chatGPT(5.5paid):
The uploaded paper is Salgueiro et al. (2026), “Ketogenic diet as a therapeutic strategy for neurodegenerative diseases: from mechanisms to translational challenges,” published in Translational Neurodegeneration. It is a review rather than a new experimental study, covering Alzheimer’s disease (AD), Parkinson’s disease (PD), Huntington’s disease (HD), amyotrophic lateral sclerosis (ALS) and multiple sclerosis (MS). The paper is open access under CC BY 4.0.
1. Summary
Central argument
The authors argue that a ketogenic diet (KD) is potentially useful in neurodegenerative disease not simply because it supplies an alternative fuel, but because ketosis simultaneously alters several processes shared across neurodegenerative disorders:
KD → ketone bodies, particularly β-hydroxybutyrate (BHB) →
- alternative neuronal energy substrate
- improved mitochondrial energetics
- reduced oxidative stress
- suppression of neuroinflammation
- altered histone modifications/gene regulation
- increased autophagy and protein clearance
- modulation of the gut microbiome
- potentially improved circadian regulation.
Their conceptual attraction to KD is therefore that it acts upstream of several shared aspects of neurodegeneration rather than targeting one disease-specific protein.
A. Bioenergetics
Under carbohydrate restriction, liver fatty-acid oxidation produces ketone bodies. Brain cells can convert BHB/acetoacetate into acetyl-CoA, feeding the TCA cycle without requiring glycolysis or pyruvate dehydrogenase.
This potentially matters particularly in disorders such as AD where cerebral glucose metabolism is impaired. The paper consequently views ketosis partly as a metabolic bypass of impaired glucose utilisation.
It also proposes that reduced glucose itself may have benefits by reducing glycolytic NADH generation, mitochondrial ROS and glycation/AGE production.
B. BHB is more than a fuel
This is one of the strongest mechanistic sections of the review. The authors distinguish BHB from acetoacetate rather than treating all ketones as interchangeable.
BHB can:
- inhibit class-I HDACs
- increase histone acetylation
- activate transcriptional programmes including PGC-1α
- act through HCA2/GPR109A
- inhibit the NLRP3 inflammasome
- influence NF-κB
- cause lysine β-hydroxybutyrylation.
The authors also highlight emerging evidence that β-hydroxybutyrylation occurs on metabolic enzymes such as citrate synthase and succinyl-CoA ligase and may increase their activity and ATP production.
That is conceptually important: ketosis is being presented as both a metabolic state and an epigenetic/post-translational signalling state.
C. Mitochondria and oxidative stress
The paper links ketosis to:
- increased mitochondrial biogenesis
- increased ATP generation
- increased NADH oxidation
- reduced electron leakage/ROS
- increased Nrf2 signalling
- increased glutathione peroxidase and HO-1
- possible UCP2-mediated reduction in mitochondrial ROS.
BHB is also linked to SIRT1 activation and downstream FOXO antioxidant pathways.
D. Neuroinflammation
BHB inhibition of NLRP3 is a major theme. The review also describes:
- reduced NF-κB activity
- reduced IL-1β, IL-6 and TNF-α
- altered macrophage/microglial phenotypes
- HCA2 signalling
- increased BDNF and possibly other neurotrophic factors.
This provides a plausible bridge between energy metabolism and inflammatory neurodegeneration.
E. Autophagy/proteostasis
The authors suggest ketosis stimulates autophagy through an interconnected:
SIRT1 → AMPK → ↓mTORC1
axis, with possible involvement of HIF-1α.
This could improve clearance of proteins such as Aβ, tau and α-synuclein. They additionally note SIRT1-mediated tau deacetylation as a possible route to improved tau clearance.
F. Circadian rhythm
A relatively unusual element of the review is the inclusion of circadian disruption as a shared component of neurodegeneration.
The authors connect BHB/KD with SIRT1 and clock-gene regulation, pointing to experimental evidence that ketone supplementation can normalise locomotor rhythmicity and clock-gene expression. They correctly qualify this by noting that some of this evidence derives from ketone supplementation rather than classical KD.
G. Gut microbiome
KD is proposed to change microbial populations, intestinal inflammation and gut permeability, potentially reducing inflammatory signals reaching the brain.
This section is interesting but much less mechanistically settled than the bioenergetic/BHB sections. The authors themselves acknowledge that changes in SCFAs differ according to dietary composition, disease and degree of ketosis.
2. Disease-specific evidence
The overall evidence hierarchy differs enormously between diseases.
| Disease | What the paper finds | My assessment of strength |
|---|---|---|
| AD | Considerable animal data; several small human interventions | Most convincing neurological case |
| PD | Good animal data; several small pilot trials | Suggestive |
| MS | Animal studies plus some relatively substantial clinical work | Interesting, especially for symptoms/inflammatory markers |
| HD | Animal data plus essentially a human case report | Very preliminary |
| ALS | SOD1 mouse studies plus essentially a human case report | Very preliminary |
Alzheimer’s disease
Animal studies report lower amyloid deposition, reduced microglial activation, improved cognition, increased ATP and restored long-term potentiation.
Human evidence includes an MCT ketogenic formulation in 20 AD subjects and a six-month MCT drink intervention in people with MCI, with improvements in several cognitive measures.
A 2021 randomized crossover trial also found sustained ketosis with improvements in daily function and quality of life.
This is encouraging, although the evidence is much stronger for short-term cognitive/metabolic effects than for slowing AD pathology or disease progression.
Parkinson’s disease
Animal studies consistently show protection of dopaminergic neurons, increased ATP and reduced oxidative/inflammatory damage.
Human studies are small:
- pilot RCT: 47 enrolled, 38 completed; improvement particularly in non-motor UPDRS measures;
- seven-person trial: cognitive improvements but no motor effect;
- seven-person longitudinal pilot: multiple reported improvements;
- 16-person MCT-KD feasibility trial: improved energy/fatigue but no significant between-group UPDRS difference.
Thus the clinical signal is intriguing, but nowhere near sufficient to demonstrate disease modification.
Huntington’s disease
The animal evidence is mixed.
A particularly interesting experiment found D-BHB extended lifespan in an HD mouse model by almost 30% and improved motor deficits. But another KD study improved weight loss/behaviour without preventing brain atrophy or improving rotarod performance.
The striking human findings — 52% improvement in motor symptoms, for example — come from one 41-year-old patient.
That should therefore be regarded as hypothesis-generating rather than efficacy evidence.
ALS
SOD1-G93A mice showed preserved motor neurons/function with KD. Caprylic triglyceride improved mitochondrial oxygen consumption and motor-neuron preservation but did not extend survival.
The human evidence cited consists of a single 64-year-old man followed for 18 months, with reported improvements/stabilisation in several measures.
Again, this is far too little evidence to infer efficacy.
Multiple sclerosis
MS arguably has some of the more interesting human evidence in the review.
Clinical studies include:
- a six-month modified Atkins KD study
- randomized studies measuring inflammatory gene expression
- reduced serum neurofilament light chain
- an 18-month randomized dietary study.
Reported outcomes include improved fatigue, depression and quality of life and changes in leptin/adiponectin.
This supports metabolic and symptomatic effects, but it remains unclear whether KD alters the long-term rate of neurological disability accumulation.
3. What is novel about this review?
There isn’t really a single new biological discovery here because it is a review. Its novelty is principally conceptual synthesis.
I would identify four particularly useful aspects.
1. Ketosis is treated as a signalling intervention, not just an alternative fuel
Older accounts often boil down to:
impaired glucose metabolism → ketones provide another fuel.
This review gives substantial weight instead to:
BHB → HDAC inhibition / β-hydroxybutyrylation / HCA2 / NLRP3 / SIRT1 / Nrf2 / mTOR/autophagy
That makes the proposed mechanism much broader.
2. It integrates metabolism with epigenetic regulation
The discussion of BHB-mediated HDAC inhibition and β-hydroxybutyrylation is particularly relevant. BHB is proposed to increase histone acetylation and activate mitochondrial/antioxidant transcriptional programmes.
The review therefore implicitly links:
whole-body metabolism → metabolite concentrations → chromatin/PTMs → gene expression → neuronal phenotype.
That is probably one of its more conceptually valuable contributions.
3. Circadian regulation is incorporated into the KD-neurodegeneration model
Many KD reviews discuss mitochondria, inflammation and autophagy. Explicitly integrating SIRT1/circadian biology with these mechanisms gives this review a somewhat wider scope.
4. It puts mechanistic and translational evidence side-by-side
The disease tables are useful because they make the gap between impressive animal mechanistic data and weak human efficacy evidence relatively visible.
The authors themselves conclude that human studies are generally short, small and sometimes self-reported.
4. Critique
A. The largest problem: it sometimes conflates ketogenic diet with ketones
Throughout the review, evidence comes variously from:
- ketogenic diets
- modified ketogenic diets
- MCT diets
- MCT drinks
- BHB supplementation
- D-BHB infusion
- ketone esters.
These are not equivalent interventions.
For example:
BHB infusion → biological effect
does not establish that:
high-fat ketogenic diet → same effect.
And conversely, benefits from KD could derive from:
- lower glucose
- lower insulin
- altered fatty acids
- altered protein/amino-acid intake
- weight loss
- caloric changes
- microbiome alteration
rather than BHB itself.
This is perhaps the most important conceptual limitation.
B. Mechanism is much stronger than clinical causality
Figure 2 effectively presents a network:
KD → BHB → ↓ROS, ↑Nrf2, ↑ATP, ↓NF-κB/NLRP3, ↑autophagy, altered microbiome.
It is a useful diagram, but visually it risks making tentative relationships look like a single established causal pathway.
In reality, different arrows derive from:
- different species
- different tissues
- different disease models
- different ketogenic interventions
- different ketone concentrations.
There is no single human experiment demonstrating the complete causal chain.
C. The clinical evidence is very underpowered
The review is strongest when it admits this.
Examples include:
- PD trials with 7 subjects
- HD evidence dominated by one case
- ALS evidence dominated by one case
- modest-sized AD trials
- relatively short interventions.
Even an apparently dramatic result such as the HD case’s 52% motor improvement cannot establish efficacy without controls.
The title and overall tone are therefore somewhat more confident than the human evidence warrants.
D. Disease modification and symptomatic improvement are not adequately separated
This distinction is critical.
Improving:
- fatigue
- mood
- body weight
- cognition
- quality of life
- inflammatory biomarkers
does not necessarily mean slowing neuronal death.
The paper sometimes moves rather easily from evidence of symptomatic/metabolic benefit to language about neuroprotection or slowing disease progression.
The correct question for a disease-modifying treatment is ultimately something like:
Does long-term ketosis alter the slope of neurodegeneration compared with an appropriate control?
The review supplies little direct human evidence for that proposition.
E. Its literature-search methodology isn’t sufficiently rigorous for a systematic review
The authors say the disease tables were derived from a structured PubMed search, focusing mainly on the preceding 15 years, selecting peer-reviewed original preclinical and clinical work with defined ketogenic interventions and measurable outcomes.
That is useful, but this is not a full systematic review methodology.
There is no clear presentation of:
- complete reproducible search strings
- multiple-database search strategy
- PRISMA flow
- formal risk-of-bias assessment
- grading of evidence quality
- quantitative meta-analysis.
Consequently, positive-study selection and narrative weighting remain possible.
F. Preclinical evidence is especially vulnerable to generalisation problems
Many animal experiments use:
- genetically extreme models
- toxin-induced PD models such as MPTP/6-OHDA
- predominantly or exclusively male mice
- interventions initiated at disease stages unlike typical human treatment.
The PD table, for example, contains several all-male animal experiments.
Protection against an acute mitochondrial toxin is biologically interesting but doesn’t necessarily predict effectiveness against decades-long sporadic human PD.
G. The review could distinguish energy supply from signalling much more experimentally
A particularly useful next-generation experiment would compare:
- ordinary diet,
- KD,
- exogenous BHB with normal carbohydrate,
- carbohydrate restriction without substantial ketosis,
- isocaloric controls.
That would allow separation of:
ketone fuel effects vs BHB signalling vs glucose/insulin reduction vs dietary-fat effects.
The current literature reviewed here generally cannot resolve those components.
5. Safety and practicality
The authors are appropriately cautious here.
Reported problems include:
- gastrointestinal symptoms
- dehydration
- “keto flu”
- constipation
- micronutrient deficiencies
- unintended weight loss
- transient hyperlipidaemia.
Potential longer-term concerns include adverse lipid profiles, reduced bone mineral density, nephrolithiasis, anaemia, neuropathy and hepatic problems.
This is particularly relevant in neurodegenerative disease because patients may already have difficulty with eating, swallowing, food preparation and maintaining body weight.
Adherence is consequently not a minor inconvenience: it may determine whether KD is clinically viable at all. The paper reports problems from cost, family/social support, holiday meals and exclusion of favoured foods.
6. An especially interesting point: acetyl-CoA and histone acetylation
There is a mechanistic issue which the paper touches but does not fully develop.
It correctly notes that BHB inhibits class-I HDACs, increasing histone acetylation.
It also describes ketone metabolism:
BHB → AcAc → acetoacetyl-CoA → acetyl-CoA → TCA cycle.
This creates two potentially independent mechanisms affecting acetylation:
BHB
→ HDAC inhibition
→ less histone deacetylation
and potentially
ketone oxidation → mitochondrial acetyl-CoA → citrate → cytosolic/nuclear acetyl-CoA
→ HAT substrate availability
→ more histone acetylation.
The review concentrates mainly on the first mechanism.
It does not seriously examine whether ketosis increases the nuclear acetyl-CoA/citrate supply available for histone acetylation. That distinction is important because histone acetylation depends on both sides of the reaction:
[
\text{Histone acetylation}
\sim
\frac{\text{HAT activity}\times[\text{acetyl-CoA}]}{\text{HDAC activity}}
]
Conceptually, therefore, ketogenic metabolism might affect histone acetylation through both acetyl donor availability and removal rate, but the paper mostly discusses the removal-rate/HDAC side.
That is, in my view, a significant mechanistic gap in what is otherwise a broad metabolism-to-epigenetics synthesis.
Overall assessment
I would rate the paper as a good and unusually broad mechanistic review, but not strong evidence that ketogenic diets are established disease-modifying treatments.
Its strongest contribution is the integrated picture:
ketosis/BHB → mitochondrial metabolism + redox control + inflammatory signalling + epigenetic regulation + autophagy + microbiome/circadian effects → potentially greater neuronal resilience.
The weakest part is the jump from that highly plausible mechanistic biology to therapeutic claims. The human data remain too small, heterogeneous and short-term, with HD and ALS in particular resting almost entirely on case-level clinical evidence. The authors themselves ultimately frame KD as a complementary metabolic intervention, rather than a cure, and call for personalised trials, improved monitoring and investigation of early preventive use.
A particularly worthwhile line to pursue from this paper is whether BHB-driven HDAC inhibition and ketone-derived acetyl-CoA/citrate availability act synergistically on histone acetylation. The review gives enough pieces of that pathway to make the hypothesis plausible, but does not connect them explicitly.