What Is the Ketogenic (Keto) Diet, and Does It Actually Work?

The ketogenic diet, usually shortened to keto, is a high-fat, very-low-carbohydrate, moderate-protein way of eating. The defining feature is how far it cuts carbohydrate: typically to roughly 20 to 50 grams a day, far below ordinary eating, in order to push the body into a metabolic state called nutritional ketosis. In that state the body runs more on fat-derived fuels than on glucose. The pattern began as a medical therapy and later spread as a popular weight-loss approach, which is why its reputation runs well ahead of what the evidence cleanly supports.

This guide is general and educational. It explains what keto is, how ketosis actually works and why nutritional ketosis is not the dangerous diabetic ketoacidosis it is sometimes confused with, what controlled studies show about weight loss, the one medical use that is genuinely established, the real risks and side effects, and who should be cautious or seek medical supervision first. The broader low-carbohydrate pattern that does not require ketosis is the subject of the separate low-carb-diet guide, and the underlying energy-balance arithmetic that governs weight change is handled by the calorie-deficit guide, so this guide stays focused on the ketosis-inducing extreme rather than repeating either.

The essentials at a glance

• Keto limits carbohydrate to roughly 20 to 50 grams a day, about 5 to 10 percent of calories, to induce nutritional ketosis (StatPearls; corroborated by Bueno 2013, Harvard, and Cleveland Clinic).
• Nutritional ketosis is not ketoacidosis: it runs at a normal blood pH, whereas diabetic ketoacidosis is a life-threatening medical emergency (StatPearls; Cleveland Clinic).
• The weight-loss edge over a low-fat diet is small and adherence-driven: a meta-analysis found about 0.9 kg (about 2 lb) more loss at a year or more, and that gap did not clearly persist as compliance declined (Bueno 2013; Harvard).
• The one established medical use is drug-resistant epilepsy, mainly in children, as a clinically supervised therapy at low to very low certainty of evidence (Cochrane review).
• Lipids are mixed: the same meta-analysis found lower triglycerides and higher HDL, but it also found keto raised LDL cholesterol in some people (Bueno 2013; Harvard).
• Keto is contraindicated or needs medical supervision for several groups, and long-term, high-quality data on it remains limited (StatPearls; Harvard).

What the ketogenic diet is and how ketosis works

At the level of food, keto is defined by a carbohydrate ceiling rather than by any single permitted food list. StatPearls describes the classic ketogenic diet as limiting carbohydrate intake to roughly 20 to 50 grams daily, on the order of 5 to 10 percent of energy, with the rest of calories coming mostly from fat and a moderate amount from protein. The exact macronutrient split is best read as an approximate range, not a fixed formula: StatPearls gives figures spanning roughly 55 to 75 percent of calories from fat, about 25 to 35 percent from protein, and about 5 to 10 percent from carbohydrate, while the Harvard T.H. Chan School of Public Health describes an overlapping split of around 70 to 80 percent fat. The common thread is that carbohydrate is kept very low and fat supplies most of the energy.

The physiology behind the carbohydrate cut is what makes keto distinct. When carbohydrate intake falls, blood glucose and the hormone insulin drop, which according to StatPearls promotes lipolysis, the release of fatty acids from fat tissue. Those fatty acids travel to the liver, which converts them into ketone bodies, namely acetoacetate, beta-hydroxybutyrate, and acetone, that tissues including the brain can use as an alternative fuel. The general machinery of how the body switches between glucose and fat for fuel is covered in the guide on how the body uses energy; the point specific to keto is that sustained low carbohydrate shifts the fuel mix toward these ketone bodies, the state called nutritional ketosis.

That state is frequently confused with a dangerous condition, and the distinction matters. Nutritional ketosis raises blood ketones modestly while blood pH stays normal. StatPearls states plainly that nutritional ketosis, with a normal blood pH, differs from ketoacidosis, which is characterized by metabolic acidosis. Diabetic ketoacidosis is a separate, life-threatening emergency seen mainly in uncontrolled diabetes, in which ketones and blood acidity climb to dangerous levels. The Cleveland Clinic makes the same separation, describing nutritional ketosis as not harmful for most people while diabetes-related ketoacidosis is life-threatening. The two share a word root but are not the same event.

What the evidence shows for weight loss

The honest summary is that keto can help with weight loss, but its measured advantage over other diets is small and does not look durable. The most cited evidence is a 2013 meta-analysis by Bueno and colleagues, which pooled 13 randomized controlled trials covering 1,415 patients with at least 12 months of follow-up. It found that people assigned to a very-low-carbohydrate ketogenic diet achieved a greater long-term weight loss than those assigned to a low-fat diet, with a weighted mean difference in body weight of about -0.91 kg, roughly 2 lb, in favor of keto. That is a real difference, but a modest one across a year or more.

What that number means in practice is best read with the convergence framing from Harvard, which reviewed the same body of work. Harvard notes that the small weight-loss difference of about 2 pounds, together with compliance to the ketogenic diet declining over time, explained the more significant difference seen at one year but not at two years. In other words, the early edge appears to be driven largely by how well people stuck to the diet, and it tends to shrink as adherence fades. There is no good evidence that keto melts fat through some metabolic magic independent of calories; the underlying driver of weight change is energy balance over time, which the calorie-deficit guide explains in detail and which the TDEE calculator on this site helps estimate. Keto is best understood as one way some people find it easier to eat less, not as a diet that suspends the rules of energy balance.

It is also worth being candid about the limits of the evidence. Harvard notes that available research on the ketogenic diet for weight loss is still limited, and that most studies so far have had small numbers of participants and were short-term. The reasonable read is that keto is a viable option for people who prefer it and can sustain it, with a small short-term advantage that should not be oversold and that fades without long-term adherence.

Established medical use: drug-resistant epilepsy

Keto did not start as a weight-loss diet. The classic ketogenic diet is a long-standing, clinically supervised therapy for drug-resistant epilepsy, and this is the one use with genuine guideline support. StatPearls states that epilepsy remains the diet's only universally accepted, guideline-supported indication, which is a useful reality check against the wider claims sometimes made for keto.

The strength of that evidence should be stated precisely rather than inflated. The 2020 Cochrane systematic review by Martin-McGill and colleagues examined ketogenic diets for drug-resistant epilepsy across 13 studies with 932 participants, of whom 711 were children and 221 were adults. In children, the diets were associated with greater seizure freedom, at a relative risk of about 3.16, which the reviewers graded as very-low-certainty evidence, and with a 50 percent or greater reduction in seizures, at a relative risk of about 5.80, graded as low-certainty evidence. The review concluded that for people who have drug-resistant epilepsy or who are unsuitable for surgical intervention, ketogenic diets remain a valid option, while noting that the adult evidence is more limited and uncertain.

Two qualifications follow directly from that. First, this is a supervised therapeutic diet managed by a clinical team, not a do-it-yourself lifestyle plan, and the evidence is strongest in children rather than the general adult population. Second, the certainty of the evidence is low to very low even where the effect is real, so the epilepsy use is a reason keto exists as a medical tool, not a reason for the general population to adopt it for unrelated goals. Outside epilepsy, claims that keto prevents or treats other diseases are not supported by guideline-level evidence and are not made here.

Risks and side effects

The most familiar downside of starting keto is the transient cluster of symptoms popularly called keto flu. In the early days of the transition, as the body adapts to very low carbohydrate, some people experience fatigue, headache, nausea, constipation, irritability, and a foggy feeling sometimes described as brain fog. The Cleveland Clinic uses the term keto flu for this short-lived adjustment phase, and Harvard describes the same transient symptom cluster. StatPearls similarly lists nausea, vomiting, constipation, headache, fatigue, dizziness, and transient low blood sugar among the diet's reported early side effects. These symptoms are generally temporary, but they are a real part of the experience for many people.

Beyond the transition, StatPearls flags several longer-running risks: deficiencies in multiple micronutrients, an increased tendency toward kidney stones, and reduced bone mineral density. These are reasons the diet benefits from planning and, in clinical use, from monitoring. The cardiometabolic picture is genuinely mixed and should not be reduced to a single headline. Bueno and colleagues found that the very-low-carbohydrate ketogenic diet lowered triglycerides and diastolic blood pressure and raised HDL cholesterol, all generally regarded as favorable, but the same meta-analysis also found that it raised LDL cholesterol, by a weighted mean difference of about 0.12 mmol/L. Harvard adds that the saturated fat common in many keto menus can have adverse effects on LDL cholesterol. The plain reading is that keto improves some markers while worsening LDL in some people, so anyone with cardiovascular risk factors has a clear reason to involve a clinician and to monitor lipids rather than assume the effect is uniformly positive.

One specific medical context deserves its own mention. A 2015 review in the journal Nutrition by Feinman and colleagues argued that dietary carbohydrate restriction reliably reduces high blood glucose and can lead to a reduction or elimination of diabetes medication. That is a meaningful point for glycemic control, but it is precisely why people with diabetes must not adjust a carbohydrate-restricted diet on their own: because the diet itself lowers blood glucose, anyone taking glucose-lowering medication needs clinical supervision so that medication can be adjusted safely, and the safety section below covers the groups for whom keto is outright contraindicated.

Keto versus ordinary low-carb

Keto and low-carb are often used as if they mean the same thing, but they are not interchangeable. Keto is the ketosis-inducing extreme of carbohydrate restriction: the roughly 20 to 50 grams of carbohydrate a day described above is low enough to shift the body into nutritional ketosis. Ordinary low-carbohydrate eating is a broader, looser pattern that simply reduces carbohydrate, commonly to somewhere under about 130 grams a day, without necessarily reaching ketosis at all. Many people eat low-carb without ever being in ketosis.

The practical upshot is that the two approaches differ in how strict they are and in what they ask of the body's metabolism. The broader low-carbohydrate pattern, including its general trade-offs and how it compares with other ways of eating, is covered in the low-carb-diet guide, so this guide does not repeat that ground. The takeaway specific to keto is that its results and its risks both flow from the depth of the carbohydrate cut, which is what separates it from low-carb eating in general.

Frequently asked questions

How many carbs can I eat on keto?

The classic ketogenic diet limits carbohydrate to roughly 20 to 50 grams a day, about 5 to 10 percent of calories, which is the threshold StatPearls describes as low enough to induce nutritional ketosis. The exact amount that keeps a given person in ketosis varies, but that range is the standard target cited across sources, and it is well below ordinary carbohydrate intake.

Is the keto diet safe?

For many healthy adults keto is tolerated in the short term, but it carries real caveats. A transient keto flu is common at the start, the diet can raise LDL cholesterol in some people, and StatPearls lists risks including micronutrient deficiency, kidney stones, and reduced bone density. It is contraindicated or needs medical supervision for several groups, and long-term, high-quality data remains limited (Harvard). Anyone with health conditions or on medication should ask a clinician before starting.

Does keto work better than other diets for weight loss?

Only by a small margin that does not clearly last. Bueno and colleagues' 2013 meta-analysis found a very-low-carbohydrate ketogenic diet produced about 0.9 kg, roughly 2 lb, more weight loss than a low-fat diet at a year or more, but Harvard's read is that this gap was driven by adherence and did not persist at two years. Weight change is governed by energy balance over time, the mechanism explained in the calorie-deficit guide, rather than by anything unique to keto.

What is keto flu?

Keto flu is the popular name for a transient cluster of symptoms some people experience in the first days of switching to very low carbohydrate, including fatigue, headache, nausea, constipation, irritability, and brain fog. The Cleveland Clinic uses the term for this short-lived adjustment phase, Harvard describes the same transient symptom cluster, and it generally eases as the body adapts.

Who should not do the keto diet?

StatPearls lists keto as contraindicated in people with acute or chronic pancreatitis, liver failure, or certain fat-metabolism disorders, and states it is not recommended during pregnancy and is contraindicated in those with a history of eating disorders. People with diabetes who use SGLT2 inhibitors should avoid it, and anyone on glucose-lowering medication should have clinical supervision. These groups should get clinical input before considering keto.

Is keto the same as low-carb?

No. Keto is the ketosis-inducing extreme of carbohydrate restriction, at roughly 20 to 50 grams of carbohydrate a day, low enough to shift the body into nutritional ketosis. Ordinary low-carbohydrate eating is a broader pattern that reduces carbohydrate, commonly to under about 130 grams a day, without necessarily reaching ketosis. The broader pattern is covered in the low-carb-diet guide.

References

1. 1The Ketogenic Diet: Clinical Applications, Evidence-based Indications, and Implementation (Masood W, Annamaraju P, Khan Suheb MZ, Uppaluri KR; StatPearls, NCBI Bookshelf NBK499830) · StatPearls Publishing / National Library of Medicine. Accessed 2026-06-05.
2. 2Very-low-carbohydrate ketogenic diet v. low-fat diet for long-term weight loss: a meta-analysis of randomised controlled trials (Bueno NB, de Melo ISV, de Oliveira SL, da Rocha Ataide T), Br J Nutr 2013;110(7):1178-1187 (PubMed 23651522) · British Journal of Nutrition (via PubMed, National Library of Medicine). Accessed 2026-06-05.
3. 3Ketogenic diets for drug-resistant epilepsy (Martin-McGill KJ, Bresnahan R, Levy RG, Cooper PN), Cochrane Database Syst Rev 2020, CD001903.pub5 (PMC7387249) · Cochrane Database of Systematic Reviews (via PMC, National Library of Medicine). Accessed 2026-06-05.
4. 4Dietary carbohydrate restriction as the first approach in diabetes management: critical review and evidence base (Feinman RD, Pogozelski WK, Astrup A, et al.), Nutrition 2015;31(1):1-13 (PubMed 25287761) · Nutrition (via PubMed, National Library of Medicine). Accessed 2026-06-05.
5. 5Diet Review: Ketogenic Diet for Weight Loss (The Nutrition Source) · Harvard T.H. Chan School of Public Health. Accessed 2026-06-05.
6. 6Ketosis · Cleveland Clinic. Accessed 2026-06-05.
7. 7Euglycemic Diabetic Ketoacidosis Associated With SGLT2 Inhibitors and the Ketogenic Diet (case series), AACE Clinical Case Reports (PubMed 33851013) · AACE Clinical Case Reports (via PubMed, National Library of Medicine). Accessed 2026-06-05.