High blood glucose together with vomiting, belly pain or fast, deep breathing is an emergency, call 911.
The rest of this guide covers what the number reflects during fasting and carbohydrate restriction, how nutritional ketosis differs from ketoacidosis, which markers belong alongside it, and when a reading is worth a prompt call.
It measures the concentration of beta-hydroxybutyrate in a blood sample, the ketone body that makes up most of the ketones circulating when your body is running largely on fat.
The mechanism is straightforward. When carbohydrate intake drops, when a fast runs long, or when insulin signalling falls away, fat tissue releases fatty acids and the liver converts them into ketone bodies that other organs can burn. Three are produced, acetoacetate, beta-hydroxybutyrate and acetone, and beta-hydroxybutyrate is the one that dominates in blood during most states of ketosis 1. Chemically BHB is a hydroxy acid rather than a true ketone, but labs regard it as a ketone body because it belongs to the same metabolic pathway and behaves the same way clinically.
Ketones are not waste. The brain, heart and skeletal muscle can oxidise them, which is why a person can go a long time without carbohydrate and still think clearly. A ketone test is therefore not a test for illness, it is a test of fuel state. The number tells you how far your metabolism has shifted toward fat as its primary energy source at the moment the sample was taken.
Most meters and many labs report BHB in mmol/L. Some laboratories report serum beta-hydroxybutyric acid in mg/dL instead, so always read the units printed next to the result before comparing one report to another.
Because it is the most abundant ketone in blood and the one that tracks the current state of ketosis most closely.
Acetoacetate converts to beta-hydroxybutyrate inside cells, and the balance between the two shifts with how reduced the cell environment is. In deeper or more severe ketosis the balance tips further toward BHB, so a method that measures only acetoacetate can underestimate how much ketone is actually present 1.
Reviews of ketone measurement describe point-of-care blood BHB meters as the method that reflects circulating ketone status most directly, with laboratory enzymatic assays used when a precise plasma value is needed 2. That is the practical reason clinicians moved toward blood testing for anyone whose ketone level genuinely needs to be known.
In an adult eating regular meals, blood beta-hydroxybutyrate usually sits low, generally under about 0.6 mmol/L, and climbs with fasting, carbohydrate restriction or prolonged exertion.
The bands below are the interpretation ranges commonly printed with ketone meters and laboratory reports. They are typical ranges, not diagnostic cut-offs, and they vary by lab, by meter and by the clinical situation.
| Blood beta-hydroxybutyrate | Commonly described as | What it often reflects |
|---|---|---|
| Below 0.6 mmol/L | Typical fed-state level | Carbohydrate is still the main fuel |
| 0.6 to 1.5 mmol/L | Mildly raised | Fasting, carbohydrate restriction, prolonged exercise or a long overnight gap |
| 1.5 to 3.0 mmol/L | Clearly raised | Deeper ketosis, or a ketone rise worth discussing with a clinician, particularly alongside high glucose |
| Above 3.0 mmol/L | Markedly raised | Worth discussing with a clinician promptly, and urgent assessment if symptoms are present |
Typical lab reference ranges vary by laboratory, age and sex, and a single reading is a snapshot of one moment, not a trend.
Because a ketone number only makes sense next to a glucose number, the companion marker is worth showing too.
| Marker | Standard lab reference range | Optimal range |
|---|---|---|
| Fasting glucose | 65 to 99 mg/dL | 75 to 86 mg/dL |
The optimal range shown is FitScript's optimization target rather than a diagnostic cut-off, and standard reference ranges still vary by laboratory, age and sex.
Nutritional ketosis is a regulated fuel switch with low or normal glucose, while diabetic ketoacidosis is an unregulated ketone build-up that occurs with high glucose and acid accumulation in the blood.
The difference is insulin. Even a small amount of circulating insulin puts a brake on fat release and on liver ketone production, so ketones plateau in a fasting or low-carbohydrate state rather than climbing without limit. Reviews of ketone physiology describe ketoacidosis as what happens when that brake is effectively absent and counter-regulatory hormones drive ketone production far past the body's buffering capacity 1. The result is a very high ketone level, high glucose in most cases, and a drop in blood pH.
The practical tells differ too. Nutritional ketosis produces no vomiting, no abdominal pain and no laboured breathing. Ketoacidosis typically does, alongside thirst, frequent urination, drowsiness and a sweet or acetone smell on the breath. Research suggests that plasma beta-hydroxybutyrate is useful for defining when diabetic ketoacidosis has actually resolved, rather than relying on indirect markers that lag behind 4.
It is also a mistake to assume ketoacidosis only happens in type 1 diabetes. Research suggests that about one third of adults presenting with new-onset diabetic ketosis fit a ketosis-prone type 2 diabetes pattern rather than classic type 1 3. A raised ketone reading with high glucose is never something to interpret alone.
A high reading reflects heavy reliance on fat for fuel, and the list of reasons runs from entirely ordinary to needing same-day care.
Common benign drivers include an overnight or extended fast, sustained carbohydrate restriction, a day of low food intake during illness, and prolonged endurance exercise. Studies suggest that supervised long fasts produce a predictable, regulated rise in ketones, with metabolic parameters and safety measures tracked across a large cohort of participants 5. In those situations ketones rise because fat is doing the work, not because something has broken.
Less ordinary drivers matter more. Insufficient insulin signalling, acute illness with fever or infection, heavy alcohol intake paired with little food, and vomiting that prevents eating can all push ketones up while the body is also under strain. International expert recommendations on continuous ketone monitoring note that some diabetes medicines are associated with ketone elevation even when glucose readings look close to normal, which is why ketones and glucose are interpreted together rather than in isolation 6.
The useful sequence is observation, then meaning, then next step. The observation is the number. The meaning depends entirely on context, what you ate, how long since you ate, whether you are unwell, and what your glucose is doing. The next step, for any reading that is high and unexplained, or high alongside symptoms, is a prompt conversation with a clinician. Management belongs to the clinician, not to a meter.
A low reading simply means carbohydrate is still your main fuel at that moment, which is the normal state for most people who have eaten in the past several hours.
There is no recognised deficiency state for ketones in a person eating normally. A value near zero after breakfast is expected, not concerning. Low is only interesting in one narrow context, when someone is deliberately fasting or restricting carbohydrate and wants to know whether the metabolic switch has happened. In that case a persistently low reading usually means glycogen stores and insulin levels are still high enough to suppress liver ketone production, often because of hidden carbohydrate, a shorter fast than intended, or simply individual variation in how fast the switch occurs.
What a low reading does not do is rule anything out. It does not confirm metabolic health, it does not confirm fat loss, and it does not say anything about insulin sensitivity on its own. Someone can have a low ketone level and a fasting glucose outside the standard range at the same time, which is exactly why ketones are read as part of a panel rather than as a standalone score. Timing matters too, since ketones commonly peak in the early morning or late in a fasted window and fall after eating. A single low number taken at the wrong hour tells you very little about the rest of the day.
Because both lower insulin, and low insulin is the signal that releases stored fat and switches the liver into ketone production.
In the first hours after a meal, insulin is high, glucose is plentiful and the liver stores rather than makes fuel. As hours pass without food, liver glycogen drains, insulin falls and fat tissue begins releasing fatty acids. The liver takes those fatty acids, oxidises them and packages the surplus as ketone bodies that cross into the brain, which cannot burn fatty acids directly. Carbohydrate restriction produces the same end state through a different route, by keeping insulin low without requiring total food restriction.
The rise is gradual and self-limiting. Studies suggest that during extended, supervised fasting, ketone levels climb and then stabilise within a regulated range rather than escalating indefinitely 5. That self-limiting quality is the single clearest difference between fasting physiology and the uncontrolled ketone accumulation seen in ketoacidosis.
This guide does not prescribe any eating pattern. Fasting and carbohydrate restriction are not appropriate for everyone, and in people with diabetes, a history of eating disorders, pregnancy, or any condition affecting kidney or liver function, the decision belongs with a clinician who knows the full picture. If you test ketones while changing how you eat, treat the number as feedback on fuel state, not as a target to chase.
Prolonged exercise can raise ketones, and the rise is often largest in the hours after the session rather than during it.
The classic description of this pattern is post-exercise ketosis, where ketone concentrations climb during recovery from sustained exertion as fatty acid release outpaces immediate energy demand 7. A long run, a long ride or a hard session done in a fasted state can all produce a reading well above a person's usual resting value, with no clinical significance whatsoever.
That has one practical consequence. If you test shortly after a long or fasted training session, expect a higher number and interpret it accordingly. Comparing a post-workout reading to a resting morning reading is comparing two different physiological states.
The context that changes this is diabetes. In people with diabetes, exercise plus a ketone rise plus high glucose is a combination that deserves attention rather than reassurance, because the same reading can arise from insulin insufficiency rather than from fuel demand. Symptoms separate the two. Feeling tired after a long effort is expected. Nausea, vomiting, abdominal pain or breathlessness is not, and needs urgent assessment. For anyone training hard while also tracking ketones for a medical reason, the testing pattern itself is a clinician conversation.
It is done either as a fingerstick using a handheld meter with ketone-specific strips, or as a venous blood draw analysed in a laboratory.
The fingerstick version takes a minute. You wash and dry your hands, insert a ketone strip, lance the side of a fingertip, apply the drop to the strip and read the result in seconds. Hand washing matters more than people expect, because residue from food on the fingertip can skew the reading. Ketone strips are separate from glucose strips, are sensitive to heat and humidity, and have their own expiry date, so a strip stored in a hot car is a strip worth discarding.
The laboratory version uses an enzymatic assay on serum or plasma and gives a precise value, which is the method of choice when a result will drive clinical decisions 2. In hospital settings, capillary ketone testing is used for specific indications rather than as a routine screen, and an audit of hospitalised patients examined how those indications map onto outcomes 8.
Preparation depends on the question. If the point is to see a resting baseline, test fasted in the morning before food or coffee with anything in it. If the point is to see how a meal, a fast or a session of exercise moved things, time the test to that event and note what you did. Record the time alongside the number, because a ketone value without a timestamp is close to uninterpretable.
Blood testing tells you most, because it measures beta-hydroxybutyrate directly.
| Method | What it measures | Strengths | Limits |
|---|---|---|---|
| Blood (fingerstick or lab) | Beta-hydroxybutyrate | Reflects current circulating ketone level, used clinically when the number matters | Strips cost more, requires a lancet |
| Breath analyser | Acetone in exhaled air | Reusable, no consumables, painless | Indirect, affected by breathing pattern and recent alcohol, less used clinically |
A newer option sits alongside these. Continuous ketone monitoring uses a wearable sensor to report ketone levels over time instead of at single moments, and clinical commentary describes it as a development with meaningful implications for how ketone risk is spotted early 9. International expert recommendations set out how such a technology is applied in diabetes care and who it suits 6. Whether any monitoring device is appropriate is a decision for a licensed provider.
Several everyday factors move the number without any change in underlying health, which is why one reading is rarely acted on alone.
Time since your last meal. Ketones drift upward across a long overnight gap and fall after carbohydrate. The same person can read 0.2 mmol/L at 9am after breakfast and well above that at 7am before it.
Hydration and illness. Fever, infection and vomiting raise fuel demand while reducing intake, so ketones can climb during a bad week with no dietary change at all.
Alcohol. Heavy intake combined with little food is a recognised route to a ketone rise, and the combination is worth taking seriously rather than dismissing as a hangover.
Pregnancy and breastfeeding. Fuel demand is higher, and ketones can appear after shorter gaps without food. Any raised reading during pregnancy is worth discussing with a clinician promptly.
Technique and storage. Unwashed fingers, expired or heat-damaged strips, a squeezed fingertip producing tissue fluid rather than blood, and comparing two different meters all introduce variation.
Medicines. Some prescribed medicines influence ketone production, and a licensed provider evaluates what that means for an individual. Treatment decisions belong to a licensed provider.
The habit that makes ketone data useful is consistency. Same meter, same time of day, same conditions, written down.
Fasting glucose is the number that gives a ketone reading its meaning, because the same ketone value means something different at a low glucose than at a high one.
High ketones with low or normal glucose generally points toward fasting or carbohydrate restriction. High ketones with high glucose points toward insufficient insulin signalling and belongs in a clinician conversation without delay. Low ketones with high glucose points somewhere else again.
What fasting glucose is, how it is measured, what moves it and how to read a result in detail is covered on its own page. See the Fasting Glucose guide for the full picture, including how it compares with longer-term glucose markers and what a single fasting value can and cannot show.
The markers that matter most are the ones describing acid-base status and glucose handling, because those are what turn a ketone number into a usable picture.
Bicarbonate, reported on most panels as CO2, reflects the buffer the body uses to absorb acid. When ketone acids accumulate faster than they can be buffered, bicarbonate falls, which is one of the earliest laboratory signs that a ketone rise has moved beyond a simple fuel switch. The anion gap, calculated from measured electrolytes, widens when unmeasured acids such as ketone acids build up, and a normal gap alongside a mildly raised ketone level is reassuring in a way that a wide gap is not.
Around those sit fasting glucose, a longer-term glucose marker, electrolytes including sodium and potassium, and kidney function. Each adds a dimension. Glucose says what the fuel supply looks like, electrolytes say how the body is coping with fluid and acid shifts, and kidney markers say how well the system is clearing what it needs to clear.
This is the argument for panels rather than single tests. A FitScript membership includes 150+ biomarkers tested per year, and Atlas, the AI health advisor, reads them together rather than one marker at a time, which is how a ketone value is meant to be interpreted anyway.
Any high ketone reading accompanied by symptoms, or any high reading with high glucose, is worth contacting a clinician about promptly rather than waiting to retest.
Call 911 or go to emergency care for high glucose together with vomiting, abdominal pain, rapid deep breathing, confusion, drowsiness, or a sweet acetone smell on the breath. That combination does not wait for a clinic appointment.
Contact a clinician the same day for a markedly raised ketone reading that has no obvious explanation, for a ketone rise during illness in a person with diabetes, for persistent vomiting that prevents eating or drinking, or for a raised reading during pregnancy. Bring the actual numbers, the times they were taken and the glucose values alongside them, because the pattern carries more information than any single result.
Book a routine conversation for a ketone level that keeps reading higher than expected during ordinary eating, for a fasting glucose outside the standard range found alongside ketone testing, or before starting any substantial change in how you eat if you have an existing condition. Never ignore a result that falls outside the standard range, and never assume a meter is wrong because the number is inconvenient.
This guide is educational and is not medical advice. Discuss your results, and any question about treatment, with a clinician or licensed provider who can see your full history.
It cannot diagnose anything on its own, and it cannot tell you why the number is what it is.
A ketone value is a measure of fuel state at one moment. It does not distinguish a long fast from an insulin problem without glucose and clinical context beside it. It says nothing about diet quality, metabolic health in general, or long-term risk.
It also carries ordinary measurement noise. Meters vary against laboratory assays, strips vary between lots, and technique varies between tests. Two readings twenty minutes apart can differ enough to change which band they fall into without anything having changed in the person.
What the test does well is answer one specific question clearly, how far toward fat fuel the body has shifted right now, and whether that shift is in a range that calls for clinical attention. Used that way, inside a panel and with a timestamp, it is a genuinely useful marker. Used alone as a score to optimise, it mostly generates anxiety.
For the markers that sit closest to a ketone result, the co2 guide explains how bicarbonate reflects the body's acid buffer and why it falls when acid accumulates, and the anion gap guide explains how laboratories use the gap between measured electrolytes to flag unmeasured acids such as ketone acids.
In an adult eating regular meals, blood beta-hydroxybutyrate usually sits below about 0.6 mmol/L. Levels climb with fasting, carbohydrate restriction and long exertion, and the bands printed with meters and lab reports are typical interpretation ranges rather than diagnostic cut-offs. Ranges vary by lab, meter, age and sex, so read your own report's units and reference range.
A blood test measures beta-hydroxybutyrate, the ketone that dominates in circulation. A urine strip detects acetoacetate instead, so it reflects an earlier time window and is affected by how hydrated you are. That is why blood testing is used when the actual number needs to be known.
It depends on the question being asked. For a resting baseline, test in the morning before food or drinks containing calories. If you want to see how a meal, a fast or a workout moved things, time the test to that event and write down what you did, because a ketone number without a timestamp is hard to interpret.
Yes. A long fast, sustained carbohydrate restriction, a day of illness with little food, heavy alcohol intake with no food, and prolonged endurance exercise can all raise ketones in someone without diabetes. A high reading that has no obvious explanation is still worth discussing with a clinician.
Readings above roughly 3.0 mmol/L are described as markedly raised and are worth discussing with a clinician promptly. Any raised reading combined with vomiting, abdominal pain, rapid deep breathing, confusion or a sweet smell on the breath, especially with high glucose, needs emergency care, call 911.
Fasting, carbohydrate restriction and long exercise raise ketones while glucose stays normal, which is ordinary physiology. There are also medical situations where ketones rise with near-normal glucose, so if the pattern is unexplained or comes with symptoms, contact a clinician rather than assuming it is diet related.
It can. Ketones often rise in the hours after a long or fasted training session, so a post-workout reading will not match a resting morning one. Compare like with like by testing under the same conditions each time, and treat a post-exercise number as its own category.
For most people with no medical reason to track them, ketones are not a routine test. Where there is a clinical reason, how often to check is set by the clinician managing your care, usually tied to glucose readings, illness or specific symptoms. Any next step, including whether to monitor at all, is a decision for a licensed provider.