The ketogenic diet has one medical indication that clinical guidelines fully accept, and it is not weight loss — it is epilepsy, and it has been used for that since 1920. Everything else about keto sits on evidence that is real but heavily qualified. This page separates the two, gives the actual macronutrient numbers, and lists the documented side effects, which most beginner guides skip.
The short version
- Standard composition: roughly 55–60% fat, 30–35% protein, 5–10% carbohydrate. Weight-loss versions push fat to 60–75% and carbohydrate to 20–50 g a day.
- Only fully accepted indication: drug-resistant epilepsy, with guideline support behind it.
- For weight loss and blood sugar there are trials showing benefit — and the same sources call the effects modest, adherence poor, and find little difference versus balanced-carbohydrate diets at two years.
- Documented side effects run from bad breath and constipation to kidney stones and reduced bone density.
- The Dietary Guidelines do not mention ketogenic or low-carbohydrate eating at all.
What the diet actually is, in numbers
A clinical reference on the NIH’s NCBI Bookshelf gives the standard composition as about 55% to 60% fat, 30% to 35% protein, and 5% to 10% carbohydrates. The version used for weight loss is stricter still: 60% to 75% of calories from fat with carbohydrate held to 20 to 50 grams daily.
To see what those limits mean in practice, our cards carry the carbohydrate figures per 100 g: avocado, eggs, salmon, walnuts, olive oil and broccoli sit inside a 20–50 g daily budget, while rice, oats and bread use most of it in one portion.
For scale, a separate reference sets the tiers: very low-carbohydrate is under 10% of calories or 20 to 50 g/d; low-carbohydrate is under 26% or less than 130 g/d. Compare that with the general recommendation of 45–65% of calories from carbohydrate and you can see how far outside ordinary guidance keto sits. That is not an argument against it — it is the reason it needs its own evidence, which it partly has.
The one established use
NINDS describes the ketogenic diet as a form of therapy for seizures that involves a brief period of fasting followed by a very carefully controlled high fat-low protein and carbohydrate diet, and states that studies suggest it is useful in children whose seizures could not be controlled by any of the medicines available.
The clinical reference is blunter about the boundary: epilepsy is its only universally accepted, guideline-supported indication, with consistent reductions in seizure burden across pediatric and adult populations. And this is not a trend — ketogenic diets have been used successfully since 1920, before the existence of medications for epilepsy.
Worth holding onto: the diet with the strongest clinical pedigree in this whole space earned it treating a neurological condition under medical supervision, not in a body-transformation context.
Weight loss and blood sugar: the qualified version
The honest summary is that the evidence exists and is weaker than the marketing. The same clinical source reports that multiple randomized controlled trials and meta-analyses have demonstrated its effectiveness for weight loss and blood sugar control — and immediately qualifies it: effects on HbA1c were modest, and adherence remained a challenge.
Set against that, a diabetes chapter on the same Bookshelf reports that between low-carbohydrate and balanced-carbohydrate diets there is probably little to no difference in weight reduction and changes in cardiovascular risk factors up to two years’ follow-up. And NIDDK still lists low-carbohydrate eating for blood glucose as a question under study rather than a recommendation.
So: keto can produce weight loss. So can other approaches, and by two years the difference washes out in the trials. If keto is the pattern you can actually stick to, that is a legitimate reason to pick it. “It works better” is not supported.
The side effects list
Beginner guides tend to mention “keto flu” and stop. The clinical source is more complete. Short term: nausea, vomiting, constipation, diarrhea, halitosis, headache, fatigue, dizziness. Longer term: dyslipidemia, hepatic steatosis, cardiomyopathy, hypercalciuria and kidney stones, reduced bone mineral density, and growth impairment in children.
There is also a population-level signal worth stating carefully: one reference notes that epidemiological studies and meta-analyses have shown an increased mortality risk with a carbohydrate intake of less than 40%, while adding that contradictory findings exist. A diabetes guideline chapter puts both ends together: both very high and low intakes are associated with premature mortality.
One thing that is not on the risk list: a clinical source states there is no evidence that very low-carb intake produces metabolic ketoacidosis and remains safe in patients, even those with type 2 diabetes. Nutritional ketosis and diabetic ketoacidosis are different things, and conflating them is a common scare.
What the Dietary Guidelines say about keto
Nothing. We checked the Guidelines’ own questions page and it does not mention ketogenic or low-carbohydrate patterns at all. What it does say is that the guidance is a customizable framework of core elements within which individuals can make tailored and affordable, nutrient-dense choices that meet their personal preferences and cultural traditions.
Read that as neither endorsement nor prohibition. If a very-low-carbohydrate pattern meets your nutrient needs, nothing federal forbids it — and nothing federal recommends it either.
What we are not claiming
- “Keto is proven for weight loss.” Trials show effect; the same sources call it modest, poorly adhered to, and no different from balanced-carb diets at two years.
- “Ketosis burns fat faster.” No source we could reach makes that mechanistic comparison at matched calories.
- That keto treats anything other than epilepsy with guideline support. The clinical reference is explicit that epilepsy is the only such indication.
- That keto is dangerous. The side-effect list is real, and so is the statement that very-low-carb intake does not produce ketoacidosis. Both belong here.
- Specific macro targets for you. The ranges above are what the literature describes, not a prescription.
FAQ
How many carbs on keto? 20 to 50 grams a day in the weight-loss formulation, or 5–10% of calories in the standard clinical one.
Is keto better than other diets for losing weight? Not according to the sources we could verify — at up to two years, little to no difference versus balanced-carbohydrate diets.
What is keto actually proven for? Drug-resistant epilepsy, where it has guideline support and a century of clinical use.
Is keto safe long term? The documented longer-term risks include dyslipidemia, fatty liver, kidney stones and reduced bone mineral density. That is a reason for medical supervision, not automatic avoidance.
Will keto cause ketoacidosis? A clinical source states there is no evidence that very-low-carbohydrate intake produces metabolic ketoacidosis, including in people with type 2 diabetes.
What to do with this
If you are considering keto for weight loss, judge it on one question: can you eat this way for a year? The evidence says the outcome will be similar to other patterns you could stick to, so adherence is the deciding variable rather than metabolic magic. If you have epilepsy, kidney disease, diabetes, or take medication, this is a diet to start with a clinician, not with a blog.
If you want the arithmetic done for you, the keto macro calculator sets the three numbers, and the calorie calculator sets the level they divide.
Related: what federal guidance says about carbohydrate, what fat loss actually has evidence for.
Educational content, not medical advice.
This content is for educational purposes only and is not medical advice. Always consult a qualified healthcare professional for personalised guidance.
