2025-03-01 · nutrition, low-carb, keto, ketogenic, diet, weight-loss, insulin, adherence

Updated 2026-07-30

What's the difference between low-carb and keto?

Low-carb is a spectrum — moderate low-carb runs 100–150 g of carbs a day, standard low-carb 50–100 g, and very-low-carb ketogenic under 50 g. Only the very-low-carb tier reliably induces nutritional ketosis, which typically starts 2–4 days after carbs drop below ~50 g. All three cut carbs, but only ketogenic patterns shift the body's primary fuel from glucose to ketones.

How long does it take to get into ketosis?

Most people reach nutritional ketosis 2–4 days after dropping to under 50 g of carbs per day, once liver glycogen is depleted. Blood ketones typically read 0.5–3.0 mmol/L on the ketogenic range. Exercise, an overnight fast, or MCT oil can shorten that window; a lingering evening snack or hidden carbs in sauces and drinks can push it out a full week.

Do low-carb diets actually work better than calorie counting?

Not at 12 months. DIETFITS (Gardner 2018, JAMA, 609 adults, 12 months) found low-carb and low-fat arms lost equivalent weight (~5–6 kg both arms). DIRECT (Shai 2008, NEJM, 2 years) showed a small low-carb edge that narrowed by year two. Low-carb wins the early sprint from water-weight loss, but the long-run delta versus a matched calorie deficit is small in trials.

What is keto flu and how do I prevent it?

Keto flu is the fatigue, headache, brain fog, and cramping that peaks days 3–7 as insulin drops, kidneys dump sodium and water, and glycogen empties. Prevent it by adding 1–2 g of extra sodium a day (broth, salted foods), 3–4 g of potassium from whole foods (avocado, leafy greens, salmon), and 300–400 mg of magnesium. Hydration alone is not enough — the mineral load is what blunts it.

Will keto raise my cholesterol?

For most people, HDL rises and triglycerides fall, and LDL is variable. A subset — often lean, insulin-sensitive people the community calls 'lean mass hyper-responders' — see LDL climb sharply on high-saturated-fat keto. The usual fix is swapping saturated fat for monounsaturated (olive oil, nuts, avocado). Recheck a full lipid panel 12 weeks after the pattern stabilizes; ApoB is more informative than LDL alone.

Can I do low-carb while still eating fruit?

Yes, at the moderate and standard low-carb tiers. Berries (about 5–8 g net carbs per half cup) fit even a 50 g ceiling; higher-sugar fruits like bananas, grapes, and mango push most of a ketogenic budget in one serving. On a strict ketogenic pattern, most people limit fruit to a handful of berries per day and treat other fruit as an occasional carb refeed rather than a daily food.

Is keto safe for people with type 2 diabetes?

For many people with type 2 diabetes, low-carb and ketogenic patterns are among the most effective non-drug interventions for glycemic control — the Virta Health trial (Bhanpuri 2018, Cardiovasc Diabetol) showed HbA1c dropping ~1.3 points and most participants reducing or stopping medications at one year. It requires clinician supervision, because insulin and sulfonylurea doses often need to come down quickly to prevent hypoglycemia.

How do I come off a keto diet without gaining the weight back?

Reintroduce carbs slowly — roughly 20–30 g extra per week, prioritizing whole foods (oats, sweet potato, beans, fruit) over refined ones. Expect 2–5 lb of water and glycogen weight to return in the first two weeks; that's not fat regain. Keep protein at 1.6 g/kg lean mass, keep resistance training, and settle at the lowest carb tier that keeps hunger and energy stable — often a moderate low-carb or Mediterranean-style hybrid.

Low-Carb and Keto Diets

Low-carb and ketogenic eating has been sold as everything from a metabolic miracle to a dangerous fad. The honest read from the trials is narrower and more useful: for the right person, cutting carbs is a durable way to eat less without counting; for the wrong person, it’s a restrictive template that quietly turns into “dirty keto” and stalls. This guide walks through what the protocols actually are, what the biggest trials show at 6 and 12 months, who they fit, who they don’t, and a 4-week starter template you can run without guessing.

Quick stats

• Short-term (3–6 mo) weight loss vs low-fat: low-carb typically wins by 2–4 kg in the first six months (Shai 2008 DIRECT; Bazzano 2014). Most of the extra loss is water and glycogen, not incremental fat.
• 12-month attrition in dietary trials: roughly 30–40% dropout on strict very-low-carb ketogenic arms; DIETFITS (Gardner 2018) reported 79% completion at 12 months on a coached “healthy low-carb” template, which is unusually high for the format.
• Ketosis onset: blood ketones typically cross 0.5 mmol/L within 2–4 days at ≤50 g of carbs per day, once liver glycogen is depleted.
• Average LDL response: neutral to modestly down at the population level, but with wide individual variance — a subset of lean, insulin-sensitive responders sees LDL climb sharply, especially on high-saturated-fat versions of the diet.
• Glycemic control in type 2 diabetes: ~1.0–1.3 point HbA1c drop at one year on ketogenic protocols in the Virta trial (Bhanpuri 2018), with most participants reducing or discontinuing sulfonylureas and insulin.

What “low-carb” and “keto” actually mean

“Low-carb” is a spectrum, not a single diet. Trials, guidelines, and the National Lipid Association’s 2019 position paper use rough thresholds that are worth keeping straight, because the biology at each tier is different.

• Moderate low-carb (100–150 g/day). About half the carbs of a standard Western diet. Usually enough to blunt post-meal glucose spikes and reduce snacking without inducing ketosis. This is the tier a Mediterranean-style low-carb hybrid usually lives at.
• Standard low-carb (50–100 g/day). The Atkins induction ceiling. Some people cross into mild ketosis at the low end, especially with exercise, but ketones are not consistent.
• Very-low-carb ketogenic (VLCKD, under 50 g/day). Reliably induces nutritional ketosis — blood ketones typically 0.5–3.0 mmol/L within 2–4 days. This is the tier the classic keto trials use.
• Strict therapeutic keto (under 20 g/day, ~4:1 fat:protein+carb ratio). The pediatric drug-resistant epilepsy protocol. Rarely used for weight management, because adherence collapses.

Nutritional ketosis is not the same thing as diabetic ketoacidosis (DKA). Ketosis is a controlled metabolic state where the liver produces ketones (mostly beta-hydroxybutyrate) as an alternative fuel; blood ketones stay in the low single digits and blood pH stays normal. DKA is a life-threatening complication mostly seen in type 1 diabetes, where ketones climb into the double digits, blood sugar is high, and blood pH drops. The two share a word and share nothing else — which is why type 1 diabetes, and any SGLT2-inhibitor user, needs clinical supervision before running a ketogenic protocol.

The “keto flu” — fatigue, headache, cramps, brain fog in days 3–7 — is not the ketones themselves. It’s insulin dropping, kidneys dumping sodium and water, and glycogen emptying. That’s why the mineral protocol below matters more than any supplement label.

Protocols at a glance

There are five low-carb patterns that show up in the literature and in clinical practice. They differ in strictness, adherence profile, and who they fit.

Most people who succeed long-term with low-carb do not stay at the strict VLCKD tier. They spend 6–12 weeks in ketosis to reset appetite and glycemic patterns, then drift up to a moderate low-carb or Mediterranean eating pattern with strategic higher-carb days around training or social meals. A separate short-window cousin worth naming is the protein-sparing modified fast — a supervised 6–12 week, ~800 kcal, near-zero-carb-and-fat protocol built around a high protein floor for pre-bariatric liver shrink and specific intensive-course use, not for long-term low-carb living. The stricter plant-free extension of VLCKD — the carnivore diet — strips out the vegetables, berries, and legumes that whole-food keto keeps for fiber, potassium, magnesium, folate, and vitamin C; its trial base is thin (essentially one selection-biased self-report survey) and it carries specific LDL-response, colorectal-cancer-signal, kidney-load, and micronutrient risks that a defensible keto template does not.

What trials actually show

Five trials do most of the work when the honest question is “how well does low-carb hold up head-to-head?” Reading them together is the fastest way to lose faith in absolutist claims from either direction.

Two threads run through the trials. First, in trials that don’t constrain calories (DIETFITS is the cleanest), low-carb and low-fat produce very similar weight loss at 12 months, and neither wins on a per-participant basis by insulin-resistance status or genotype. Second, in metabolic-ward trials that hold calories constant (Hall 2016), swapping carbs for fat isocalorically produces essentially no independent fat-loss advantage. Low-carb does deliver a bigger early scale drop from water and glycogen — real, useful for motivation, but not fat.

Where low-carb genuinely does more than a matched calorie deficit is in glycemic control. The Virta Health non-randomized trial (Bhanpuri 2018) put type 2 diabetes patients on a sustained ketogenic protocol with remote clinician support: HbA1c dropped ~1.3 points at one year, most participants reduced or stopped sulfonylureas and insulin, and CV risk markers improved. That’s a bigger effect than most first-line pharmacology outside GLP-1s.

Does low-carb actually beat a calorie deficit?

The honest answer is: for weight loss alone, no. For glycemic control in T2D and for appetite suppression in certain responders, yes — but through mechanisms that ultimately still route back through a calorie deficit.

The strongest version of the “carbs cause weight gain” argument is the carbohydrate-insulin model (CIM), popularized by David Ludwig and others. It says elevated post-meal insulin from carbs shunts fat into storage and drives hunger, so cutting carbs at any calorie level unlocks fat loss. It’s a coherent hypothesis. The metabolic-ward trials that could actually falsify it — Hall 2016 being the cleanest — do not support it. Isocaloric carb-for-fat swaps do not produce meaningful fat-loss differences.

Two other mechanisms are more consistent with the data:

• Water and glycogen inflate early results. Every gram of glycogen holds ~3 g of water. Emptying a full 400–500 g of liver and muscle glycogen accounts for 3–5 lb of scale loss that isn’t fat. Trials with 4–8 week endpoints wildly overstate low-carb’s fat-loss edge for this reason.
• Appetite suppression on VLCKD is real for some people. Sumithran 2013 (Eur J Clin Nutr) documented that ketosis blunts the post-weight-loss rise in ghrelin — the hunger signal that normally rebounds after a diet. That’s a mechanistic reason some people find ketogenic patterns easier to stick with, and it may explain a large chunk of the free-living weight-loss advantage in unblinded trials.

So the practical framing is: low-carb doesn’t override a calorie surplus, but for a subset of people it makes the deficit easier to hit without tracking. That subset is real, and it is not everyone.

For a wider head-to-head against fasting, meal replacements, Mediterranean, and calorie counting, see the best diet for weight loss comparison.

Health-marker changes to expect

Low-carb and ketogenic patterns change more than the scale, and the direction of the changes is fairly consistent across trials.

• HDL: up. Almost universally in trials; often by 5–15%.
• Triglycerides: down. Often by 20–30%, sometimes more if starting values are elevated. Along with HDL, this is the strongest cardiometabolic case for low-carb.
• LDL cholesterol: variable. Population-average change is small, but individual variance is wide. A minority — often lean, insulin-sensitive “lean mass hyper-responders” — see LDL rise sharply. The National Lipid Association’s 2019 position paper on low-carb and lipids recommends monitoring, and the usual first-line fix is swapping saturated fat for monounsaturated (olive oil, avocado, nuts).
• ApoB / small-dense LDL particles: usually down. Even when LDL-C is neutral, particle size tends to shift larger and less atherogenic on low-carb — this is why ApoB is more informative than LDL alone.
• Fasting insulin and HOMA-IR: down. Consistent across trials; often the change that makes low-carb feel “right” subjectively (less crash-and-crave cycling).
• HbA1c in T2D: down. ~0.5–1.3 point drops in the well-designed trials, larger on strict ketogenic protocols with medication de-prescribing (Virta / Bhanpuri 2018).
• Blood pressure: modestly down. Especially in the first 8–12 weeks; some of the drop is diuretic and reverses if sodium is under-supplemented too aggressively.

Practical rule: recheck a full lipid panel and fasting metabolic panel about 12 weeks after the eating pattern stabilizes — early labs pulled in weeks 2–4 catch the water-and-flux period rather than the steady state. If ApoB or LDL rise significantly, the swap-to-mono-fat move usually resolves it; if it doesn’t, moving to a moderate low-carb or Mediterranean template is a reasonable step. Deeper coverage of the LDL-response conversation is in cholesterol and weight loss.

Who low-carb / keto fits best

Four profiles are well-supported by the trial literature:

• Type 2 diabetes or prediabetes needing glycemic control. The largest independent benefit outside of a matched calorie deficit shows up here — HbA1c drops, medication reductions, and blood-glucose smoothing. See type 2 diabetes and weight loss for the medication-adjustment context.
• PCOS with insulin resistance. Insulin is the metabolic hinge in a large fraction of PCOS presentations. Reducing carbs typically reduces insulin more than a matched-calorie moderate-carb diet, and small trials show improved ovulation and menstrual regularity.
• Strong “appetite responders.” People who report that protein and fat leave them full for hours while carbs leave them hungry an hour later. VLCKD’s ghrelin-blunting effect is a plausible mechanism, and self-selected adherence tends to be high.
• Migraine or drug-resistant epilepsy. Therapeutic ketogenic protocols have a long track record in pediatric drug-resistant epilepsy and a growing case in migraine prophylaxis. Weight loss is a side effect, not the target.

Who should not do keto

• Pregnancy or breastfeeding. Nutritional demands are too high, and there is not enough safety data.
• Type 1 diabetes without endocrine supervision. Real DKA risk if insulin dosing is mistimed, especially in early adaptation.
• History of eating disorders. Rigid food rules are a common relapse pathway — including for people who previously restricted through calorie counting.
• History of pancreatitis. Very high fat intake is a plausible trigger; a moderate low-carb pattern is safer if low-carb is the goal.
• Athletes needing glycogen-dependent power output during the 4–8 week adaptation window. Sprints, top-end lifting, and repeated high-intensity intervals suffer measurably until fat-adaptation stabilizes; a targeted or cyclical keto is the usual workaround.
• Anyone on SGLT2 inhibitors without clinician oversight — euglycemic DKA is a documented risk.

4-week starter protocol

If you decide low-carb or keto fits, easing in reduces the sodium crash, avoids most of the keto flu, and gives you a real data point on whether the pattern actually reduces your total intake or just shifts calories.

• Week 1 — cut sugar and refined starches (about 150 g carbs → about 100 g). Bread, pasta, rice, cereal, sweet drinks, most desserts go. Vegetables, whole fruit, beans, and yogurt stay. Expect the first hunger and fatigue signals, and add 500 mL of water above your normal baseline. This week is where most people quietly bail; make sure the kitchen is stocked before you start.
• Week 2 — drop to about 50 g carbs, hit the mineral targets. This is the ketosis-onset week. Add 1–2 g extra sodium (broth, salted foods, or 1/4–1/2 tsp of salt across meals), 3–4 g potassium from whole foods (avocado, leafy greens, salmon), and 300–400 mg of magnesium. Skip the mineral load and the keto flu will hit hard. For the full sodium + potassium + magnesium replacement protocol that prevents most keto flu symptoms, see hydration and electrolytes for weight loss.
• Week 3 — stabilize protein at 1.6 g/kg lean body mass. Prioritize whole-food fat (olive oil, nuts, salmon, avocado, dairy if tolerated) over processed keto snacks. This is when appetite-suppression usually clicks in for responders; if it hasn’t after 10–14 days at under 50 g, VLCKD may not be the right template.
• Week 4 — reassess. Pull a full lipid panel plus fasting glucose and HbA1c if you started with metabolic concerns. Look at scale trend, energy, cravings, sleep, and workout performance. Three honest exits: continue VLCKD if it’s working; drift to a moderate low-carb (100 g) if energy and workouts are suffering; or slide to a Mediterranean-style low-carb hybrid for long-term maintenance.

Two rules make the ramp safer. First, keep hydration and electrolytes running — most “keto isn’t working” complaints in weeks 2–4 are undersalting, not the diet failing. Second, plan the exit before you start. A ketogenic pattern with no re-entry template usually ends in a rebound eating pattern that undoes the loss.

For coordinated calorie math while you’re doing this, see TDEE and calorie deficit for beginners. If you’re pairing low-carb with time-restricted eating, intermittent fasting covers the window mechanics without duplicating the food-choice guidance here.

What “low-carb done wrong” looks like

Most low-carb failures fall into four patterns that trials and clinical practice keep flagging:

• Unlimited processed meats, zero vegetables. Technically ketogenic, functionally a fiber, potassium, and antioxidant crater. Constipation is the near-guaranteed early consequence; long-term, colorectal risk becomes the concern the American Institute for Cancer Research position papers raise.
• Loading dairy fats and forgetting fiber. Butter coffee, cheese-heavy plates, and cream-based sauces are appealing on VLCKD and easy to over-eat. Fiber intake often crashes to under 10 g/day; the digestive-tract cost shows up in weeks 2–3.
• “Dirty keto” bar-and-shake patterns. Packaged keto bars, shakes, and low-carb tortillas hit macros on paper but often deliver high sodium, seed oils, artificial sweeteners — often heavy on erythritol, which carries the Witkowski 2023 CV signal at higher daily intake — and a micronutrient profile no whole-food eating pattern would produce. Trials that improved cardiometabolic markers used whole-food templates, not packaged ones.
• No re-entry plan. Abandoning VLCKD without ramping carbs back gradually causes a 3–5 lb water and glycogen rebound in the first two weeks, which reads as “the diet stopped working” and often triggers a full return to pre-diet eating. Reintroducing whole-food carbs at ~20–30 g per week protects against this.

Bottom line

Low-carb and ketogenic diets are legitimate tools for the people they fit — especially for glycemic control in type 2 diabetes, PCOS with insulin resistance, and strong appetite responders. They do not outperform a matched calorie deficit for pure weight loss in 12-month head-to-head trials (DIETFITS is the cleanest example). Use them if the eating pattern is sustainable for you, run the sodium and potassium protocol during induction, recheck lipids at 12 weeks, and plan the exit before you start.

Sources at a glance

• Shai I, et al. Weight loss with a low-carbohydrate, Mediterranean, or low-fat diet (DIRECT). New England Journal of Medicine (2008).
• Gardner CD, et al. Effect of low-fat vs low-carbohydrate diet on 12-month weight loss in overweight adults (DIETFITS). JAMA (2018).
• Gardner CD, et al. Comparison of the Atkins, Zone, Ornish, and LEARN diets for change in weight and related risk factors (A TO Z). JAMA (2007).
• Bazzano LA, et al. Effects of low-carbohydrate and low-fat diets: a randomized trial. Annals of Internal Medicine (2014).
• Hall KD, et al. Energy expenditure and body composition changes after an isocaloric ketogenic diet in overweight and obese men. American Journal of Clinical Nutrition (2016).
• Bhanpuri NH, et al. Cardiovascular disease risk factor responses to a type 2 diabetes care model including nutritional ketosis (Virta Health). Cardiovascular Diabetology (2018).
• Sumithran P, et al. Ketosis and appetite-mediating nutrients and hormones after weight loss. European Journal of Clinical Nutrition (2013).
• Kirkpatrick CF, et al. Review of current evidence and clinical recommendations on the effects of low-carbohydrate and very-low-carbohydrate diets on cardiovascular disease risk (NLA position paper). Journal of Clinical Lipidology (2019).

Frequently asked questions

What’s the difference between low-carb and keto? Low-carb is a spectrum — moderate low-carb runs 100–150 g of carbs a day, standard low-carb 50–100 g, and very-low-carb ketogenic under 50 g. Only the very-low-carb tier reliably induces nutritional ketosis, which typically starts 2–4 days after carbs drop below ~50 g. All three cut carbs, but only ketogenic patterns shift the body’s primary fuel from glucose to ketones.

How long does it take to get into ketosis? Most people reach nutritional ketosis 2–4 days after dropping to under 50 g of carbs per day, once liver glycogen is depleted. Blood ketones typically read 0.5–3.0 mmol/L on the ketogenic range. Exercise, an overnight fast, or MCT oil can shorten that window; a lingering evening snack or hidden carbs in sauces and drinks can push it out a full week.

Do low-carb diets actually work better than calorie counting? Not at 12 months. DIETFITS (Gardner 2018, JAMA, 609 adults, 12 months) found low-carb and low-fat arms lost equivalent weight (~5–6 kg both arms). DIRECT (Shai 2008, NEJM, 2 years) showed a small low-carb edge that narrowed by year two. Low-carb wins the early sprint from water-weight loss, but the long-run delta versus a matched calorie deficit is small in trials.

What is keto flu and how do I prevent it? Keto flu is the fatigue, headache, brain fog, and cramping that peaks days 3–7 as insulin drops, kidneys dump sodium and water, and glycogen empties. Prevent it by adding 1–2 g of extra sodium a day (broth, salted foods), 3–4 g of potassium from whole foods (avocado, leafy greens, salmon), and 300–400 mg of magnesium. Hydration alone is not enough — the mineral load is what blunts it.

Will keto raise my cholesterol? For most people, HDL rises and triglycerides fall, and LDL is variable. A subset — often lean, insulin-sensitive people the community calls “lean mass hyper-responders” — see LDL climb sharply on high-saturated-fat keto. The usual fix is swapping saturated fat for monounsaturated (olive oil, nuts, avocado). Recheck a full lipid panel 12 weeks after the pattern stabilizes; ApoB is more informative than LDL alone.

Can I do low-carb while still eating fruit? Yes, at the moderate and standard low-carb tiers. Berries (about 5–8 g net carbs per half cup) fit even a 50 g ceiling; higher-sugar fruits like bananas, grapes, and mango push most of a ketogenic budget in one serving. On a strict ketogenic pattern, most people limit fruit to a handful of berries per day and treat other fruit as an occasional carb refeed rather than a daily food.

Is keto safe for people with type 2 diabetes? For many people with type 2 diabetes, low-carb and ketogenic patterns are among the most effective non-drug interventions for glycemic control — the Virta Health trial (Bhanpuri 2018, Cardiovasc Diabetol) showed HbA1c dropping ~1.3 points and most participants reducing or stopping medications at one year. It requires clinician supervision, because insulin and sulfonylurea doses often need to come down quickly to prevent hypoglycemia.

How do I come off a keto diet without gaining the weight back? Reintroduce carbs slowly — roughly 20–30 g extra per week, prioritizing whole foods (oats, sweet potato, beans, fruit) over refined ones. Expect 2–5 lb of water and glycogen weight to return in the first two weeks; that’s not fat regain. Keep protein at 1.6 g/kg lean mass, keep resistance training, and settle at the lowest carb tier that keeps hunger and energy stable — often a moderate low-carb or Mediterranean-style hybrid.

How this compares to other options

• Compared with the Mediterranean eating pattern, low-carb hits glycemic markers faster but has weaker long-term cardiovascular evidence — a Mediterranean-style low-carb hybrid is often the durable landing spot.
• Compared with intermittent fasting, low-carb changes what you eat while fasting changes when; stacking the two is common and works well for appetite responders.
• Compared with calorie-restricted eating, low-carb removes the tracking burden for many people, but head-to-head trials show equivalent 12-month weight loss when adherence is matched.
• If you’re still deciding between approaches, our best diet for weight loss comparison covers keto against Mediterranean, fasting, calorie counting, and meal replacements side by side.
• For a time-boxed, elimination-based framework rather than a macro-defined one, the Whole30 diet for weight loss guide covers a 30-day whole-food reset with a structured reintroduction phase. For the ancestral-elimination framework that is lower-carb than a standard diet but not ketogenic — grains and legumes out, fruit and starchy tubers in — see our paleo diet for weight loss guide.