times higher the odds of response were with the diet; 95% CI 1.06–9.45; stable or progressive disease 11.3% vs 26.9%); pathological complete response 10.8% vs 12.7% (OR 0.83; 95% CI 0.28–2.44) (de Groot et al., 2020).

Protection from Chemotherapy Toxicity ⚠️ Conflicted

Fasting protected healthy cells from chemotherapy in mice (Raffaghello et al., 2008). In DIRECT, severe toxicity did not differ despite dexamethasone (an anti-nausea steroid) being omitted in the diet arm, though T-cell DNA damage was lower (de Groot et al., 2020). Net reading: no clinical toxicity reduction is shown.

Magnitude: Grade III/IV (severe or life-threatening) toxicity in 47.7% with the diet vs 56.3% on the regular diet in the paper’s Table 2 (not significant); the paper’s text gives 75.4% vs 65.6% for the same comparison, a discrepancy it does not explain (de Groot et al., 2020).

Multiple Sclerosis Quality of Life ⚠️ Conflicted

In a 60-patient randomized multiple sclerosis pilot, one 7-day FMD followed by a Mediterranean diet (plant-rich, olive-oil-based) improved quality of life; evidence is indirect, as FMD was not tested alone (Choi et al., 2016). Related calorie-restriction pilots found no patient-reported benefit (Roman et al., 2020). Net reading: benefit is unproven.

Magnitude: Not quantified in available studies. The pilot reports its quality-of-life improvements only graphically and as “clinically meaningful,” without a between-group figure in its text.

Speculative 🟨

Lifespan and Healthspan Extension

In mice, bimonthly cycles started in middle age extended lifespan, lowered cancer incidence and slowed bone loss (Brandhorst et al., 2015). No human lifespan data exist; the basis is animal work only.

Lower Estimated Biological Age

Three cycles lowered a blood-chemistry estimate of biological age by a median 2.5 years in Longo’s analysis (Brandhorst et al., 2024). Such estimates are not validated as targets of intervention.

Lower IGF-1 Levels

Cycles lower IGF-1, a growth signal linked in observational data to some cancers, by about 19 ng/mL in pooled trials (Mohammadzadeh et al., 2025). Whether this lowers disease risk is untested.

Autophagy Induction

An L-Nutra-sponsored 30-person pilot found increased autophagic flux (the rate of cellular self-recycling) in blood immune cells after six diet days (Espinoza et al., 2026). Clinical relevance is unknown.

Stem-Cell Regeneration and Immune Rejuvenation

Refeeding after FMD expanded stem and progenitor (immature precursor) cells in mice (Brandhorst et al., 2015); in humans, the lymphoid-to-myeloid ratio (an immune-aging marker) rose (Brandhorst et al., 2024). The basis is mechanistic.

Neuroprotection and Cognition

Cycles improved cognition in old mice (Brandhorst et al., 2015) and slowed decline in Alzheimer’s disease mouse models (Rangan et al., 2022). No controlled human cognitive trials exist; the basis is animal work.

Pancreatic Beta-Cell Regeneration

In diabetic mice, cycles regenerated insulin-producing beta cells and restored glucose control (Cheng et al., 2017). Regeneration has not been shown in people; the basis is animal and cell work.

Benefit-Modifying Factors

  • Genetic polymorphisms: No study has tested whether variants such as APOE4 (a variant of the cholesterol-transport gene APOE that raises Alzheimer’s risk) modify FMD benefits; a planned Cedars-Sinai trial lists APOE4 among its study conditions (NCT06682767).
  • Baseline biomarkers: Benefits concentrate in people with elevated starting values; in Wei’s post hoc analysis, glucose, blood pressure, IGF-1, cholesterol and CRP fell mainly in those above risk cutoffs, with little change in normal ranges (Wei et al., 2017).
  • Body fat: Participants with obesity lost more BMI (body mass index, weight relative to height) than lean participants in Wei’s trial (Wei et al., 2017), so lean, metabolically healthy adults have less room for measurable gain.
  • Sex: No sex-specific efficacy analyses have been published; women made up 82% of dropouts in Wei’s trial (Wei et al., 2017), so tolerability, rather than proven efficacy, may differ by sex.
  • Pre-existing conditions: Type 2 diabetes, fatty liver and Crohn’s disease show the clearest measured gains (van den Burg et al., 2024; van den Burg et al., 2025; Kulkarni et al., 2026); cancer benefits remain unproven.
  • Age: Trials enrolled adults aged 18–75; no data exist beyond that range, and in older adults with sarcopenia (age-related muscle loss) further weight loss may offset metabolic gains (theoretical).

Potential Risks & Side Effects

High 🟥 🟥 🟥

Transient Energy-Deficit Symptoms

Fatigue, weakness, headache, dizziness, hunger and nausea are common during the five diet days and resolve after refeeding. Reported by Wei and Longo’s USC (University of Southern California) group, all grade 1–2 (mild to moderate) (Wei et al., 2017), van den Burg’s Leiden group (van den Burg et al., 2024) and Sulaj’s Heidelberg group, where 71–95% reported no adverse effect depending on the symptom (Sulaj et al., 2022). Dislike of the food also drives dropout.

Magnitude: Over 12 months, adverse events were registered in 19 participants on the diet vs 18 on usual care (51 vs 49 randomized); cycle-day symptoms were reported without a control-group rate (van den Burg et al., 2024).

Medium 🟥 🟥

No risk reaches Medium: beyond transient symptoms, the harms rest on small trials outweighed by larger null data, on drug labels extrapolated to the diet, or on animal work.

Low 🟥

Lean-Mass Loss ⚠️ Conflicted

Absolute lean mass fell in Wei’s randomized arm, but not relative to body mass (Wei et al., 2017). A Leiden trial found no fat-free-mass loss (van den Burg et al., 2024); a Verona trial found no muscle-force loss (Nardon et al., 2022). Net reading: meaningful muscle loss is unproven.

Magnitude: Fat-free mass 0.3 kg lower than usual care at 12 months (95% CI −1.1 to +0.6; not significant) (van den Burg et al., 2024).

Serious Adverse Events ⚠️ Conflicted

In the Leiden trial, eight serious adverse events occurred in five diet participants and none in controls; investigators judged none diet-related (van den Burg et al., 2024). Wei’s trial recorded no grade 3 or higher events (Wei et al., 2017). Net reading: no causal serious harm is shown.

Magnitude: Serious adverse events in 5 of 51 participants on the diet vs 0 of 49 on usual care, none attributed to the diet (van den Burg et al., 2024); 16% vs 8% in a 20-per-arm multiple sclerosis pilot (Choi et al., 2016).

Hypoglycemia with Glucose-Lowering Drugs ⚠️ Conflicted

Low intake raises hypoglycemia (low blood sugar) risk with insulin or sulfonylureas (drugs that force insulin release), per labeling (Glucotrol XL label). A Heidelberg trial halving long-acting insulin and pausing oral drugs found no excess (Sulaj et al., 2022). Net reading: pre-cycle dose reduction appears to contain risk.

Magnitude: Grade 1 hypoglycemia in 10% with the diet vs 11% on a Mediterranean diet; grade 2 in 5% vs 11% (Sulaj et al., 2022).

Speculative 🟨

Refeeding-Linked Tumor Promotion

In mice, refeeding after water-only fasting, not the FMD, boosted intestinal stem-cell proliferation and, with loss of the tumor-suppressor gene Apc, increased tumor formation (Imada et al., 2024). The basis is indirect animal data only.

Transient Immune-Cell Reduction

Cycles temporarily cut white-blood-cell counts by 40–50% in mice, with slight lymphocyte reductions in people (Choi et al., 2016). Infection susceptibility is untested; the basis is mechanistic.

Triggering Disordered Eating

Cyclic restriction could reinforce restrictive or binge eating in susceptible people; Prolon lists eating disorders as a contraindication (Prolon product page). No controlled data exist; the concern is theoretical.

Risk-Modifying Factors

  • Genetic polymorphisms: No gene variant has been studied; people with inherited disorders of fat burning or glucose production cannot adapt to fasting, and DIRECT excluded metabolic disorders affecting fasting adaptation (NCT02126449).
  • Baseline biomarkers: Low BMI, low blood pressure, low fasting glucose or protein deficiency magnify undernutrition, dizziness and hypoglycemia risks; Prolon excludes a BMI under 18.5 and prior protein deficiency (Prolon product page).
  • Sex: Women made up 82% of dropouts in Wei’s trial (Wei et al., 2017); no sex difference in adverse-event rates has been reported. Pregnancy and breastfeeding are exclusions.
  • Pre-existing conditions: Diabetes treated with insulin or sulfonylureas, heart failure, a history of fainting, advanced kidney disease and cancer-related wasting raise risk; trials excluded several of these (NCT03811587; NCT04147585).
  • Age: No data exist beyond age 75; older adults with sarcopenia or frailty have less reserve against lean-mass loss and dizziness, and Prolon excludes users over 70 (Prolon product page).

Key Interactions & Contraindications

Prescription drugs

  • Insulin and sulfonylureas (glipizide, glimepiride, glyburide): Reduced intake raises hypoglycemia risk; glipizide labeling states hypoglycemia is more likely when caloric intake is deficient (Glucotrol XL label). Mitigation: prescriber-directed dose reduction before cycles and home glucose checks.
  • SGLT2 (sodium-glucose cotransporter 2) inhibitors (drugs making kidneys excrete glucose: empagliflozin, dapagliflozin, canagliflozin): Ketoacidosis (dangerous blood-acid buildup) risk, even with glucose below 250 mg/dL; labeling lists reduced caloric intake and ketogenic (very-low-carbohydrate) diets as precipitants (Jardiance label). Mitigation: discussing withholding with the prescriber.
  • Metformin: In the Leiden trial, metformin users completed monthly cycles without diet-related serious harm while needing less glucose-lowering medication (van den Burg et al., 2024). Consequence: falling drug requirement. Mitigation: periodic HbA1c review with the prescriber.
  • Antihypertensives (blood-pressure drugs: lisinopril, losartan, amlodipine): Additive pressure lowering on diet days may cause dizziness or fainting (theoretical, given the diet’s blood-pressure effect in Wei et al., 2017). Mitigation: home blood-pressure checks and prescriber review if readings fall.
  • Diuretics (drugs that increase urine output: furosemide, hydrochlorothiazide): Lower sodium and potassium intake combined with diuretic losses may cause electrolyte imbalance (theoretical). Mitigation: electrolytes checked at baseline and after cycles.
  • Corticosteroids during chemotherapy (dexamethasone, an anti-inflammatory, anti-nausea steroid): Steroids raise blood glucose, countering the diet’s glucose lowering (theoretical); DIRECT omitted dexamethasone in the diet arm without excess toxicity (de Groot et al., 2020). Mitigation: steroid decisions only with the oncology team.
  • GLP-1 (glucagon-like peptide 1) receptor agonists (gut-hormone-mimicking weight-loss drugs: semaglutide, tirzepatide): Combined appetite suppression and calorie deficit may deepen lean-mass loss and nausea (theoretical). Mitigation: protein-adequate eating between cycles and body-composition tracking.

Over-the-counter medications

  • Acetaminophen (paracetamol): In a hospital case series, recent fasting was associated with liver injury at 4–10 g/day (Whitcomb & Block, 1994). Mitigation: staying within the labeled daily maximum during cycles and avoiding alcohol.
  • NSAIDs (non-steroidal anti-inflammatory drugs: ibuprofen, naproxen): Taken with little food, they may increase stomach irritation (theoretical). Mitigation: taking them with diet-day meals or using alternatives.

Supplements

  • Glucose-lowering supplements (berberine, chromium, cinnamon, alpha-lipoic acid): Additive glucose lowering may cause hypoglycemia, especially alongside diabetes drugs (theoretical). Mitigation: pausing them during cycles or monitoring glucose.
  • Blood-pressure-lowering supplements (magnesium, garlic extract, beetroot nitrate): Additive pressure lowering on diet days may cause dizziness (theoretical). Mitigation: pausing during cycles or home blood-pressure checks.
  • Protein and amino acid supplements (whey, branched-chain amino acids, leucine): Extra protein counteracts the diet’s low-protein design, likely blunting the IGF-1 and mTOR reductions (theoretical). Mitigation: deferring them to refeeding days.

Other interventions

  • Alcohol: Raises hypoglycemia risk with sulfonylureas (Glucotrol XL label) and fasting-related acetaminophen liver injury (Whitcomb & Block, 1994). Mitigation: avoiding alcohol during cycles.
  • Intense or endurance training: Low energy availability on diet days may cause dizziness and impaired performance (theoretical). Mitigation: light activity during cycles, harder training between them.

Thresholds below come from trial exclusion criteria and from the seller’s own safety list (Prolon, sold by L-Nutra).

Populations who should avoid Fasting-Mimicking Diet:

  • Pregnant or breastfeeding women (Prolon product page; NCT02158897)
  • People under 18 or over 70 years (Prolon product page)
  • BMI under 18.5 kg/m² (Prolon product page); trials excluded BMI under 19 kg/m² (NCT02158897)
  • Known, diagnosed or suspected eating disorder (Prolon product page)
  • Fever, active infection, malnutrition or prior protein deficiency (Prolon product page)
  • Allergy to nuts, sesame, soy or another diet ingredient (NCT03811587)
  • Myocardial infarction (heart attack) within the past 6 months (NCT03811587)
  • Creatinine clearance (a kidney-filtration measure) below 30 mL/min/1.73 m² (NCT03811587)
  • Heart failure of NYHA (New York Heart Association symptom scale) class 2 or higher, or LVEF (left ventricular ejection fraction, the share of blood pumped per beat) below 40% (NCT04147585)
  • History of syncope (fainting), including with past calorie restriction (NCT03811587; NCT04147585)
  • Diabetes treated with drugs that carry hypoglycemia risk (NCT04147585)
  • Liver or kidney disorders that low glucose and protein intake could worsen (NCT04147585)
  • Metabolic disorders impairing glucose production or adaptation to short fasting (NCT02126449)

Risk Mitigation Strategies

Doses, timings and thresholds below follow common practice unless cited.

  • Medication review before the first cycle: Prevents hypoglycemia; insulin and sulfonylurea doses are reduced by the prescriber beforehand, since labeling links deficient intake to hypoglycemia (Glucotrol XL label).
  • Glucose stop threshold: Prevents severe hypoglycemia; the cycle is stopped and carbohydrate eaten if glucose falls below 70 mg/dL, the ADA (American Diabetes Association) level 1 threshold (ADA, 2025); this cutoff earns ADA members no direct revenue.
  • SGLT2 inhibitor pause: Prevents ketoacidosis; withholding is discussed with the prescriber, mirroring the label’s advice to withhold at least 3 days before procedures with prolonged fasting (Jardiance label).
  • Gentle activity on diet days: Reduces dizziness and fainting; walking or yoga replaces intense training, and rising slowly from lying or sitting limits blood-pressure drops.
  • Hydration: Reduces headache and dizziness; about 2–3 L of water or herbal tea daily during cycles.
  • Protein-adequate refeeding: Limits lean-mass loss; about 1.0–1.2 g protein per kg body weight daily between cycles, with resistance training two or more times weekly.
  • Body-composition tracking: Detects lean-mass loss; a DEXA (dual-energy X-ray absorptiometry, a low-dose body-composition scan) every 6–12 months, reducing cycle frequency if lean mass falls from baseline.
  • Weight floor: Prevents undernutrition; cycling pauses if BMI approaches 18.5 kg/m², Prolon’s exclusion threshold (Prolon product page).
  • Skipping cycles during illness: Addresses the immune-cell reduction concern; cycles are deferred during fever or active infection, a listed Prolon contraindication (Prolon product page).
  • Oncology supervision: Prevents malnutrition and treatment conflicts; use during cancer therapy occurs only within oncology-team plans, as in DIRECT (de Groot et al., 2020).

Therapeutic Protocol

Doses are cited to their source; other parameters without a citation (timing, refeeding, titration) reflect common practice.

  • Standard Prolon regimen: Five consecutive days: day 1 ≈1,100 kcal (11% protein, 46% fat, 43% carbohydrate), days 2–5 ≈720 kcal (9% protein, 44% fat, 47% carbohydrate) of plant-based soups, bars, teas and a supplement (Brandhorst et al., 2024).
  • Research schedules: Monthly for three cycles in generally healthy adults (Wei et al., 2017); monthly for 12 months alongside primary diabetes care (van den Burg et al., 2024).
  • Maintenance schedule (Longo): Valter Longo, who popularized the diet, pairs a mostly plant-based, pescatarian (fish but no other meat) everyday diet with five-day fasting-mimicking periods four times a year (The Longevity Diet).
  • Oncology schedule: In DIRECT, the diet ran for 3 days before and during each chemotherapy cycle (de Groot et al., 2020); Milan oncologists used cyclic five-day FMD with standard therapy (Vernieri et al., 2022).
  • Competing approaches: Water-only fasting, continuous calorie restriction and time-restricted eating are alternatives; in mice, FMD, not water fasting, reversed bowel pathology (Rangan et al., 2019); human weight loss did not differ significantly from daily restriction (Sadeghian et al., 2021).
  • Homemade versions: Some practitioners replicate the published calorie and macronutrient targets with whole foods; these versions have not been tested in trials.
  • Time of day: Each day’s allotment may be eaten whenever participants choose; trial kits were boxed by day so items could be spread across breakfast, lunch, dinner and snacks (Brandhorst et al., 2024).
  • Refeeding: Day 6 typically starts with light, easily digested meals such as soups, rice and vegetables before returning to normal portions.
  • Genetic polymorphisms: No genotype-guided protocol exists; no trial has adjusted calories, frequency or composition by gene variants such as APOE4.
  • Sex: No sex-specific protocol exists; given higher dropout among women in Wei’s trial (Wei et al., 2017), dietitian support may matter for adherence.
  • Age: Trials enrolled adults up to 70–75 years; Prolon excludes users over 70 (Prolon product page), and older adults typically use fewer cycles with protein-rich refeeding.
  • Baseline biomarkers: People with elevated glucose, blood pressure, IGF-1 or CRP show the largest changes (Wei et al., 2017); those with normal values may use less frequent cycles.
  • Pre-existing conditions: In type 2 diabetes, monthly cycles were combined with general-practitioner medication adjustment (van den Burg et al., 2024); in cancer, the diet is used only within oncology protocols.

Discontinuation & Cycling

  • Lifelong or short-term: The diet is designed as open-ended periodic cycles; trials ran 3–12 months, and lifelong use has not been studied.
  • Withdrawal effects: None are known; diet-day symptoms resolve between cycles after refeeding (van den Burg et al., 2024).
  • Tapering: Not required; each cycle ends with a light day-6 refeed rather than a taper.
  • Durability after stopping: Weight, blood-pressure and IGF-1 changes persisted about 3 months after the last cycle in volunteers (Wei et al., 2017); in type 2 diabetes, gains had generally faded a year after stopping (van den Burg et al., 2023).
  • Cycling for efficacy: Cycling is intrinsic to the method; frequency ranges from monthly to quarterly, and no trial has compared frequencies.

Sourcing and Quality

  • Commercial kit: L-Nutra supplied the diet in Longo’s USC trials (Brandhorst et al., 2024) and supplies it in registered trials elsewhere (NCT05832086); its proprietary formulation means trial results apply most directly to that product.
  • Formulation: Kits contain plant-based soups, bars, olives, crackers, teas and a vitamin-mineral supplement; nut, sesame and soy content led trials to exclude people with these allergies (NCT03811587).
  • Regulatory status of the product: The seller states that Prolon’s ingredients are generally recognized as safe and that its claims are not evaluated by the FDA (U.S. Food and Drug Administration) (Prolon product page); the FDA does not approve food products (FDA consumer update).
  • Third-party testing: No independent certification program covers FMD kits; the manufacturer’s compliance with food-safety rules is the main quality safeguard.
  • Homemade versions: Whole-food versions must hit low protein (about 9–11% of energy) and low sugar to mimic fasting; exceeding these targets likely blunts the intended growth-signal reductions.

Practical Considerations

  • Time to effect: Ketones rise and IGF-1 falls by day 5 of cycle one (Wei et al., 2017); weight and blood pressure change after three cycles; HbA1c and liver fat by 6–12 months (van den Burg et al., 2024; van den Burg et al., 2025).
  • Common pitfalls: Adding protein snacks, training hard on diet days, not adjusting diabetes medication and overeating on refeeding days undermine safety or effect.
  • Adherence: A quarter of participants dropped out in Wei’s trial (Wei et al., 2017), and only 20% of DIRECT patients completed the diet with every chemotherapy cycle (de Groot et al., 2020).
  • Regulatory status: The seller labels Prolon as not intended to diagnose, treat, cure or prevent any disease (Prolon product page), and the FDA does not approve food products (FDA consumer update), so disease uses are investigational.
  • Cost: A five-day Prolon kit lists at $215, or $182.75 by subscription (Prolon product page); four kits a year cost roughly $730–860, and insurers generally do not reimburse them.
  • Payer incentives: In Dutch primary care, the program cost €2,241 more than usual care in year one; a lifetime model estimated €8,369 per quality-adjusted life-year (a year in full health) (van den Burg et al., 2025); short budget horizons can bias payers toward reimbursed drugs.

Interaction with Foundational Habits

  • Sleep: Indirect and unmeasured; no trial has assessed sleep. Hunger and ketosis may disturb sleep on diet days in some people (theoretical). Eating the last allotment in the evening and avoiding caffeine after midday are common practical adjustments.
  • Nutrition: Direct; the diet is itself a nutrition intervention. Between cycles, Longo pairs it with a mostly plant-based, pescatarian everyday diet (The Longevity Diet); adequate protein between cycles protects lean mass, especially in older adults.
  • Exercise: Potentially blunting on diet days through low energy and protein intake; across three cycles, physically active young men lost no muscle force or volume (Nardon et al., 2022). Light activity on diet days and resistance training between cycles are common practice.
  • Stress management: Indirect; fasting acts as a physiological stressor, and FMD raised the stress hormone corticosterone in mice (Choi et al., 2016). Irritability on diet days is commonly reported, so cycles are often scheduled away from high-demand periods.

Monitoring Protocol & Defining Success

Baseline testing: Before the first cycle, a baseline assessment documents starting values and screens for exclusions: weight, waist and blood pressure; fasting glucose and HbA1c; electrolytes and kidney function; IGF-1; and body composition by DEXA where available. People taking glucose-lowering or blood-pressure drugs review dosing with their prescriber first.

Ongoing monitoring: During each cycle, weight, symptoms and, for people on glucose-lowering drugs, home glucose are checked daily. Laboratory tests are repeated 5–7 days after the third cycle, the timepoint used in trials (Wei et al., 2017), then every 6–12 months while cycling continues, with DEXA every 6–12 months. Success is defined as lower weight, waist, glucose, HbA1c or liver fat than at baseline, without lean-mass loss or recurrent hypoglycemia.

Biomarker Optimal Functional Range Why Measure It? Context/Notes
Fasting glucose 70–99 mg/dL (standard reference range) Expected to change; safety (hypoglycemia) Draw after an 8–12 h fast; for people on glucose-lowering drugs, home readings below 70 mg/dL meet the ADA (American Diabetes Association) hypoglycemia threshold (ADA, 2025), a cutoff that earns ADA members no direct revenue
HbA1c 4.0–5.6% (standard reference range) Expected to change in diabetes and prediabetes Reflects about 3 months of glucose; repeating no sooner than 3 months apart; pairs with fasting glucose
Blood pressure Below 120/80 mmHg (“normal,” 2025 AHA/ACC guideline) Expected to change; safety (dizziness with blood-pressure drugs) Source: AHA (American Heart Association) and ACC (American College of Cardiology) guideline (Jones et al., 2025); conventional clinical reference is the same normal category; these professional bodies’ members treat hypertension, so lower cutoffs can widen treated populations
Body weight and waist No established target; tracking change from own baseline Expected to change; safety (undernutrition) Measured fasted at the same time of day; a BMI near 18.5 kg/m² (Prolon exclusion threshold) signals a pause (Prolon product page)
Lean mass (DEXA) No established target; tracking change from own baseline Safety (a fall changes cycle frequency) Absolute lean mass fell in one randomized arm (Wei et al., 2017); same scanner and hydration state for repeat scans
IGF-1 No established target; tracking change from own baseline Expected to change Lab reference ranges are age-specific; participants above 225 ng/mL showed the largest drops (Wei et al., 2017); measuring 5–7 days after a cycle
Sodium, potassium and creatinine Sodium 135–145 mmol/L, potassium 3.5–5.0 mmol/L, creatinine about 0.6–1.3 mg/dL, lower in women (standard reference range) Safety (electrolytes and kidney function) Creatinine clearance below 30 mL/min/1.73 m² was a trial exclusion (NCT03811587); most relevant with diuretics
β-hydroxybutyrate (blood or urine ketones) No established target; a rise during the cycle confirms a fasting-like state Expected to change; safety with SGLT2 inhibitors Rose by day 5 in trials (Wei et al., 2017); with SGLT2 inhibitors, ketoacidosis can occur with glucose below 250 mg/dL (Jardiance label)

Qualitative markers:

  • Energy, hunger and mood on diet days versus refeeding days
  • Dizziness or light-headedness on standing
  • Sleep quality during cycles
  • Exercise capacity and strength between cycles
  • Bowel symptoms in people with inflammatory bowel disease
  • Adherence: number of completed five-day cycles

Emerging Research

  • Colorectal cancer recurrence: A Fudan University randomized trial of perioperative (around-surgery) FMD in 602 patients; primary endpoint 3-year disease-free survival (share alive without cancer return); completion October 2027 (NCT05384444). A positive result would give the first survival-type benefit; a null result would weaken the cancer-adjunct rationale.
  • Triple-negative breast cancer (tumors lacking estrogen, progesterone and HER2 receptors; LESLIE): A KU Leuven phase 2 trial in 356 patients testing FMD plus exercise during neoadjuvant (pre-surgery) therapy; primary endpoint pathological complete response; completion 2031 (NCT06831955). A positive result would upgrade tumor-response evidence, without separating diet from exercise; a null would echo DIRECT.
  • Prostate cancer control: A phase 2 trial in 138 men comparing FMD with a standard anticancer diet; primary endpoint PSA (prostate-specific antigen) falling to 0.2 or lower within 6 months; completion 2029 (NCT05832086). Positive results would extend benefit to hormone-sensitive cancer; null results would narrow it.
  • Aging-biomarker follow-up: A Fondazione Valter Longo trial in 135 adults cycling FMD every 3 months for a year; primary endpoint maintenance of fat loss; completion June 2027 (NCT07255300). Positive results support quarterly maintenance; null results suggest regain. The sponsor has a stake in the diet.
  • Parent aging trial: The 501-participant trial of FMD with or without Longo’s everyday diet (NCT05698654) passed its registered completion in January 2025; its status is listed as unknown, with no results posted or published as of October 2026.
  • FMD versus rapamycin (NIBBLE): A Cedars-Sinai phase 1/2 trial in 60 adults comparing an approximately 8-week FMD program or low-dose rapamycin with placebo; primary endpoint adverse events; not yet recruiting, completion 2028 (NCT06682767). A clean safety profile would support combined longevity trials; excess events would argue caution.
  • Ulcerative colitis: A Stanford trial in 75 patients comparing three monthly cycles with a regular diet; primary endpoint partial Mayo score response (a colitis activity score); completion 2028 (NCT03615690). A positive result would extend the Crohn’s benefit; a null would confine it.
  • Durability of benefit: Benefits in type 2 diabetes generally faded after periodic fasting stopped (van den Burg et al., 2023); longer trials with clinical endpoints such as cardiovascular events could strengthen or weaken the case.
  • Refeeding and cancer risk: Mouse work showing post-fast refeeding promotes intestinal tumors (Imada et al., 2024) raises a question human cancer-screening data from cycling cohorts could answer.
  • Independence of evidence: Critiques of the developer-led evidence base (Sofi, 2020) make trials without L-Nutra funding or product supply the key test of whether effects replicate.

Conclusion

The fasting-mimicking diet is a short, repeated period of very low-calorie, low-protein, plant-based eating meant to reproduce the effects of fasting while still allowing food. For health-focused adults willing to tolerate five demanding days at a time, the clearest human evidence concerns modest weight and body-fat loss and better blood sugar control in type 2 diabetes, with more limited support for reduced liver fat, a small blood-pressure drop in people with raised pressure, and calmer symptoms in Crohn’s disease, a bowel disease. Other claims, including longer life, a younger biological age, tissue regeneration, brain protection and help during cancer treatment, rest mainly on animal work, unproven blood markers or mixed small trials.

The main downsides are tiredness, headache, dizziness and hunger during the diet days, a risk of low blood sugar for people on certain diabetes drugs that appears manageable when doses are lowered, and uncertainty about muscle loss in older or frail adults. Benefits seem to fade once cycles stop, so any gain likely depends on continued use.

The evidence base is young, mostly short-term, and largely produced by the diet’s developer, who holds a stake in the company selling the leading kit, or by groups using that company’s product or funding. Results from fully independent research teams remain few, so the overall picture is promising for risk factors such as weight, blood sugar and liver fat but still uncertain for long-term health and lifespan.

Top - Benefits - Risks - Protocol