Ketosis and Low Carbohydrate: What Changes When Fuel Shifts

Bottom line

Ketosis is a natural physiological state in which, when carbohydrate is very low, the body switches to fat and ketone bodies as fuel. It is a mature treatment for drug-resistant epilepsy. It is not a fat-loss shortcut that fits everyone, and it is not a proven longevity formula.

For a person whose insulin is still working, nutritional ketosis and the dangerous state of diabetic ketoacidosis are not the same thing ¹. When calories and protein are matched, a low-carbohydrate diet has no extra fat-burning advantage ²². What most deserves watching is LDL cholesterol: in pooled results it rises a little on average ²³. Triglycerides usually fall, and HDL usually rises ²³. In a minority of people the rise is very large ²⁶.

One view now tests it inside a broader dietary frame ⁴⁵. Another is neutral and more interested in the brain and its potential for mental health ⁹¹⁰. A third body of writing weights ketone bodies as signals, which fasting and exercise can also raise ¹³.

PearHealth: a tool with a defined use, not a creed. Anyone planning to try it long-term is better off first reviewing their own lipids and medications with a clinician.

What it is

Ketosis means the body switches from burning mostly glucose to burning mostly fat, while the liver turns part of that fat into "ketone bodies" for the rest of the body to use.

Ketone bodies are the family name for three small molecules: acetoacetate, beta-hydroxybutyrate, and acetone ¹. Of these, beta-hydroxybutyrate (BHB) is the most abundant in blood and the one most often measured. The brain can use it directly.

The brain is about 2% of body weight and uses about 20% of the body's energy. When food is unavailable for a long time, ketone bodies are its reserve fuel ¹.

There are several roads into ketosis: going a long time without eating, a ketogenic diet, exercise (especially fasted), a class of glucose-lowering drugs called SGLT2 inhibitors, and medium-chain triglycerides, all of which can raise ketone bodies ¹³.

Nutritional ketosis, in a conversation with D'Agostino, is described as blood ketones (BHB) reaching at least 0.5 millimoles per liter, with a band they called ideal of 1 to 3 millimoles per liter ¹⁶. Ketoacidosis is the emergency in which, when insulin is severely lacking, ketone bodies surge out of control and the blood turns acidic. It is most common in people with type 1 diabetes ³³. One long piece was written to separate the two ¹.

A low-carbohydrate diet is the broader term: carbohydrate is reduced, but not necessarily far enough to enter ketosis. A ketogenic diet is the strictest step inside low carbohydrate.

Why it matters

First, it is a medical therapy. The ketogenic diet was originally designed to control epilepsy ¹⁰, and drug-resistant epilepsy in children has randomized-trial support ¹⁹.

Second, it affects glucose management. In a one-year study, people with type 2 diabetes who entered nutritional ketosis under a medical team's guidance improved both glycated hemoglobin, a reflection of average glucose over the past two or three months, and body weight by a clear amount ²⁴.

Third, it may push cardiovascular risk up in some people. An observational study in the UK Biobank found that people who reported a low-carbohydrate, high-fat diet had higher LDL cholesterol and apolipoprotein B, and more major cardiovascular events during follow-up ²⁷. Apolipoprotein B (ApoB) counts the particles that carry cholesterol. More particles, more chance of a blockage.

Different perspectives

Main positions

One position: it is a natural state that gets misunderstood, and the diet does not fit everyone. A long piece on ketone physiology was written, mainly to separate nutritional ketosis from ketoacidosis ¹.

One piece said that no strictly controlled trial had answered whether ketosis raises calorie expenditure ². The judgment: possible gains in aerobic capacity and muscular endurance, no advantage for explosive power, a rise in strength unlikely ². In one self-test, exogenous ketone salts lowered oxygen use at the same cycling power, though it is a single case ⁶. One review notes that incomplete adaptation can discount exercise performance ⁷.

A later analysis of five cases of cholesterol rising sharply during ketosis pointed to the diet amplifying a genetic tendency ³. When lipids become a problem, part of the saturated fat can be swapped for monounsaturated fat, or lipid-lowering treatment discussed with a clinician ³. Any diet evaluated in that later frame is checked against five conditions: energy balance, metabolic health, enough protein, sufficient micronutrients, and the ability to keep it ⁴. Ketosis is one option under that check. There is also a caution against over-reading one short study, such as a fat-versus-carbohydrate comparison of brain reward, which cannot be pushed onto the cause of obesity ⁸.

Another view is neutral. Neurons need a large amount of glucose, and during ketosis they can switch to ketone bodies ⁹. The speaker says it is not clear which foods the pattern allows, and does not want to disparage it ⁹.

Psychiatrist Chris Palmer shared cases of ketosis as an add-on in depression, bipolar disorder, and schizophrenia, and said there is not yet a randomized trial proving efficacy in a large number of patients ¹⁰. He names the tension: "neurons love glucose," and glucose is being reduced ¹⁰. The direction he values is diet alongside existing medicines, not instead of them ¹⁰.

With Layne Norton, matched-calorie, matched-protein feeding studies showed low carbohydrate does not remove more fat ²². One public diet page does not list ketosis on its own ¹².

Other writing is more interested in BHB as a signal, not only a fuel. A topic page says BHB can inhibit histone deacetylases, the knobs on gene switches, and an inflammation switch called NLRP3, and that it is tied to longer life in worms. Human trials are few and mixed ¹³. Fasting and exercise can raise it too.

With Eric Verdin, one conversation discussed mice that, from midlife, alternated a ketogenic week with an ordinary week: fewer midlife deaths, better memory in old age, no extension of maximum lifespan ³¹. Verdin stresses that human genetic differences make the carry-over hard, and that extra protein during ketosis may cancel a benefit ¹⁵. One addition is that variants such as PPAR-alpha may make safe entry into ketosis hard for some people ¹⁵.

On exercise, a pool of nine studies found no performance gain for endurance athletes from periodic carbohydrate restriction, and high-intensity work usually needs more carbohydrate ¹⁴. A 4-day study dropped muscle glycogen about 40% and raised a marker of bone resorption ¹⁴. A well-designed ketogenic diet, in the D'Agostino conversation, includes plenty of fiber, not only meat and butter ¹⁶. In a conversation on brain injury, BHB and lactate are described as possible alternative therapeutic fuels ¹⁸. Ronald Krauss grants that some high responders are better off staying away from saturated fat ¹⁷.

Where views agree and differ

Ketosis is a capacity the body already has. It is not a fraud. People differ, and one cut does not fit everyone. Nutritional ketosis is not ketoacidosis ¹¹⁶. Drug-resistant epilepsy has a clear role ¹⁹¹⁰. Genes decide a large share of the difference: lipid background in one account ³, PPAR-alpha in another ¹⁵, saturated-fat high responders for Krauss ¹⁷.

On performance, a 2013 reading thought endurance might improve ². Another body of writing sees no advantage from carbohydrate restriction ¹⁴. Elite race walkers later supported that: the ketogenic group burned more fat and used more oxygen at the same speed, the 10-kilometer result did not improve, and the high-carbohydrate group improved 6.6% ²⁹.

On focus, one account watches lipids and individual risk ³. Another watches the brain and mental health ¹⁰. Others watch BHB's signaling and do not hold that a long ketogenic diet is required to get it ¹³.

PearHealth's synthesis

Ketosis is a tool with defined settings.

"Ketosis itself is not something to fear" is adopted. The two states split on whether insulin is severely lacking ¹³³. Type 1 diabetes, and some glucose-lowering drugs, are a different situation ³³¹³.

"No fat-burning magic" is adopted. Matched calories and protein, no extra fat loss ²². Eating freely in an inpatient study, a low-fat plant diet ran about 689 kilocalories a day below a ketogenic diet ²¹. Longer trials: the ketogenic group lost only 0.91 kilograms more ²³. The effect depends on sticking with it.

Lipids stay under watch. LDL rises a little on average ²³, individuals differ enormously ²⁶, and a low-carbohydrate high-fat pattern has been associated with more cardiovascular events ²⁷. That is not proof of cause. It is enough to make a repeat lipid panel part of long-term ketosis.

Longevity, cancer, and psychiatric claims stay open. Mouse results are only a hint ³¹. The psychiatric data are a retrospective look at 31 inpatients with no control group ³⁰. Neither supports an ordinary person staying ketogenic for those reasons. Large cohorts tie replacing carbohydrate with animal sources to higher death risk, and plant sources to the opposite ²⁸. What replaces the carbohydrate may matter as much as how much was removed.

How strong is the evidence

The firmest evidence is drug-resistant epilepsy in children, then short-term improvement in glucose and weight. Claims about longevity, cancer, and psychiatric illness mostly stay at animal research or early human research.

Epilepsy: there is a single randomized trial ¹⁹, and Cochrane pooled 13 randomized trials and 932 participants. The chance of seizures falling by half reached as high as six times usual care. Trials are small, and certainty is low to very low ²⁰.

Weight loss: longer randomized trials show a small ketogenic advantage ²³. Strict feeding studies show no fat-burning advantage. The gap is appetite and how much is eaten, not a metabolic trick ²².

Type 2 diabetes: the one-year study shows a large improvement ²⁴. It is not a randomized trial. Participation was voluntary, so the result may be optimistic. In the same research, LDL cholesterol rose about 9.9%, and ApoB was essentially unchanged ²⁵.

Lipid risk: in the pooled randomized trials, LDL cholesterol rose 0.12 millimoles per liter on average ²³. In a case report, 3 of 5 patients had total cholesterol rise above 500 milligrams per deciliter, and it fell clearly after a return to a less strict diet ²⁶. Evidence on cardiovascular events comes from observational research, and the diet was self-reported ²⁷.

Mortality: 15,428 people followed a median of 25 years, plus pooled prospective cohorts, showed higher death risk at both the lowest and highest carbohydrate ends, hazard ratios about 1.20 and 1.23. Observational ²⁸.

Other: the race-walker study is small and carefully designed ²⁹. Mental health has a retrospective look at 31 inpatients and no control group ³⁰. Longevity has only mouse data ³¹.

What you can do

The point is not "the less carbohydrate, the better." Name the purpose, find a safety boundary, and check the response on a blood panel.

• Get clear on the goal: glucose control, weight loss, or exercise performance? Many people start by cutting refined carbohydrate and sugary drinks, rather than going all the way at once. Whether it can be kept up is itself one of the conditions used in one public frame to evaluate a diet ⁴.
• People using a glucose-lowering medicine, including people with type 1 diabetes, talk with a clinician before changing the diet. In the one-year glucose-control study, insulin and sulfonylurea drugs were mostly reduced or stopped, and that needs a clinician in charge ²⁴. In the inpatient study, some people also withdrew because of low blood glucose ²¹. SGLT2 inhibitors themselves raise ketone bodies ¹³.
• Check lipids once before starting, and check them again after a few months for a comparison.
• If low carbohydrate is the choice, plant sources of fat and protein, and vegetables rich in fiber, are worth weighting more heavily ²⁸¹⁶.
• People doing high-intensity training or competing can arrange carbohydrate by training intensity, rather than one cut for the long term ¹⁴²⁹.
• Exogenous ketone drinks and a ketogenic diet are two different things. They cannot stand in for each other ²⁶.

Questions worth raising with a clinician: whether a current medicine needs a change if carbohydrate falls; whether familial high cholesterol means checking ApoB or genes first; when to repeat lipids and glucose; and, if LDL cholesterol or ApoB rises clearly, whether to change the fat source, add a medicine, or stop.

Common misconceptions

• "Ketosis is acidosis." When insulin is still present, nutritional ketosis and ketoacidosis are two different things. The latter happens mainly when insulin is severely lacking ¹³³.
• "Ketosis has a metabolic advantage, so any amount of food still leads to leanness, and a ketone drink does the same." Strict feeding studies do not support a metabolic advantage ²². One account also notes that exogenous ketones and nutritional ketosis differ, and that the self-test is one person's data ²⁶.
• "If HDL is up and triglycerides are down, the lipids are better across the board." Those changes are indeed common ²³²⁴. LDL cholesterol can rise at the same time, and in a minority of people the rise is startling ²⁶. The whole picture matters, ApoB in particular.
• "Ketosis makes every athlete stronger." In races that are about speed, ketosis may lower movement economy and cancel training progress ²⁹.
• "If the mice lived longer, people will too." That experiment lowered midlife death and did not extend maximum lifespan ³¹. Human genetic differences still block a straight carry-over ¹⁵.

How it connects to your blood work

Ketosis leaves traces on a routine panel. Some often improve. Some may worsen. The panel is the main check on whether this way of eating fits.

• LDL cholesterol (ldl-c) and total cholesterol (total-cholesterol): a small rise on average, a large rise in a minority. These are the ones most in need of a repeat check ²³²⁶.
• Apolipoprotein B (apob) and non-HDL cholesterol (non-hdl-c): ApoB counts atherogenic particles. In one study LDL cholesterol rose and ApoB did not ²⁵. In an observational study both were higher ²⁷.
• LDL particle number by NMR (ldl-p-nmr): in the one-year glucose-control study, small dense particles fell about 20.8% ²⁵.
• Triglycerides (triglycerides) and HDL cholesterol (hdl-c): on low carbohydrate the first usually falls and the second usually rises ²³²⁴.
• Fasting glucose (fasting-glucose) and glycated hemoglobin (hba1c): readings of glucose control, clearly improved in the type 2 diabetes research ²⁴.
• Fasting insulin (fasting-insulin) and an insulin-resistance index (homa-ir): in the one-year study the insulin-resistance index fell 55% ²⁴.
• High-sensitivity C-reactive protein (hs-crp): an inflammation marker, down 39% in the one-year study ²⁴.
• Bicarbonate (bicarbonate): falls when blood turns acidic. It flags ketoacidosis, not nutritional ketosis.

Blood ketones (BHB) are not on PearLab's current marker list. Confirming ketosis depends on a dedicated blood-ketone test. A home fingertip ketone meter can measure it too ³².