Abstract
Fasting is part and parcel of the practice of various religious groups. In the scientific literature, the term intermittent fasting (IF) was first reported in a 1946 paper demonstrating its association with increased longevity in rodents. Research has extended IF, particularly time-restricted eating (TRE), to Alzheimer's disease (AD), a progressive neurodegenerative disease characterized by neuritic plaques, neurofibrillary tangles, and neuronal loss. AD manifests in asymptomatic, mild cognitive impairment (MCI), and dementia phases. Delaying progression from MCI to dementia by one year could reduce dementia prevalence by millions. Currently, no pharmacological treatments can reverse or arrest MCI progression to dementia, making exploration of non-pharmacological interventions critical. TRE is a promising approach. AD brains exhibit decreased glucose uptake, while ketone utilization remains intact. Fasting for at least 8–12 h induces a cascade of molecuar events that lead to a metabolic switch from glucose to ketone utilization, providing an alternative energy source for AD brains. Preclinical studies demonstrate that TRE enhances cognitive function via hippocampal neurogenesis, autophagy, and reduced neuroinflammation. Human studies on TRE in MCI are limited but promising, often focusing on cardiometabolic outcomes, with little known about TRE targeting MCI. This review synthesizes current evidence on TRE and cognitive outcomes in humans, non-human primates, and rodents, and describes ongoing trials in MCI patients. We propose a theoretical model of direct and indirect pathways linking TRE with resistance to AD in the brain parenchyma, and identify gaps in knowledge regarding long-term cognitive effects and mechanistic pathways of TRE in MCI, urging rigorous clinical trials to establish TRE as a safe and possibly effective strategy to delay MCI progression to dementia.
Keywords
Alzheimer's disease (AD), the leading cause of dementia, affects approximately 7.2 million Americans 1 and 55 million people worldwide, with projections estimating 150 million dementia cases by 2050. 2 AD is characterized by amyloid-β (Aβ) plaque deposition, tau protein tangles, neuroinflammation, and neurodegeneration, leading to progressive cognitive and functional decline. 3 AD progresses through three phases: an asymptomatic phase, mild cognitive impairment (MCI), and dementia. 3 MCI, the intermediate stage between normal aging and dementia, is associated with increased dementia risk, and cardiovascular and metabolic risk factors such as insulin resistance accelerate progress.4–6 Delaying MCI progression by one year could reduce dementia prevalence by approximately 9 million cases, alleviating its $1.3 trillion global economic burden.7,8
Currently, no pharmacological treatments can reverse or halt the progression of MCI to dementia. Lifestyle interventions, including nutritional strategies, are promising, safe, and accessible options to mitigate neurodegeneration. 9 Fasting is part of the practice of various religious groups. 10 However, in the scientific literature, the term intermittent fasting (IF) was first reported in 1946 in a peer-reviewed publication showing that IF increases the longevity of rodents. 11 Time-restricted eating (TRE), a form of IF, restricts daily food intake to a specific window (e.g., 6–10 h), leading to a metabolic switch from glucose to ketone utilization 12 (Figure 1). This is critical as glucose uptake is impaired, while ketone utilization remains intact in AD; the former is likely due to defective glucose transporters. 13 The decreased glucose uptake extends to asymptomatic AD; a 1996 study showed reduced glucose uptake in the temporoparietal region of the brain in cognitively unimpaired apolipoprotein E ε4 (APOE ε4) homozygotes. 14
Fasting for at least 8 to 12 h triggers a metabolic switch that involves a cascade of molecular events leading to ketone bodies production. FFAS: free fatty acids; BDNF: brain-derived neurotrophic factor.
TRE enhances cognitive function through ketogenesis, autophagy, reduced neuroinflammation, and improved insulin sensitivity.15–18 TRE induces lipolysis, releasing free fatty acids which are metabolized into ketone bodies, primarily β-hydroxybutyrate, in the liver and astrocytes. Ketones reduce mitochondrial oxidative stress and stimulate brain-derived neurotrophic factor (BDNF) expression, promoting hippocampal neurogenesis.19,20 Additionally, TRE enhances neuronal stress resistance via increased superoxide dismutase 2 (SOD2) activity, autophagy, DNA repair, and reduced inflammation, while improving insulin sensitivity and cardiovascular health.17,21,22 These mechanisms position TRE as a potential intervention for MCI and AD.
This paper reviews past and current literature on TRE, focusing on cognitive outcomes in humans, non-human primates, and rodents. We highlight knowledge gaps in long-term human studies and mechanistic pathways, describe ongoing trials in MCI patients as identified on ClinicalTrials.gov, and advocate for rigorous, long-term randomized controlled trials to investigate TRE's potential benefits in AD prevention.
Methods
The scope of this review is narrative and interpretive. We searched four databases (i.e., PubMed, Scopus, Web of Science, and Google Scholar) to identify relevant studies published from 1946 (the first report of IF effects) through October 2025. Search terms included combinations of “time-restricted eating”, “intermittent fasting,” “mild cognitive impairment,” “Alzheimer's disease,” “dementia,” “cognitive function,” “cognition,” “neurodegeneration,” “ketone bodies,” “metabolic switch,” and “brain health”. With regard to clinical trial identification, we searched ClinicalTrials.gov (October 2025) using search terms “time-restricted eating,” “intermittent fasting,” “mild cognitive impairment,” and “Alzheimer's disease” to identify ongoing and completed trials examining TRE interventions in MCI or AD populations.
Similarities and differences between caloric restriction and intermittent fasting
Caloric restriction (CR) and TRE differ fundamentally in the mechanisms by which they lead to metabolic benefits. CR involves a sustained 20–40% reduction in daily caloric intake without altering meal timing. Importantly, moderate CR (approximately 20–25% reduction) does not reliably deplete hepatic glycogen stores sufficiently to trigger robust ketogenesis, as the liver can maintain adequate glycogen reserves under conditions of modest energy deficit. 23
In contrast, TRE restricts the daily eating window to 6–10 h, requiring fasting periods of 14–18 h without mandating CR. This extended fasting duration can be sufficient to deplete hepatic glycogen stores (which can occur after 8–12 h of fasting in humans), thereby triggering a metabolic switch from glucose to ketone body utilization. 23 Fasting durations of ≥14 h stimulate lipolysis, releasing free fatty acids that undergo hepatic β-oxidation, resulting in the production of ketone bodies (primarily β-hydroxybutyrate). 23 Circulating ketone concentrations may rise modestly (≈0.2–0.5 mM) within 8–12 h and can reach 1–2 mM by 48 h.24,25 Rodent studies show even more rapid ketone kinetics, with millimolar concentrations achieved within 24 h of fasting. 23
TRE can achieve metabolic benefits independent of weight loss, making it particularly suitable for MCI patients without obesity who may not tolerate sustained calorie deficit. While both CR and TRE improve insulin sensitivity and reduce oxidative stress, only TRE consistently induces the metabolic switch that may directly support brain bioenergetics in AD.
The 5:2 IF regimen is a distinct approach that combines elements of CR and IF. In this regimen, individuals consume approximately 500–600 kcal on two nonconsecutive fasting days per week while eating ad libitum on the remaining five days. This approach differs mechanistically from daily TRE in several ways. First, 5:2 fasting achieves metabolic benefits primarily through intermittent severe energy deficit rather than circadian optimization of meal timing. In addition, the substantial CR on fasting days typically results in greater weight loss (clinically meaningful reductions of approximately 3–5% of body weight) compared with daily TRE. Lastly, differences in adherence patterns have important implications for different populations; the 5:2 regimen requires episodic willpower to endure severe restriction, whereas TRE requires consistent daily discipline in meal timing. 23 26–29
In the context of MCI and AD, these distinctions are clinically relevant. Daily TRE may be preferable for older adults with MCI because (a) TRE does not mandate CR that could result in unintentional weight loss or nutrient insufficiency, (b) TRE provides consistent daily metabolic switching that may optimize circadian rhythms, typically disrupted in AD (particularly in advanced stages), and (c) TRE may be simpler to implement and sustain than the 5:2 regimen in people with cognitive impairments. However, future comparative effectiveness trials are needed to directly assess whether daily TRE or 5:2 fasting offers superior cognitive and metabolic outcomes in patients with MCI. 30
Circadian timing of TRE: Early versus late eating windows
In addition to fasting duration, the timing of the eating window is an important consideration and has implications relative to circadian biology. Early TRE (eTRE) restricts food intake to morning and early afternoon hours (e.g., 7 AM-3 PM), whereas later TRE shifts eating to afternoon and evening hours (e.g., 12 PM-8 PM).
A meta-analysis of 12 RCTs (n = 730) showed that both eTRE and later TRE led to moderate reductions in insulin resistance compared to participants not engaging in TRE. However, eTRE showed more effectiveness than later TRE in improving insulin resistance (early versus later TRE: −0.44; 95% CI, −0.86 to −0.02; p < 0.05). There was no significant difference in weight loss, fasting blood glucose, blood pressure, or lipid profile between eTRE and later TRE. 31
Circadian disruption is a hallmark of AD. Patients with AD exhibit reduced circadian amplitude, phase delays (later sleep and wake times), irregular sleep-wake patterns, and the sundowning syndrome, characterized by increased agitation, aggression, and confusion during late afternoon and evening hours.32–34 Late eating windows could potentially exacerbate circadian misalignment and worsen sundowning symptoms by conflicting with already-delayed rhythms and potentially stimulating alertness during desired sleep periods. However, this hypothesis requires direct empirical testing in AD populations with systematic assessment of sundowning severity across different TRE timing protocols.
In addition, eTRE (e.g., 7 AM- 3 PM) may pose significant social and practical challenges, as it excludes evening family meals and social dining; activities that are important for quality of life and which may protect against social isolation in older adults. Balancing circadian optimization with quality of life, social engagement, and long-term adherence will thus be critical in protocol design. Flexible protocols that allow occasional “social exceptions” (e.g., 1–2 evenings per week) while maintaining the eating window on most days may represent a pragmatic compromise that optimizes both biological benefits and real-world sustainability. 35
Reviews on TRE and cognitive outcomes
Since the original publication in 1946, 11 subsequent studies explored TRE in the context of aging and AD. 36 Mark Mattson popularized TRE in AD research via his book, The Intermittent Fasting Revolution, targeting the lay public. 37 Mattson's work suggests that TRE directly enhances neuroplasticity and reduces inflammation, while indirectly optimizing brain health through metabolic improvements.17,38 Metabolic switching induced by TRE can directly and indirectly impact the brain as illustrated in Figure 2. TRE, alongside exercise and cognitive stimulation, promotes neuronal resilience by increasing BDNF expression, enhancing neurogenesis, and reducing oxidative stress, while indirectly improving insulin sensitivity and cardiovascular health. This counteracts AD-related pathology such as neuroinflammation, Aβ accumulation, and tau aggregation. 39 Building on this, reviews have also explored TRE's broader cognitive benefits. One review found that TRE may improve cognitive function and reduce oxidative stress and inflammation in individuals with MCI.40,41 Researchers also proposed that TRE could reverse or delay AD pathological processes, 42 while others suggest that TRE reduces neuroinflammatory markers and protects cognition in AD. 43 TRE also has been explored for preventing age-related vascular cognitive impairment and dementia. 44 Collectively, these reviews hypothesize that TRE has potential to mitigate AD pathology, warranting further empirical investigation.
A proposed theoretical model of the direct and indirect pathways linking TRE with resistance to Alzheimer's disease in the brain parenchyma. TRE: time-restricted eating; BDNF: brain-derived neurotrophic factor.
Original research on TRE and cognitive outcomes in humans
Human studies on TRE, mainly conducted in Asia, Europe, or North America, vary in sample size (15 to >1300 participants) and age distribution, reflecting diverse objectives. These studies assess TRE's association with MCI or AD frequency, cognitive scores (e.g., Mini-Mental State Examination, Montreal Cognitive Assessment), and biomarkers (e.g., SOD2, inflammation). For example, an observational study of 99 patients with MCI showed that those regularly practicing TRE had improved cognitive function and increased SOD2 levels after 3 years, 18 with follow-up analysis reporting reduced DNA damage and inflammation. 45 A cross-sectional study of 883 older Italian adults found that a < 10-h eating window was associated with lower cognitive impairment frequency. 46 Conversely, a study of 1353 older Chinese adults reported that a < 10-h eating window was associated with higher cognitive impairment frequency and worse Mini-Mental State Examination performance in “orientation to place” and “attention/ calculation”. 47 Of note, the Mini-Mental State Examination is a bedside cognitive screening tool, not a standard neuropsychological battery, which may explain discrepancies. The conflicting results may also stem from differences in study design, population characteristics, or unadjusted confounders like diet quality. A single-arm study using an 8-week, 14-h nightly fasting intervention in 18 adults aged ≥ 65 years with self-reported memory decline revealed improved cognitive function and sleep. 48 A qualitative study reported high feasibility and acceptability of prolonged nightly fasting, despite some barriers. 49 These human studies, while promising, highlight the need for standardized TRE protocols and larger trials to resolve conflicting findings. Please refer to Table 1 for an overview of studies on TRE and cognitive outcomes in humans.
Overview of studies on TRE and cognitive outcomes in humans.
AD: Alzheimer's disease; ADL: Activities of Daily Living; APOE: apolipoprotein E; AVLT-H: Auditory Verbal Learning Test-Huashan version; BMI: body mass index; CDR-SB: Clinical Dementia Rating Scale Sum of Boxes; CI: confidence interval; CRP: C-reactive protein; CSF: cerebrospinal fluid; FFQ: food frequency questionnaire; HDL: high density lipoprotein; IF: intermittent fasting; MAPS: Memory and Attention Phone Screener; MCI: mild cognitive impairment; MMSE: Mini-Mental Examination State; MoCA: Montreal Cognitive Assessment; NVC: Neurovascular coupling; OR: odds ratio; RAVLT: Rey Auditory Verbal Learning Test; RCT: randomized controlled trial; SOD: superoxide dismutase; TRE: time-restricted eating; TRF: time-restricted feeding.
Original research on TRE and cognitive outcomes in non-human primates and rodents
Preclinical studies, primarily conducted in the USA, examine effects of TRE on AD pathology (Aβ, tau), cognitive performance (memory tasks), and metabolic markers in rodents and non-human primates. Rodent studies demonstrate reduced Aβ and tau pathology, improved maze task memory, increased BDNF, and reduced inflammation. 55 In triple-transgenic AD mouse models, TRE was shown to promote hippocampal neurogenesis and reduce oxidative stress. Notably, their learning and memory performance was restored to that of wild type mice that were fed ad libitum, whereas AD mice who were fed ad libitum were severely impaired. 56 These findings provide a strong mechanistic rationale for exploring TRE's potential in humans with MCI. Non-human primate studies conducted with rhesus monkeys focused on long-term CR rather than TRE; however, these long-term CR studies had limited AD-specific cognitive data.57–60
TRE and mild cognitive impairment: Studies identified from ClinicalTrials.gov
Two ongoing trials, both named Time-Restricted Eating in Alzheimer's Disease (TREAD), investigate TRE in AD; one is being conducted at Barrow Neurological Institute in Phoenix, Arizona (NCT06429124), and the other at the University of California San Diego (UCSD) Alzheimer's Disease Research Center in San Diego, California (NCT06548191). The Barrow trial is a pilot study evaluating a 16-h nightly fast and an 8-h eating window for 5 days per week over 3 months in patients aged 55–89 years with MCI, focusing on feasibility (recruitment, retention, acceptability, adherence, safety) and secondary outcomes in cognitive, behavioral, metabolic, and sleep measures. The UCSD study is a randomized controlled trial with an immediate versus delayed start control group, evaluating a 14-h nightly fast and a 10-h eating window (10 AM-8 PM) daily in adults aged ≥ 60 years with MCI or early AD, with follow-up up to 12 months and assessments at 3 and 6 months. This study focuses on feasibility, safety, cognitive performance, sleep quality, AD biomarkers, and metabolic health.
Two other trials have similarities but also differences from the above two TREAD trials. “Lifestyle for the BRAin Health - Time-restricted eating and mindfulness (LIBRA REMIND)” (NCT07042087) is an ongoing randomized two-arm clinical trial sponsored by St Anne's University Hospital Brno in the Czech Republic. This 9-month study includes adults aged 60–80 years and evaluates the effects of combining yoga-based mindfulness techniques with TRE on brain health, cognition, and metabolic outcomes in older adults with subjective cognitive impairment or MCI. NCT05858008 is another ongoing single-arm trial utilizing a 14:10 TRE intervention in women aged 45–95 years with MCI with a goal of assessing feasibility. “Calorie Restriction and Brain Function in Mild Cognitive Impairment” (NCT03872375) is a completed trial sponsored by Iowa State University in Ames, Iowa. This study investigated intermittent CR in participants with MCI, aiming to reduce metabolic dysfunction, improve glucose metabolism, and enhance brain and cognitive function. Participants engaged in calorie-restricted fasting protocols to evaluate impacts on executive function and other cognitive domains, providing data on the efficacy of calorie-focused intermittent fasting in MCI populations.
Comparing these studies to the aforementioned TREAD trials reveals key similarities and differences. Commonalities include a shared focus on non-pharmacological dietary interventions (forms of IF or CR) to improve cognitive and metabolic health in older adults at risk for or experiencing cognitive decline, such as MCI or early AD. All trials also emphasize feasibility, safety, and outcomes such as cognitive performance, sleep, and biomarkers, aligning with the goal of delaying neurodegeneration through metabolic switching. However, there are notable differences in intervention, design, and scope. For example, the TREAD studies prioritize standard TRE without calorie restriction, with shorter durations (3–6 months) and US-based locations targeting MCI and/ or AD specifically. In contrast, NCT07042087 integrates TRE with mindfulness (yoga), extends to 9 months, and focuses on adults with subjective cognitive impairment (not exclusively MCI/ AD), incorporating brain response assessments in a European setting. NCT03872375 emphasizes intermittent CR using a 5–2 schedule rather than TRE, and specifically addressed metabolic dysfunction in MCI, potentially offering weight-dependent benefits unlike the weight-independent focus of TRE in TREAD. These variations highlight potentially complementary approaches, with TREAD emphasizing timing-based fasting feasibility in AD populations, while the others explore integrated or calorie-focused strategies.
With regard to implications of circadian timing of TRE in ongoing trials, the Barrow TREAD trial employs a 16:8 TRE protocol with flexibility in eating window timing, while the UCSD TREAD trial specifies a 10 AM-8 PM eating window. Future trials should explicitly test early versus late eating windows with matched fasting durations (e.g., 7 AM-3 PM versus 12 PM-8 PM, both 8-h windows) to isolate the effects of circadian timing independent of fasting length. Moreover, interventions should consider individual chronotype (morning versus evening preference), as chrono-mismatched TRE may reduce adherence and blunt benefits. Assessing outcomes such as objective sleep quality (actigraphy, polysomnography), sundowning severity (behavioral assessments), and circadian biomarkers (dim light melatonin onset) will be essential for optimizing TRE protocols for AD populations.
Mechanistic biomarkers to track response to TRE
Comprehensive biomarker panels are important to establish TRE's mechanisms of action in MCI/ AD and identify responders versus non-responders. For example, β-hydroxybutyrate (βHB) is the primary ketone body produced during fasting and serves as a direct marker of metabolic switching. Fasting βHB levels should be measured throughout an intervention to determine whether the TRE protocol achieves adequate ketogenesis (e.g., target: ≥0.5 mM). Continuous ketone monitoring using wearable biosensors (e.g., Levels, Nutrisense) may provide real-time feedback on adherence and individual metabolic responses, while measuring acetone in breath offers a non-invasive alternative to blood ketone testing.
In addition, BDNF is a neurotrophin critical for synaptic plasticity, neurogenesis, and memory formation. Low serum BDNF is associated with cognitive decline and AD risk in humans, while TRE increases BDNF expression in animal models.61,62 Serum BDNF levels should be assessed as a potential mediator of TRE's cognitive effects, with the hypothesis that TRE increases BDNF and that BDNF changes correlate with cognitive improvement. AD-specific plasma biomarkers enable non-invasive assessment of AD pathology with high accuracy. Plasma p-tau217 is the most specific marker for AD pathology, particularly at scores of ≥ 0.34 units, showing excellent concordance with amyloid PET and tau PET. 63 Plasma Aβ42/Aβ40 ratio reflects brain amyloid burden, while plasma neurofilament light chain (NfL) indicates axonal injury and neurodegeneration. A reduction in p-tau217 or slowing of NfL accumulation with TRE would provide mechanistic support for disease modification. Inflammatory markers may also be relevant, as chronic neuroinflammation drives AD progression. 64 TRE reduces systemic inflammation in animal models and human metabolic studies. High-sensitivity CRP, IL-6, TNF-α, and IL-1β should be measured to assess TRE's anti-inflammatory effects. Additionally, emerging markers such as glial fibrillary acidic protein (GFAP), a marker of reactive astrogliosis, may provide insights into neuroinflammatory changes. 64
Metabolically, the influence of TRE on insulin resistance and glucose metabolism in MCI/ AD can be evaluated with homeostasis model assessment of insulin resistance (HOMA-IR), fasting insulin, and hemoglobin A1c (HbA1c). Continuous glucose monitoring (CGM) captures 24-h glucose variability and time-in-range, providing detailed metabolic phenotyping and responses to TRE.
Furthermore, TRE reduces oxidative damage in preclinical models. Biomarkers such as 8-OHdG (8-hydroxy-2′-deoxyguanosine) for DNA damage, malondialdehyde (MDA) for lipid peroxidation, and antioxidant enzyme activity (SOD, catalase, glutathione peroxidase) can assess oxidative stress modulation. It may also be informative to include circadian rhythm markers. Dim light melatonin onset (DLMO), salivary cortisol awakening response, and core body temperature rhythms provide objective measures of circadian phase and amplitude. Actigraphy with circadian rhythm analysis (cosinor analysis, non-parametric circadian rhythm analysis) can quantify rest-activity patterns and detect improvements in circadian robustness.
Finally, statistical mediation analyses can test whether biomarker changes (e.g., increased BHB, reduced p-tau217, improved HOMA-IR) mediate the relationship between TRE adherence and cognitive outcomes. 65 This will elucidate mechanisms and identify predictors of treatment response.
Safety considerations in older adults with MCI/AD
While TRE holds promises for MCI/AD, safety concerns unique to older adults with cognitive impairment may include unintentional weight loss, sarcopenia, worsening frailty, and metabolic disturbances. TRE frequently results in spontaneous calorie reduction averaging 167–300 kcal/day even without instructions to follow a CR regimen. While this modest deficit may benefit individuals who are overweight, it may pose a risk for older adults with MCI who may already experience age-related sarcopenia and unintentional weight loss. Loss of lean muscle mass is a critical concern, as sarcopenia is associated with increased frailty, falls, disability, and mortality in older adults.66–69 To mitigate sarcopenia risk, TRE interventions in MCI/ AD populations may need to incorporate dietary counseling and adherence to physical activity guidelines. For example, the Barrow clinical trial has a registered dietitian who contacts study participants on a weekly basis and provides dietary counseling on balanced nutrition, including sufficient calorie and protein intake during the eating window. 70 Additionally, the Barrow study measures sedentary and physical activity levels at baseline and at follow-up. 54
Overall, recommendations for TREAD trials include: (1) Ideally, monitor body composition using DXA or bioelectrical impedance at baseline and at the end of the intervention to track lean body mass changes; (2) provide individualized dietary counseling to ensure protein intake of ≥ 1.2 g/kg/day is distributed across meals within the eating window; (3) recommend physical activity, including resistance training; (4) establish stopping rules for participants who experience excessive weight loss or significant lean mass reduction; and (5) consider excluding individuals with pre-existing frailty (e.g., FRAIL scale score ≥ 3) or BMI < 20 kg/m2 who may be at highest risk.
Of note, there may be interactions between hypoglycemia and medications. Older adults with diabetes taking insulin or sulfonylureas face increased hypoglycemia risk during prolonged fasting periods. Therefore, many TRE studies exclude patients with type 1 diabetes and insulin-requiring patients with type 2 diabetes. 54 Additional considerations for eligibility are the use of GLP-1 receptor agonist medications, which can exacerbate hypoglycemia in TRE and may independently influence cognitive outcomes and confound the effects of TRE in MCI and AD.71,72 Antihypertensive medications may require dose adjustment as TRE can reduce blood pressure, potentially causing orthostatic hypotension and falls. Close medical supervision with medication adjustment protocols is essential, including glucose monitoring for participants with diabetes and regular blood pressure assessments.
Furthermore, individuals with MCI may struggle with the cognitive demands of tracking eating windows, particularly if protocols are complex or frequently modified. A pilot study 49 emphasized that clear, simple instructions and regular support are critical for adherence in older adults with memory decline. Strategies to enhance feasibility include: (1) using visual aids, alarms, and smartphone apps with reminders; (2) involving caregivers or family members in intervention delivery; (3) using simpler protocols (e.g., 12:12 or 14:10 rather than 16:8 or 18:6); and (4) providing regular check-ins via phone or telehealth. In addition, baseline malnutrition screening using validated tools (e.g., Mini Nutritional Assessment) is important, with ongoing monitoring throughout the intervention. Participants at risk of malnutrition should either be excluded or receive intensive nutritional counseling and supplementation.
In summary, TRE in older adults with MCI/ AD requires careful safety monitoring, individualized dietary counseling, medication management, and psychosocial support to maximize benefits while minimizing risks. Future trials should thus prioritize comprehensive safety protocols and transparent reporting of adverse events.
Individual variability and need for personalized TRE approaches
Substantial inter-individual variability in metabolic, genetic, behavioral, and circadian factors may determine TRE response and tolerability. For example, the APOE ε4 allele is the strongest genetic risk factor for late-onset AD, affecting 40–65% of patients with AD. APOE ε4 carriers exhibit impaired brain glucose metabolism,14,73 altered lipid metabolism, and heightened neuroinflammation compared to non-carriers. Whether APOE ε4 carriers respond differently to TRE-induced ketogenesis compared to non-carriers remains unknown. Stratification by APOE genotype in TRE trials could reveal whether personalized fasting durations for APOE ε4 carriers is beneficial.
Individuals also vary in their circadian preference (chronotype), ranging from “early risers” to “night owls.” Imposing a rigid early TRE window (e.g., 7 AM-3 PM) on an evening chronotype individual may reduce adherence and potentially worsen circadian misalignment. Conversely, aligning TRE windows with individual chronotype preferences may enhance adherence and metabolic benefits. Therefore, assessment of chronotype using validated questionnaires (e.g., Munich ChronoType Questionnaire, Morningness-Eveningness Questionnaire) may provide insight into whether chronotype-informed TRE improves outcomes.
In addition, TRE's metabolic benefits may differ based on baseline metabolic health. Individuals with obesity and insulin resistance may experience greater improvements in glucose metabolism and weight loss compared to lean, metabolically healthy individuals. However, lean older adults with MCI may still benefit from TRE's effects, such as ketogenesis, autophagy, and circadian entrainment, even without weight loss.
Emerging evidence also suggests sex-specific responses to TRE and ketogenic interventions in APOE ε4 models, with female APOE ε4 mice showing greater cognitive improvements, enhanced synaptic plasticity, and more robust microbiome-metabolite axis modulation compared to males following ketogenic diet. 74 However, most TRE studies in older adults have been underpowered to detect sex differences or have not reported sex-stratified analyses, highlighting the need for adequately powered trials with pre-specified sex-stratified analyses and exploration of potential mechanisms (e.g., hormonal regulation, body composition differences, circadian rhythm sex differences).74–77
Furthermore, polypharmacy is common in older adults, with medications potentially interacting with TRE's metabolic effects. For example, metformin enhances insulin sensitivity and may synergize with TRE, whereas corticosteroids impair glucose metabolism and could blunt TRE benefits. Detailed medication tracking and exploratory analyses of medication-TRE interactions will inform clinical implementation.
In conclusion, future research will need to include precision nutrition approaches that tailor TRE based on one or more of the following factors: (1) APOE genotype, (2) chronotype, (3) baseline metabolic health, (4) sex, (5) medication use, (6) frailty status, and (7) psychosocial factors (e.g., caregiver support). Adaptive trial designs (e.g., SMART trials - Sequential Multiple Assignment Randomized Trials) could systematically test personalized TRE algorithms to optimize individual outcomes.
Limitations of current studies and future directions
There is a lack of long-term TRE human studies, with most of the current evidence stemming from preclinical models or relatively short-term human studies. Thus, there is a need for large-scale, long-term randomized controlled trials. Variability in TRE regimen and outcome measures limit comparability. Few studies adjusted for critical confounders, such as physical activity or diet quality, which may explain conflicting findings. 47 Preclinical studies, particularly in primates, often lack AD-specific cognitive endpoints. Human studies also vary considerably in cognitive outcome measures, leading to heterogeneity in results. Studies identified in ClinicalTrials.gov underscore the need for standardized protocols and advanced biomarkers that are critical for validating TRE's role in MCI.
Discussion
TRE is a promising potential non-pharmacological intervention for MCI.17,55,56 Human studies indicate that TRE elicits resistance to AD through direct and indirect mechanisms (Figure 2). The direct pathway involves TRE leading to a metabolic switch from glucose to ketone body production that induces a cascade of events that may enhance plasticity in the hippocampus. The indirect pathway decreases insulin resistance, which is the underlying disease of various syndromes that are associated with cognitive impairment, including type 2 diabetes, neuroinflammation, and stroke.18,45 Pilot studies on prolonged nightly fasting in adults with cognitive decline 48 and qualitative research 49 indicate TRE's feasibility and potential cognitive benefits. TRE mitigates cardiometabolic risks and promotes neuroprotection, thus extending its relevance to healthy aging.41,78
TRE's mechanisms, including ketogenesis, BDNF upregulation, decreased neuroinflammation, and autophagy, align with AD pathology targets.17,19,20 TRE's weight-independent benefits enhance its applicability for patients with MCI.21,79 Challenges include adherence barriers and cultural resistance, which may be addressable with personalized coaching and tracking. One study's contradictory findings 47 underscore the need for standardized protocols to control for confounders such as physical activity. Ongoing trials will likely provide critical data on feasibility and long-term outcomes. Study participants reported that clear instructions and guidance were key for success in prolonged nightly fasting (14-h fasting, 10-h eating window). 49
The limited human data on TRE and cognition show inconsistent findings that warrant careful interpretation. While Currenti and colleagues 46 found lower cognitive impairment prevalence in Italian older adults with ≤10-h eating windows, Li and colleagues 47 reported the opposite, i.e., they observed higher cognitive impairment frequency in Chinese older adults with ≤10-h eating windows compared to those with >10-h windows.
Several methodological factors may explain this discrepancy. First, both studies used cross-sectional designs, precluding causal inference and raising concerns about reverse causality; 80 individuals with existing cognitive impairment may alter their eating patterns due to appetite dysregulation, meal preparation difficulties, or caregiver influences. Second, neither study verified eating windows through objective methods such as continuous glucose monitoring; and self-reported eating duration is prone to recall bias, particularly among cognitively impaired populations. Third, confounding by diet quality, total caloric intake, physical activity, and socioeconomic status was not controlled in the analyses. Italian and Chinese populations also differ substantially in dietary patterns, genetic backgrounds, and lifestyle factors, which could influence TRE's effects. Fourth, measurement bias may exist since Currenti and colleagues used screening instruments of unspecified type, while Li and colleagues used Mini-Mental State Examination, which is a bedside cognitive screening instrument primarily used to determine whether a patient needs referral for neuropsychological testing. The clinical examination and neuropsychological assessment are then used to classify a person as cognitively impaired, e.g., as having MCI or not.
In addition to measurement bias, Li and colleagues’ finding of worse performance specifically on Mini-Mental State Examination orientation and attention/ calculation subscales may reflect acute effects of fasting on attention rather than long-term cognitive decline. This raises the possibility that testing cognitively impaired individuals during their fasting window could artificially worsen performance due to hunger, hypoglycemia, or reduced alertness.
Limitations of current research include a scarcity of long-term human clinical trials, variability in TRE regimens, and inconsistent outcome measures. Future research should thus explore optimal fasting durations while also considering dietary quality and timing. Incorporation of advanced AD biomarkers (e.g., plasma p-tau 217) 81 will also be necessary to establish an optimal TRE protocol for individuals with MCI due to AD. Multicenter, double-blind randomized controlled trials in diverse MCI populations, using qualitative insights for adherence 81 and comprehensive neuropsychological evaluation to measure cognitive outcomes, are needed to establish evidence-based strategies to delay the progression of MCI.
Footnotes
Acknowledgements
The authors would like to thank Cassie Todd of Neuroscience Publications at Barrow Neurological Institute for her assistance with the figures of this paper.
Author contribution(s)
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding was provided by Barrow Neurological Foundation.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Yonas Geda is an Editorial Board Member of this journal but was not involved in the peer-review process of this article nor had access to any information regarding its peer-review.