Human Autophagy and Autophagic Flux Responses to Fasting Interventions: Evidence and Methodological Nuances

In adults, what direct evidence shows that time-restricted eating, intermittent fasting, prolonged fasting, or subsequent refeeding alters autophagy or autophagic flux in human tissues or cells, and how strong is this evidence?

Direct measurement of autophagic flux in humans shows that multi-day fasting (3-5 days) activates autophagy in metabolically active tissues like skeletal muscle and specific immune cell populations, but daily time-restricted eating and overnight fasting produce inconsistent or minimal effects that vary substantially by tissue type, measurement method, and individual health status, with the strongest evidence coming from studies using lysosomal inhibitors to measure flux rather than static marker abundance.

Abstract

Fourteen studies examining autophagy responses to fasting interventions in adults reveal heterogeneous evidence quality and tissue-specific responses. Studies using validated autophagic flux assays with lysosomal inhibitors provide the strongest evidence: a 5-day fasting-mimicking diet enhanced autophagic flux in peripheral blood mononuclear cells compared to controls (p=0.0191) [1], and 4-day complete fasting activated flux in neutrophils when measured appropriately [2]. In contrast, 12-hour overnight fasting followed by refeeding did not alter flux in two studies of 42 healthy adults [3, 4]. Prolonged fasting (72 hours) consistently activated autophagy pathway components in skeletal muscle, including ~50% decreased mTOR phosphorylation [5] and increased ULK1 expression [6], though static marker measurements showing simultaneous increases in LC3B-II and p62 prevented definitive flux determination [5, 6]. Daily time-restricted eating (6-8 hour windows) produced modest effects after 4 days to 8 weeks, increasing LC3A expression by 22% [7] and ATG-5 levels [8]. Tissue-specific responses emerged, with blood autophagy markers largely unaffected by 12-hour fasting in critically ill patients and healthy controls [9], while fasting increased autophagic monocytes but not lymphocytes [10]. Chronic kidney disease abolished the autophagy activation observed in healthy subjects during overnight fasting (γLC3 ratio 0.92 vs 1.78, p<0.0001) [11]. Exercise intensity proved more significant than fasting state in activating skeletal muscle autophagy through AMPK-dependent pathways [12]. The evidence suggests multi-day fasting activates autophagy in metabolically active tissues when measured with validated flux methods, while shorter daily fasting periods show inconsistent effects that vary by tissue type, measurement timing, and individual metabolic capacity.

Flow Diagram

Paper search

PubMed Corpus

We performed a keyword search across the PubMed corpus.

(autophagy OR autophagic flux OR LC3 OR SQSTM1 OR p62 OR ULK1 OR Beclin-1 OR ATG) AND (fasting OR starvation OR "time-restricted eating" OR "time restricted feeding" OR intermittent fasting) AND humans

The search returned 300 total results from PubMed.

Elicit Corpus

We performed a semantic search across over 138 million academic papers from the Elicit search engine, which includes all of Semantic Scholar and OpenAlex.

We ran these queries:

• "human fasting skeletal muscle blood leukocyte PBMC autophagy LC3 p62"
• "prolonged fasting refeeding human autophagy autophagic flux biomarker study"

The searches returned 600 total results from Elicit.

We retrieved 794 papers most relevant to the query for screening.

Screening

Abstract screening

We screened in sources based on their abstracts that met these criteria:

• Human adults: Does the title or abstract report a study conducted in human adults (18 years or older), or a human mixed-age cohort that can plausibly include adults?
• Primary completed study: Is this a primary report of a completed human study, rather than a review, protocol, design paper, editorial, erratum, or conference abstract?
• Fasting exposure: Does the title or abstract evaluate a defined fasting, time-restricted eating/feeding, intermittent fasting, alternate-day fasting, or refeeding-after-fasting exposure?
• Autophagy-related outcome: Does the title or abstract state that it measures autophagy, autophagic flux, or an autophagy-related molecular/cellular marker?

Papers that failed any strict criterion were automatically excluded. For the remaining papers, we considered all screening questions together and made a holistic judgement about whether to screen in each paper.

36 papers passed abstract screening and moved to full-text screening.

At abstract screening, the number of papers excluded for each primary reason was:

• Human adults: n = 463
• Primary completed study: n = 67
• Fasting exposure: n = 144
• Autophagy-related outcome: n = 82
• Other / below screening threshold: n = 2

Full-text screening

We then screened papers based on their full text using these additional criteria:

• Human adult data: Include only reports with analyzable data from adults aged 18 years or older.
• Defined exposure: Include only studies with a sufficiently described fasting or refeeding exposure, including duration or daily eating window.
• Direct autophagy measure: Include only studies measuring autophagy or autophagic flux directly in a human biological specimen, using at least one molecular, protein, imaging, or lysosomal/autophagosome marker. Exclude studies with metabolic, inflammatory, longevity, or clinical outcomes only.
• Primary completed report: Include only completed primary research reports with results, excluding reviews, protocols, narrative pieces, and conference abstracts.

Papers that failed any strict criterion were automatically excluded. For the remaining papers, we considered all screening questions together and made a holistic judgement about whether to include each paper in the final analysis.

14 papers passed full-text screening and moved to data extraction.

At full-text screening, the number of papers excluded for each primary reason was:

• Human adult data: n = 1
• Defined exposure: n = 1
• Direct autophagy measure: n = 4
• No full text: n = 16

Data extraction

We asked a large language model to extract each data column below from each paper. We gave the model the extraction instructions shown below for each column.

• Study identity and design: Extract citation, country if reported, design (randomized, crossover, controlled feeding, prospective mechanistic, or other), sample size analyzed, and population characteristics.
• Exposure and comparator: Extract fasting regimen precisely: daily eating window, fasting duration, energy intake permitted, number of cycles/days, physical activity instructions, and comparator or baseline.
• Biological specimen and timing: Extract specimen/tissue/cell type, collection timing relative to last caloric intake and refeeding, and whether repeated within-person sampling occurred.
• Autophagy measurement and interpretability: Extract each autophagy-related assay/marker, whether it assesses flux versus a static abundance/expression proxy, direction of change, and author-stated assay limitations.
• Primary autophagy findings: Extract numerical results where available, uncertainty/statistical test, and authors' conclusion regarding autophagy.
• Metabolic context and safety: Extract ketones, glucose, insulin, weight change, adverse events, and discontinuations only if reported.
• Risk-of-bias signals: Extract design features relevant to bias: allocation/randomization, control/comparator, blinding of laboratory outcomes, missing data, pre-specification, and conflicts of interest.

Results

Characteristics of included studies

Fourteen studies examining autophagy in response to fasting or time-restricted eating were included. Full texts were retrieved for all studies. Sample sizes ranged from 7 to 70 participants, with most studies involving healthy adults. Studies were conducted primarily in Europe (Belgium, Denmark) and the United States, with additional studies from Taiwan and Russia. Study designs included randomized crossover trials, controlled feeding studies, and prospective mechanistic studies.

The studies examined diverse fasting regimens ranging from 8-hour overnight fasts to 4-day complete fasts, with measurements taken in various tissues including skeletal muscle, blood cells, and peripheral blood mononuclear cells. Most studies employed crossover designs where participants served as their own controls, with measurements before and after fasting interventions.

Autophagy measurement methods and interpretability

Studies employed heterogeneous assay methods with varying capacity to distinguish autophagic flux from static marker abundance. LC3 (microtubule-associated protein 1A/1B-light chain 3) processing was the most commonly reported marker, assessed through LC3B-II/LC3B-I ratios [1, 5, 10, 13], LC3 lipidation [2], or LC3A/LC3B gene expression [7]. Additional markers included p62/SQSTM1 (sequestosome 1) [5, 12, 13], ULK1 (unc-51 like autophagy activating kinase 1) phosphorylation [5, 6, 12], ATG (autophagy-related) proteins [8], Beclin-1 [8, 11, 14], and mTOR (mechanistic target of rapamycin) signaling [5].

Critical methodological distinctions emerged regarding flux versus static measurements. Several studies measured autophagic flux using lysosomal inhibitors: Espinoza et al. treated PBMCs with chloroquine ex vivo, measuring LC3B-II accumulation as a validated flux indicator [1]. Similarly, Pietrocola et al. injected leupeptin before blood collection in mice and cultured human leukocytes ex vivo with leupeptin, finding autophagy activation detectable only with protease inhibition [2]. Singh et al. 2024 and 2025 added lysosomal inhibitors directly to whole blood, quantifying LC3B-II build-up via ELISA as a physiologically relevant flux measure [3, 4]. In contrast, most studies measured static abundance of autophagy markers without flux assessment, leading to interpretive challenges.

The LC3B-II/LC3B-I ratio exemplifies this limitation. While commonly used as an autophagy indicator, elevated LC3B-II can reflect either increased autophagosome formation or impaired autophagosome degradation [5]. Vendelbo et al. found LC3B-II increased ~30% during 72-hour fasting but p62 also increased ~10%, complicating interpretation since p62 is typically degraded during autophagy [5]. The authors noted they could not definitively determine autophagic flux from these conflicting markers [5]. Bak et al. reported similar patterns: ULK1 increased and its inhibitory phosphorylation decreased with fasting, suggesting autophagy activation, yet p62 mRNA and protein also increased [6]. Both studies acknowledged these data do not directly measure flux [5, 6].

Studies using complementary markers provided more robust evidence. Schwalm et al. 2015 measured ULK1 phosphorylation, LC3bII/I ratio, p62/SQSTM1 protein, and autophagy-related gene expression, finding decreased LC3bII and p62 after high-intensity exercise, indicating increased flux [12]. The decreased levels of both markers, rather than just one, more convincingly demonstrated autophagic degradation. Similarly, Kröpfl et al. combined flow cytometry for LC3B-positive cells with Western blotting for LC3BII/I ratios, though they found a negative correlation between these measures in monocytes [10], suggesting complex regulation not fully captured by single markers.

Multiple studies highlighted tissue-specific autophagy responses. Van Dyck et al. found autophagy markers in blood samples largely unaffected by 12-hour fasting in critically ill patients and healthy controls, leading authors to conclude blood may not reflect autophagy in other tissues [9]. Kröpfl et al. observed differential responses in lymphocytes versus monocytes: fasting increased autophagic monocyte concentrations but not lymphocytes, while exercise increased autophagic lymphocytes but not monocytes [10]. Chen et al. demonstrated that overnight fasting activated autophagy in healthy subjects' leukocytes but not in chronic kidney disease patients [11], indicating disease state can abolish typical autophagy responses.

Author-stated limitations acknowledged these measurement challenges. Multiple studies noted that gene expression may not reflect protein levels or functional activity [7], single time points may miss dynamic changes [4], and blood autophagy may not represent responses in metabolically active tissues like liver or muscle [4, 9]. Espinoza et al. cited small sample size and short assessment period as limitations [1], while Karahan et al. noted their assessment of ATG-5 and Beclin-1 alone, without p62/SQSTM1 evaluation, limited autophagy flux interpretation [8].

Autophagy findings by fasting exposure

Time-restricted eating (daily eating windows)

Two studies examined daily time-restricted eating with 6-8 hour eating windows. Jamshed et al. found early time-restricted feeding (eTRF; 8am-2pm, 6-hour window) increased LC3A gene expression by 22% in the morning compared to a 12-hour control window (8am-8pm), with p=0.001 [7]. This increase occurred in whole blood cells after 4 days of eTRF [7]. However, LC3A is a static gene expression marker rather than a direct flux measure [7], and the authors noted gene expression may not reflect protein levels [7]. Metabolic context included decreased 24-hour glucose levels by 4±1 mg/dL (p=0.0003) and increased morning ketones by 0.03±0.01 mM (p=0.009) [7].

Karahan et al. compared energy restriction alone versus energy restriction combined with 16:8 time-restricted eating (eating window 10am-8pm) in 48 adults with metabolic dysfunction-associated fatty liver disease [8]. ATG-5 levels increased significantly only in the time-restricted group after 8 weeks, from 0.74 to 0.95 ng/mL (p=0.03) [8], with a positive correlation between serum FGF-21 and ATG-5 (R=0.343, p=0.02) [8]. ATG-5 represents a static serum marker [8], and the study acknowledged limitations in assessing flux without evaluating p62/SQSTM1 [8]. Both groups showed weight loss and improved metabolic markers including decreased fasting glucose and liver enzymes [8].

Intermittent fasting and overnight fasts

Studies of 8-15 hour overnight fasts revealed tissue-specific and context-dependent autophagy responses. Chen et al. demonstrated that 12-hour overnight fasting induced autophagy activation in healthy subjects, as indicated by LC3-I to LC3-II conversion (γLC3 ratio 1.78) and increased Atg5 and Beclin-1 mRNA [11]. This response was absent in chronic kidney disease patients (γLC3 ratio 0.92-1.22, p<0.0001) [11], and hemodialysis did not restore autophagy activation [11]. Measurements were taken in leukocytes collected after overnight fasting and 2 hours after breakfast [11].

Two studies by Singh et al. examined whether a high-protein meal after 12-hour overnight fasting altered autophagy in 42 healthy adults. Using lysosomal inhibitor-enhanced ELISA to measure autophagic flux in PBMCs [3, 4], both studies found no change in flux between the fasting state and 1 hour post-meal consumption [3, 4]. The authors noted this challenges assumptions that acute nutritional interventions alter autophagy in humans [4] and suggested the 1-hour post-meal sampling may not have been optimal timing [4]. Notably, females had higher autophagic flux than males (p=0.0031) [3], representing sexual dimorphism independent of the meal intervention.

Schwalm et al. 2017 examined 8-hour overnight fasting combined with 2-hour high-intensity cycling in 7 trained athletes [13]. In vastus lateralis muscle, the LC3bII/LC3bI ratio decreased postexercise only in the fasted state (p=0.019) [13], while mitochondrial LC3bII and p62/SQSTM1 remained unchanged [13]. The authors concluded mitophagy was not activated during or early after exercise regardless of nutritional state, though muscle cells in the fed state appeared to prepare mitochondria for degradation at later time points [13].

Kröpfl et al. found 14-15 hour fasting increased circulating autophagic monocyte concentrations (p=0.026) but not lymphocytes compared to a fed condition with standardized breakfast [10]. Conversely, acute exercise elevated autophagic lymphocyte concentrations (p=0.043) but not monocytes [10]. Western blot analysis showed higher LC3BII/I ratios negatively correlated with autophagic cell numbers (r=-0.74, p=0.02) [10], suggesting complex regulation where increased autophagy at the protein level associated with fewer autophagic cells in circulation [10].

Prolonged fasting (24+ hours)

Studies of 72-hour to 4-day fasts provided evidence of autophagy pathway activation with persistent interpretive challenges. Vendelbo et al. studied 8 healthy males during 72-hour fasting with only water permitted [5]. In skeletal muscle biopsies, LC3B-II expression increased ~30%, but p62 protein also increased ~10% [5]. mTOR phosphorylation decreased ~50%, along with reduced downstream signaling including 4EBP1, ULK1, and rpS6 phosphorylation [5]. The conflicting LC3B-II and p62 patterns made autophagic flux determination problematic [5], though decreased mTOR signaling supported autophagy pathway activation. Muscle net phenylalanine release increased significantly [5], occurring in the context of decreased glucose by 25% and initiated ketogenesis [5].

Bak et al. compared 12-hour versus 72-hour fasting in 9 lean and 9 obese young men [6]. After 72 hours, ULK1 mRNA and protein levels increased significantly in muscle and adipose tissue [6], while the ratio of inhibitory ULK1 phosphorylation (Ser 757) to total ULK1 decreased [6], indicating reduced autophagy suppression. However, p62 mRNA and protein levels also increased [6]. Obese subjects exhibited increased whole body lipolysis and decreased urea production rates both basally and after 72 hours fasting, compatible with muscle protein preservation [6]. The study noted limitations in expressing results per body weight, potentially underestimating metabolic fluxes in obese subjects [6].

Pietrocola et al. examined 4-day zero-calorie fasting (water, tea, coffee only) in 9 healthy volunteers [2]. Fasting provoked major changes in plasma metabolome yet only minor alterations in leukocyte intracellular metabolome [2]. White blood cells showed marked reduction in protein lysine acetylation in both nuclear and cytoplasmic compartments [2]. Autophagy activation, measured by LC3B lipidation, became detectable in neutrophils only after ex vivo culturing with leupeptin [2], indicating autophagic flux occurred but required lysosomal inhibition for detection [2]. Human volunteers experienced <2% weight loss despite the 4-day fast [2], contrasting with 20% weight loss in mice fasted 48 hours [2].

Fasting-mimicking diets and refeeding

Espinoza et al. conducted a randomized trial of two fasting-mimicking diet (FMD) formulations versus control in 30 healthy adults aged 25-65 years [1]. The FMD consisted of 1100 kcal on day 1 and 700-800 kcal on days 2-5, followed by refeeding [1]. Using chloroquine-treated PBMCs to assess autophagic flux [1], the control group showed higher LC3B-II/LC3B-I ratios than FMD groups by day 6 (p=0.0191) [1], indicating FMD enhanced autophagic degradation through increased flux [1]. Significant between-group differences in flux emerged by the end of the 6-day intervention (p<0.05) [1], with a trend toward increased flux in the ProLon group persisting 48 hours after refeeding on day 8 [1]. The intervention decreased fasting glucose, insulin, and HOMA-IR while increasing β-hydroxybutyrate [1], though the small sample size and short assessment period limited generalizability [1].

Van Dyck et al. investigated whether 12-hour nutrient interruption could initiate a fasting response in 70 prolonged critically ill patients [9]. The randomized crossover design alternated 12 hours of up-to-target feeding with 12 hours fasting on day 8±1 of ICU stay [9]. The 12-hour fast significantly increased serum bilirubin and β-hydroxybutyrate and decreased insulin requirements and serum IGF-I (all p≤0.001) [9], confirming metabolic fasting response initiation. However, autophagy markers in whole blood and isolated white blood cells remained largely unaffected by fasting in both patients and 23 matched healthy subjects [9]. Authors concluded blood samples may not reliably reflect autophagy at the tissue level [9], as three patients developed severe hypoglycemia during fasting [9].

Tkhakushinov & Lysenkov examined partial food deprivation (800-1200 kcal/day) for 12 days in 20 men across three age groups [14]. Using Beclin-1 as an autophagy marker [14], they found age was the leading factor influencing autophagy activity, with activity decreasing as age increased [14]. Weight category did not significantly impact autophagy activation during caloric restriction [14]. Correlations between autophagy activity (delta-beclin-1) and lipid metabolism parameters varied by age group, with negative correlations with HDL in young adults and negative correlations with LDL and total cholesterol in elderly [14]. The study measured static Beclin-1 abundance rather than flux [14].

Exercise interactions with fasting

Two studies from the same Belgian group examined how exercise intensity and nutritional state interact to influence skeletal muscle autophagy. Schwalm et al. 2015 randomly assigned 23 well-trained athletes to control, low-intensity (55% VO2 peak), or high-intensity (70% VO2 peak) 2-hour cycling in both fed and fasted states [12]. Exercise intensity proved more significant than diet in activating autophagy [12]. High-intensity exercise decreased LC3bII and p62/SQSTM1 levels in vastus lateralis muscle, indicating increased autophagic flux [12]. This flux increase associated with increased AMPKα activity, as evidenced by higher phosphorylation of AMPKαThr72 and ACCSer79 after high-intensity exercise (p<0.001) [12]. ULK1Ser317 phosphorylation increased after exercise (p<0.001) [12], while fasting alone did not significantly enhance autophagy compared to exercise intensity [12].

Synthesis

Methodological hierarchy and flux measurement

Studies using validated flux assays provide more reliable evidence than those measuring static markers. The three studies employing lysosomal inhibitors to directly measure autophagic flux—Espinoza et al., Pietrocola et al., and Singh et al.—represent methodologically superior designs [1–4]. Espinoza et al. demonstrated that a 5-day fasting-mimicking diet enhanced flux compared to controls [1], while Pietrocola et al. showed 4-day fasting activated flux in neutrophils when measured appropriately [2]. Conversely, Singh et al. found 12-hour overnight fasting followed by refeeding did not alter flux [3, 4]. These flux-based studies suggest fasting duration is critical: multi-day fasts activate measurable autophagy, while overnight fasts may not produce detectable flux changes within the 1-hour post-refeeding measurement window.

Studies relying on LC3B-II/LC3B-I ratios or p62 abundance without flux assessment face interpretive limitations. Vendelbo et al. and Bak et al. both found increased LC3B-II alongside increased p62 during 72-hour fasting [5, 6], patterns the authors acknowledged prevent definitive flux determination [5, 6]. The simultaneous increase in both markers could reflect either autophagosome accumulation from impaired degradation or increased p62 synthesis during stress, rather than increased flux. These ambiguous findings should be weighted less heavily than direct flux measurements when assessing fasting's effects on autophagy.

Tissue-specific responses and biological compartments

Different tissues show distinct autophagy responses to fasting, explaining apparently contradictory findings. Van Dyck et al.'s conclusion that blood samples may not reflect tissue-level autophagy [9] aligns with accumulating evidence of compartment-specific responses. Studies in skeletal muscle (Vendelbo, Bak, Schwalm 2015 and 2017) consistently showed activation of autophagy pathway components including decreased mTOR signaling [5] and increased ULK1 expression [6], even when flux measurements were ambiguous. In contrast, studies measuring autophagy in circulating blood cells (Van Dyck, Singh et al., Kröpfl) found either no changes [3, 4, 9] or cell-type-specific responses limited to particular leukocyte subsets [10].

Kröpfl et al.'s finding that fasting increased autophagic monocytes but not lymphocytes, while exercise increased autophagic lymphocytes but not monocytes [10], demonstrates that even within blood, different cell types respond distinctly to metabolic stressors. Pietrocola et al. observed autophagy activation only in neutrophils among leukocyte subpopulations from fasted volunteers [2]. These tissue- and cell-type-specific responses suggest fasting activates autophagy preferentially in metabolically active tissues like muscle while blood cells, particularly lymphocytes, may be less sensitive to fasting-induced autophagy.

Temporal dynamics and measurement timing

The timing of measurements relative to fasting and refeeding substantially affects detected autophagy changes. Chen et al. measured autophagy in leukocytes both after 12-hour overnight fasting and 2 hours post-breakfast [11], finding fasting induced autophagy activation in healthy subjects [11]. However, Singh et al. measured PBMCs only 1 hour after a high-protein meal following 12-hour fasting and finding no flux change [3, 4], noting the 1-hour sampling may not have been optimal [4]. Refeeding likely suppresses autophagy through mTOR reactivation, and the kinetics of this suppression may vary by tissue and nutrient composition.

Studies of prolonged fasting (72 hours to 4 days) measured autophagy at the end of fasting without immediate refeeding [2, 5, 6], capturing maximal autophagy activation. Espinoza et al. uniquely measured autophagy during the fasting-mimicking diet (days 4 and 6) and 48 hours after refeeding (day 8) [1], observing that increased flux in the ProLon group showed a trend toward persistence after refeeding [1]. This suggests autophagy activation may not immediately reverse upon refeeding, but the 48-hour measurement may reflect early re-suppression. The exercise studies (Schwalm 2015 and 2017) measured muscle biopsies immediately and 1 hour post-exercise [12, 13], capturing acute responses that differ from chronic fasting adaptations.

Dose-response: Fasting duration and autophagy magnitude

A dose-response relationship emerges wherein longer fasting durations show clearer autophagy activation. Four-day complete fasting (Pietrocola et al.) and 5-day fasting-mimicking diet (Espinoza et al.) demonstrated measureable flux increases [1, 2]. Seventy-two-hour fasting (Vendelbo, Bak) showed autophagy pathway activation through mTOR suppression and ULK1 upregulation [5, 6], even if flux remained ambiguous. Studies of 12-15 hour overnight fasts showed variable results: Chen et al. found activation in healthy subjects [11], but Singh et al. found no flux change when measured post-refeeding [3, 4], and Van Dyck et al. found 12-hour fasting insufficient to alter blood autophagy markers [9].

Daily time-restricted eating (6-8 hour windows) showed modest effects after multiple days. Jamshed et al. found 22% increased LC3A expression after 4 days of 6-hour eating windows [7], while Karahan et al. found increased ATG-5 after 8 weeks of 16:8 time-restricted eating [8]. The cumulative exposure in these studies (28 hours total fasting over 4 days for Jamshed; 672 hours over 8 weeks for Karahan) suggests repeated daily fasting cycles may produce different autophagy patterns than single prolonged fasts.

Metabolic context modulation

The metabolic state during fasting modulates autophagy responses. Studies reporting increased ketones (β-hydroxybutyrate) showed concurrent autophagy activation: Jamshed et al. found increased morning ketones with LC3A upregulation [7], Van Dyck et al. found increased β-hydroxybutyrate from 4 hours of fasting onwards with metabolic fasting response [9], and Espinoza et al. found increased ketones alongside enhanced autophagic flux in the fasting-mimicking diet group [1]. Ketones may serve as a biomarker of the metabolic shift necessary for autophagy activation.

Schwalm et al. 2015 demonstrated exercise intensity activates autophagy through AMPK-dependent pathways independently of fasting [12]. High-intensity exercise increased AMPKα phosphorylation and autophagic flux regardless of fed versus fasted state [12], suggesting AMPK activation—whether through energy depletion during exercise or nutrient deprivation during fasting—represents a common mechanism. This mechanistic convergence explains why exercise and fasting may activate autophagy through shared signaling pathways despite different triggers.

Disease state and baseline autophagy capacity

Chen et al.'s finding that chronic kidney disease patients failed to activate autophagy during overnight fasting, while healthy controls showed robust activation [11], demonstrates disease can abolish fasting-induced autophagy. Similarly, Van Dyck et al. studied critically ill ICU patients who showed minimal autophagy marker changes despite metabolic fasting response [9]. This suggests severe illness impairs cellular capacity for autophagy induction, potentially through inflammation, oxidative stress, or metabolic dysfunction. Bak et al. found obese subjects had lower adipose tissue autophagy-related protein expression despite higher total lipolysis [6], indicating obesity may blunt tissue-level autophagy responses even when whole-body metabolism responds to fasting.

Study quality considerations

Randomized crossover designs where participants serve as their own controls (Jamshed, Schwalm 2015 and 2017, Vendelbo, Bak) minimize inter-individual variability and provide stronger evidence than single-arm pre-post studies (Singh 2024 and 2025, Tkhakushinov) or case-control designs (Chen) [3–7, 11–14]. However, even well-designed crossover studies using static markers face interpretive limitations compared to studies measuring flux with lysosomal inhibitors. The randomized trial by Espinoza et al., despite its small sample size and industry sponsorship (though funders did not participate in conduct or analysis) [1], provides high-quality flux data using validated methodology.

Integration and clinical implications

Synthesizing across these factors, the evidence suggests multi-day fasting (≥72 hours) or fasting-mimicking diets activate autophagy in metabolically active tissues, particularly skeletal muscle, with measurable flux increases when assessed using appropriate methods. This activation requires sufficient duration to induce metabolic shifts including ketogenesis and substantial mTOR suppression. Shorter daily fasting periods (12-15 hours) show inconsistent autophagy effects, possibly due to measurement timing relative to refeeding, tissue examined, and individual variation in metabolic flexibility. Blood cells, particularly lymphocytes, appear less responsive to fasting-induced autophagy than muscle or specific monocyte populations. Disease states including chronic kidney disease and critical illness may abolish or substantially blunt autophagy responses to fasting.

Human studies measuring autophagy with validated flux assays remain sparse, limiting definitive conclusions about optimal fasting protocols for autophagy induction. The discordance between preclinical models showing robust autophagy activation and human studies showing modest or absent changes suggests species differences in autophagy regulation or highlights the limitations of current human measurement techniques. Future research requires standardized flux measurement protocols, tissue-specific assessments, and careful attention to measurement timing relative to fasting and refeeding cycles.