Skip to content
The Longevity Archive
browse 192

The biology of fasting

Ketogenesis and beta-hydroxybutyrate

Also called: Ketosis, BHB, Ketone bodies

Written by Reviewed Sep 2026

The liver breaks fat down faster than it can burn the pieces, condenses the surplus into small water-soluble molecules called ketone bodies, and exports them as a fuel the brain can use when glucose is scarce.

What it is

There are three ketone bodies. Acetoacetate is the one the liver actually makes. Beta-hydroxybutyrate, usually shortened to BHB, is acetoacetate with two hydrogens added, and it is the one that dominates in blood during a fast and the one meters measure. Acetone is what acetoacetate turns into when it loses a carbon dioxide on its own, and because acetone is volatile it leaves on the breath, which is why deep ketosis is smellable. Breath acetone is a real marker but only a loosely quantitative one.

The liver makes all of this and can use none of it. Hepatocytes do not express SCOT, the enzyme every other tissue needs to feed a ketone body back into the citric acid cycle. That single omission is what turns ketogenesis into a genuine export pathway rather than a pointless internal loop: the liver manufactures a fuel it is constitutionally unable to consume, and ships every molecule out.

BHB is not only a fuel. It also inhibits a class of enzymes that control gene expression, activates a receptor on fat cells, and is itself used to write a chemical mark onto histones. Those three activities are chemically unrelated to each other and none of them requires BHB to be burned. Where they have been shown, and where they have not, is the whole point of the two sections below.

The machinery

Step one is release. Insulin normally suppresses fat breakdown, so insulin falling is permissive rather than active. Adipose triglyceride lipase, hormone sensitive lipase and monoacylglycerol lipase strip the three fatty acids off a stored triacylglycerol and release them along with one glycerol. The fatty acids travel bound to albumin. The glycerol travels free and becomes a substrate for making new glucose.

Step two is the gate. Fatty acids taken up by the liver are activated to fatty acyl-CoA, which cannot cross the inner mitochondrial membrane on its own. Carnitine palmitoyltransferase 1, CPT1, swaps the CoA for carnitine so the molecule can be carried across. CPT1 is inhibited by malonyl-CoA, and malonyl-CoA is made by acetyl-CoA carboxylase, which insulin stimulates and AMPK shuts off. So when insulin falls and AMPK activity rises, malonyl-CoA falls, CPT1 opens, and fat floods into the mitochondrial matrix. This one regulated step is where the hormonal state of a fast becomes a biochemical flux.

Step three is beta-oxidation, which chops two carbons at a time off the fatty acid chain as acetyl-CoA. A sixteen-carbon palmitate yields eight of them.

Step four is the bottleneck. Acetyl-CoA normally condenses with oxaloacetate to enter the citric acid cycle, but in a fasting liver oxaloacetate is being drawn off to make glucose, and the heavy load of reduced cofactors from beta-oxidation pushes the chemistry further away from regenerating it. The liver ends up with more acetyl-CoA than it has partners for.

Step five is ketogenesis proper. Two acetyl-CoA condense to acetoacetyl-CoA, a third is added by HMG-CoA synthase 2 to give HMG-CoA, and HMG-CoA lyase cleaves that to acetoacetate. HMGCS2 is the rate controlling enzyme and it is the mitochondrial isoform, driven transcriptionally by PPAR alpha and FOXA2 in the fasted liver. Acetoacetate is then reduced to BHB by BDH1. Because that last reaction consumes NADH, the ratio of BHB to acetoacetate in blood is a direct readout of how reduced the liver's mitochondria are.

Step six is use. Ketone bodies need no carrier in plasma and cross the blood brain barrier on monocarboxylate transporters. In a tissue that expresses SCOT, BDH1 runs backwards to regenerate acetoacetate, SCOT hands it a CoA from succinyl-CoA, and thiolase splits the result into two acetyl-CoA for the citric acid cycle.

What switches it on

Falling insulin, rising glucagon, and rising AMPK activity, all converging on malonyl-CoA and therefore on the CPT1 gate. There is also a brake built in: BHB is the endogenous ligand of the receptor HCAR2, also called GPR109A, which sits on fat cells and restrains further fatty acid release. In a person with intact insulin signalling, ketone production limits itself.

On the timing
This is one of the few places in fasting biology where a real human time course exists, and it does not look like the popular version. Beta-hydroxybutyrate was measured every six hours through the standard 72 hour diagnostic fast in 34 adults. The median change from 12 to 18 hours was zero. Essentially nothing happens to ketones in the first 18 hours. The steepest climb is between 18 and 36 hours, a median rise of 333%, then 210% from 36 to 54 hours and 167% from 54 to 72 hours. That is a percentage trajectory, not a set of absolute concentrations. No single human study publishes absolute millimolar BHB at 12, 24, 36, 48 and 72 hours, which is why the tidy concentration ladder that circulates is not reproduced on this page.

What has been measured

In people

Human ketone measurement is unusually good, because ketone bodies are in blood where they can be sampled repeatedly, and because magnetic resonance spectroscopy and cerebral catheterisation have both been used to measure them inside the living brain. The fuel role is directly measured in people. The signalling roles are not.

  • The 72 hour trajectory, 34 adults, sampled every 6 hours through the standard diagnostic fast used in endocrinology to rule out insulinoma. Median change was 0% from 12 to 18 hours, +333% from 18 to 36, +210% from 36 to 54, and +167% from 54 to 72. Using the existing criterion of BHB above 2.7 mmol/L, 74% of the people whose fast was negative crossed that level before the 72 hour mark (Journal of Clinical Endocrinology and Metabolism, 2005).
  • Absolute concentrations where they were published. Total ketone bodies, meaning acetoacetate plus BHB together, were 0.20 mM in overnight fasted subjects and averaged 5.7 mM at rest in subjects fasted 3 to 5 days (American Journal of Physiology, 1983). In 13 volunteers fasting 21 days, blood ketones went from 0.1 plus or minus 0.04 to 6.61 plus or minus 1.25 mmol/L (Scientific Reports, 2024). Total ketone bodies and BHB alone are not interchangeable numbers, and the ratio between the two shifts as a fast lengthens.
  • Inside the brain, by 4-tesla magnetic resonance spectroscopy of the occipital lobe. Brain BHB rose from 0.05 plus or minus 0.05 mmol/L non-fasted to 0.60 plus or minus 0.26 after two days and 0.98 plus or minus 0.16 after three. Brain lactate rose from 0.69 plus or minus 0.17 to 1.47 plus or minus 0.22 mmol/L. Plasma and brain BHB correlated at r = 0.86 with a brain to plasma slope of 0.26 (Journal of Cerebral Blood Flow and Metabolism, 2000).
  • Brain fuel accounting at 3.5 days, 9 healthy volunteers, dynamic PET plus an independent Fick determination. Cortical grey matter glucose metabolism fell 26%, white matter fell 27%, the fall was uniform across regions, and cerebral blood flow did not change. Net brain BHB uptake rose 13-fold. At 3.5 days ketones supply roughly one quarter of the human brain's energy requirement (Journal of Cerebral Blood Flow and Metabolism, 1994).
  • How ketones get into the brain, same 9 subjects, double indicator method. The transport capacity for glucose rose 55%, more than the lower plasma glucose alone would predict, indicating modest transporter upregulation. The transport capacity for BHB did not increase at all. BHB influx rose more than tenfold purely because blood BHB rose. Ketone entry into the human brain is governed by concentration, not by transporter induction (American Journal of Physiology, 1995).
  • The founding observation, and it is human. Cerebral vessels were catheterised in three obese patients through 5 to 6 weeks of starvation, and beta-hydroxybutyrate and acetoacetate were shown to have replaced glucose as the predominant fuel of brain metabolism (Journal of Clinical Investigation, 1967).
  • Production is self-limiting. Ketone body production rate begins to plateau after about five days, through ketones restraining their own supply by antilipolytic and insulinotropic effects combined with progressive saturation of muscle ketone uptake, which shunts what is made preferentially to the brain. Urinary ketone excretion stays below 10% of total turnover under physiological conditions (Diabetes/Metabolism Reviews, 1989).

In other species and in cell culture

Every one of BHB's non-fuel activities was characterised in mouse tissue, in mouse-derived cells, or in cultured cell lines. They are real, specific and well controlled experiments. None of them has been repeated as a measurement in human tissue at a concentration a human fast produces.

  • Histone deacetylase inhibition, mouse tissue and cultured cells. BHB is an endogenous and specific inhibitor of class I histone deacetylases. Giving mice exogenous BHB, and separately fasting them or calorie restricting them, raised global histone acetylation in tissue. The transcriptional consequence tracked oxidative stress resistance: Foxo3a and Mt2 were induced, BHB raised histone acetylation at both promoters, both genes were activated by selective depletion of HDAC1 and HDAC2, and BHB treated mice were substantially protected against oxidative stress (Science, 2013).
  • The HCAR2 receptor. D-beta-hydroxybutyrate was identified as the endogenous ligand of the receptor previously known only as the niacin receptor. It activated mouse PUMA-G and, in a receptor assay, the human orthologue HM74a, at concentrations observed in serum during fasting, and it inhibited lipolysis in mouse adipocytes in a PUMA-G dependent manner (Journal of Biological Chemistry, 2005). The lipolysis experiment was in mouse fat cells; the human protein was tested in a dish.
  • Histone beta-hydroxybutyrylation, cultured cells and mouse liver. BHB is the source of a distinct covalent histone mark. Forty-four lysine beta-hydroxybutyrylation sites were identified, a count comparable to the known histone acetylation sites. The marks were dramatically induced by elevated BHB in cultured cells and in the livers of mice under prolonged fasting or streptozotocin induced diabetic ketoacidosis, were enriched at active gene promoters, and the rise in H3K9bhb during starvation was associated with genes upregulated in starvation responsive metabolic pathways (Molecular Cell, 2016).
  • NLRP3 inflammasome suppression, mouse and cell systems. BHB, but not acetoacetate and not the structurally related short chain fatty acids butyrate or acetate, suppressed activation of the NLRP3 inflammasome by urate crystals, ATP and lipotoxic fatty acids, by preventing potassium efflux and reducing ASC oligomerisation. The effect was independent of chirality, and independent of AMPK, reactive oxygen species, autophagy and glycolytic inhibition, and it occurred without BHB being oxidised at all (Nature Medicine, 2015).

Why it matters

The fuel story is settled and it is human. A brain that stores no glycogen and cannot burn long chain fatty acids has a serious problem when glucose runs short, and ketone bodies are the solution: small, water soluble, transported without a carrier protein, and oxidised by neurons. Three independent human methods, arteriovenous catheterisation, positron emission tomography and magnetic resonance spectroscopy, all show the same switch happening.

The signalling story is the interesting frontier and it is not yet a human story. BHB inhibiting histone deacetylases, activating HCAR2 and writing its own histone mark are the reasons ketosis is discussed as something more than an emergency fuel. Each of those is a real mechanism demonstrated in a real experiment. Each was demonstrated in mouse tissue or cultured cells. Whether human tissue reaches the concentrations, and holds them for the durations, that those effects require has not been measured in a person.

A claim you will see repeated

The concentration ladder. Almost every fasting page gives specific millimolar thresholds by hour: roughly 0.1 to 0.3 after an overnight fast, under 1 at 24 hours, 1 to 3 at two to three days, 5 to 7 over weeks. Those bounds are directionally consistent with the reviews, but the research behind this page could not trace a single one of them to a primary human measurement, so none of them are published here. What is published instead is the percentage trajectory that was actually measured, the total ketone anchors at overnight and at 3 to 5 days, the 21 day figure, and the brain concentrations. Separately, ketosis and autophagy are routinely treated as the same event on the same clock. They are not. Ketogenesis is a biosynthetic pathway in liver mitochondria. Autophagy is a degradative pathway at the lysosome. They share upstream regulators and they rise over the same hours, but a ketone reading on a meter is not a measurement of autophagosome flux in any tissue.

Citations

  1. Human2005
    Increasing serum betahydroxybutyrate concentrations during the 72-hour fast: evidence against hyperinsulinemic hypoglycemia

    Journal of Clinical Endocrinology and Metabolism

    Beta-hydroxybutyrate measured every 6 hours through the standard 72 hour diagnostic fast in 34 adults whose fast was negative for insulinoma, alongside 21 with surgically confirmed insulinoma. Median change 0% from 12 to 18 hours, +333% from 18 to 36, +210% from 36 to 54, +167% from 54 to 72. 74% of those with a negative fast exceeded 2.7 mmol/L before 72 hours.

  2. Human1983
    Ketone body turnover during and after exercise in overnight-fasted and starved humans

    American Journal of Physiology

    Total ketone bodies, acetoacetate plus beta-hydroxybutyrate, were 0.20 mM in overnight fasted subjects and averaged 5.7 mM at rest in subjects fasted 3 to 5 days.

  3. Human2024
    Analysis of physiological and biochemical changes and metabolic shifts during 21-Day fasting hypometabolism

    Scientific Reports

    13 volunteers through a 34 day protocol including 21 days of complete fasting. Blood ketones rose from 0.1 plus or minus 0.04 to 6.61 plus or minus 1.25 mmol/L. Resting energy expenditure fell 20.3 plus or minus 11.13%, body weight fell 14.96 plus or minus 1.55%, and the respiratory quotient shifted toward fat.

  4. Human2000
    Human brain beta-hydroxybutyrate and lactate increase in fasting-induced ketosis

    Journal of Cerebral Blood Flow and Metabolism

    Beta-hydroxybutyrate measured directly in the occipital lobe of healthy adults by 4-tesla magnetic resonance spectroscopy. Brain BHB 0.05 plus or minus 0.05 mmol/L non-fasted, 0.60 plus or minus 0.26 after the second day, 0.98 plus or minus 0.16 after the third. Brain lactate rose from 0.69 plus or minus 0.17 to 1.47 plus or minus 0.22 mmol/L. Plasma and brain BHB correlated at r = 0.86, brain to plasma slope 0.26.

  5. Human1994
    Brain metabolism during short-term starvation in humans

    Journal of Cerebral Blood Flow and Metabolism

    9 healthy volunteers before and after 3.5 days of starvation. Cortical grey matter glucose metabolism fell 26% (0.294 to 0.217 micromol per gram per minute), white matter fell 27%, the fall was uniform across regions at 24 to 30%, and an independent Fick determination confirmed a 24% fall. Cerebral blood flow did not change. Global net brain uptake of beta-hydroxybutyrate rose 13-fold (0.012 to 0.155 micromol per gram per minute). At 3.5 days ketones account for approximately one quarter of cerebral energy requirements.

  6. Human1995
    Blood-brain barrier permeability of glucose and ketone bodies during short-term starvation in humans

    American Journal of Physiology

    Same 9 subjects, double indicator method. The permeability surface area product for glucose transport into brain rose 55%, more than the roughly 22% expected from the lower plasma glucose alone. The permeability surface area product for beta-hydroxybutyrate did not increase. BHB blood concentration and BHB influx into brain both rose more than tenfold, so ketone entry is governed by blood concentration rather than transporter induction.

  7. Human1967
    Brain metabolism during fasting

    Journal of Clinical Investigation

    Catheterisation of cerebral vessels in three obese patients undergoing 5 to 6 weeks of starvation demonstrated that beta-hydroxybutyrate and acetoacetate had replaced glucose as the predominant fuel for brain metabolism.

  8. Review1989
    Ketone body production and disposal: effects of fasting, diabetes, and exercise

    Diabetes/Metabolism Reviews

    Ketone body production rate begins to plateau after about five days of fasting, through ketone mediated restraint of lipolysis and insulin secretion combined with progressive saturation of muscle ketone uptake, which shunts ketones preferentially to the brain. Urinary ketone excretion remains below 10% of total turnover under physiological conditions.

  9. Animal2013
    Suppression of oxidative stress by beta-hydroxybutyrate, an endogenous histone deacetylase inhibitor

    Science

    Mouse tissue and cultured cells. BHB is an endogenous and specific inhibitor of class I histone deacetylases. Exogenous BHB, fasting and calorie restriction each raised global histone acetylation in mouse tissues. Foxo3a and Mt2 were induced, BHB raised histone acetylation at both promoters, both genes were activated by selective depletion of HDAC1 and HDAC2, and BHB treated mice were substantially protected against oxidative stress.

  10. Animal2005
    (D)-beta-Hydroxybutyrate inhibits adipocyte lipolysis via the nicotinic acid receptor PUMA-G

    Journal of Biological Chemistry

    D-beta-hydroxybutyrate identified as the endogenous ligand of the niacin receptor. It activated mouse PUMA-G and, in a receptor assay, the human orthologue HM74a, at concentrations observed in serum during fasting, and inhibited lipolysis in mouse adipocytes in a PUMA-G dependent manner. This is the negative feedback loop that makes ketone production self-limiting.

  11. Animal2016
    Metabolic Regulation of Gene Expression by Histone Lysine beta-Hydroxybutyrylation

    Molecular Cell

    Cultured cells and mouse liver. Lysine beta-hydroxybutyrylation identified and verified as a distinct histone modification, with 44 sites found, a count comparable to known histone acetylation sites. Marks were dramatically induced by elevated BHB in cultured cells and in the livers of mice under prolonged fasting or streptozotocin induced diabetic ketoacidosis, were enriched at active gene promoters, and the rise in H3K9bhb during starvation was associated with genes upregulated in starvation responsive pathways.

  12. Animal2015
    The ketone metabolite beta-hydroxybutyrate blocks NLRP3 inflammasome-mediated inflammatory disease

    Nature Medicine

    Mouse and cell systems. BHB, but not acetoacetate and not butyrate or acetate, suppressed NLRP3 inflammasome activation in response to urate crystals, ATP and lipotoxic fatty acids, by preventing potassium efflux and reducing ASC oligomerisation and speck formation. The effect was independent of chirality, and of AMPK, reactive oxygen species, autophagy and glycolytic inhibition, and occurred without BHB being oxidised.

  13. Review2017
    Multi-dimensional Roles of Ketone Bodies in Fuel Metabolism, Signaling, and Therapeutics

    Cell Metabolism

    Synthesis of ketone bodies as both fuel and signalling molecules, including the enzymology of HMGCS2, BDH1 and SCOT and the reason hepatocytes cannot oxidise their own product.