Low-carb can mean 20 grams or 200, and what replaces the carbs matters most

Key takeaways

• One widely used scheme calls under 130 grams of carbohydrate a day “low-carb,” while the American Diabetes Association uses the label for 26% to 45% of calories, which can top 200 grams on a 2,000-calorie day²³.
• The ADA’s 2019 consensus report found that cutting overall carbohydrate has the most evidence for improving blood sugar in diabetes².
• Summaries of a 2021 BMJ review of 23 trials report about 32 more diabetes remissions per 100 people on low-carb diets at six months, with the gains fading by a year⁵⁶.
• In data on more than 432,000 people, low-carb eating built on animal fat and protein was tied to higher death rates, and plant-based low-carb eating to lower rates¹¹.
• If you take diabetes or blood pressure medicine, or are pregnant or trying to conceive, talk with a clinician before you cut carbs sharply.

In 2022, a company called Virta Health asked a federal agency to review the evidence on low-carb and very low-carb diets for type 2 diabetes and prediabetes¹. Virta wanted the findings to help win endorsements from federal agencies¹.

The Agency for Healthcare Research and Quality, or AHRQ, turned the request down¹. Its reasons are revealing. Ten recent systematic reviews had already tackled the question, the agency found, and its review program takes no sides between competing treatments¹. A systematic review gathers the studies on a question and weighs them together.

So the research exists, and there’s plenty of it. The harder problem is that “low-carb” means very different things, from skipping soda to living near ketosis.

This guide sorts out the definitions, what trials show for blood sugar, the long fight over carbs and insulin, and what decades of data say about the foods that replace the carbs. Our companion article covers the strictest end, the ketogenic diet.

“Low-carb” has two common scorecards, and they disagree

A 2015 scheme by Feinman and colleagues, used in a 2019 guide for diabetes nurses, sorts diets by grams a day on a 2,000-calorie diet². Very low-carb, or ketogenic, means 20 to 50 grams. Low-carb means under 130 grams, or under 26% of calories².

In that scheme, moderate runs from 130 to 225 grams, and anything above is high². AHRQ used the same 130-gram line for low-carb and put very low-carb under 50 grams¹.

The ADA’s consumer guidance draws the lines differently. It calls 26% to 45% of calories a low-carb pattern³. It saves “very low-carb” for less than 26%, often 20 to 50 grams of carbohydrate not counting fiber³.

So a day with 180 grams of carbohydrate is “moderate” in one scheme and “low-carb” in the other. Individual studies set their own cutoffs, too, which makes them hard to compare, that guide noted².

When a headline says “low-carb,” look for the grams. A finding about 40 grams a day may say little about 160.

For blood sugar, cutting carbs has the deepest evidence

In its 2019 consensus report, the ADA concluded that reducing overall carbohydrate had the most evidence for improving blood sugar in people with diabetes². It called low-carb and very low-carb eating a reasonable option for some, especially people not meeting blood sugar goals or hoping to cut back on diabetes medicine².

The same report found long-term benefit from very low-carb diets unconvincing, partly because people struggle to stick with them².

The ADA’s current consumer pages list low-carb among seven meal patterns and say no single plan suits everyone³. Its Standards of Care for 2026, released in December 2025, name Mediterranean-style and low-carb patterns as having evidence for preventing type 2 diabetes⁴.

A major test of remission came in 2021. A BMJ review by Goldenberg and colleagues pooled 23 randomized trials with 1,357 adults with type 2 diabetes⁵⁶. In a randomized trial, people are assigned by chance to one diet or another. The figures below come from summaries in the journals Diabetes & Primary Care and AJMC⁵⁶.

At six months, low-carb diets produced about 32 more remissions per 100 people than control diets⁵⁶. Remission mostly meant an A1C blood test below 6.5%, whether or not people still took medicine⁵. Counting only people off diabetes medicine, the difference was small and not statistically significant⁵⁶.

By 12 months, the data were sparse and the advantage had faded⁵⁶. The authors said clinicians might consider short-term low-carb diets, while flagging safety questions and the chance that eating fewer calories explained part of the effect⁶.

Two details matter for your choices. In 18 of the 23 trials, the comparison group ate a low-fat diet, so the review says little about how low-carb stacks up against, say, a Mediterranean diet⁵. And very low-carb trials had lower remission rates than more moderate ones, which the authors tied to people not sticking with the stricter plans⁵.

Do carbs and insulin drive weight gain? Both camps are serious

The carbohydrate–insulin model says fast-digesting carbs set off hormone responses, chiefly insulin, that push the body to store fat. In this view, fat storage comes first and pushes the body to take in more energy than it burns, reversing the usual cause and effect⁹.

In 2021, 17 authors led by David Ludwig laid out the model in the American Journal of Clinical Nutrition⁹. Among them were Walter Willett, Arne Astrup and Gary Taubes⁹.

In 2022, 12 researchers led by Kevin Hall of the National Institutes of Health replied in the same journal¹⁰. Their energy balance model puts the brain in charge of body weight, steering how much we eat through hormone, metabolic and nerve signals, mostly below awareness¹⁰.

Hall’s group argued that the newest carbohydrate-insulin model had dropped a formerly central claim, that fat buildup is the main driver of weight gain¹⁰. In their view, it is now a narrower special case of their own model, focused on diets high in fast-digesting carbs¹⁰.

The best-known test came from Framingham State University in Massachusetts. In a 2018 BMJ trial, 164 adults first lost about 12% of their weight⁷. They were then randomly assigned to diets with 20%, 40% or 60% of calories from carbohydrate for 20 weeks while holding their weight steady⁷.

The low-carb group burned about 209 more calories a day than the high-carb group in the main analysis⁷. Among people who released the most insulin, the gap looked bigger still⁷. The authors cautioned that the study couldn’t prove the carb-to-fat ratio alone caused the change⁷.

Hall and a colleague then reran the numbers using the analysis plan first registered for the trial⁸. The Framingham team had switched plans before results were unmasked, calling the registered version an error⁷⁸.

Under the original plan, the gap shrank to roughly 100 calories a day and was no longer statistically significant, and neither was the extra effect in people who released the most insulin⁸. The reanalysis was posted in 2018 on bioRxiv, a site for preprints, which are papers shared before peer review; its page lists no journal version⁸.

Where does that leave you? With a real but contested signal. The theory question is open, and our keto article describes an NIH inpatient study in which people ate less on a high-carb, low-fat diet. Meanwhile, the long-term data below suggest that food quality matters under either model.

Over decades, what replaces the carbs matters more than the cut

The Atherosclerosis Risk in Communities study, called ARIC, enrolled 15,428 adults ages 45 to 64 in four U.S. communities from 1987 to 1989¹¹. Over a median of 25 years, 6,283 died¹¹.

Death rates traced a U shape. Risk was lowest when carbohydrates supplied 50% to 55% of calories, and higher at both the low and high ends¹¹.

The team then pooled ARIC with seven other cohorts, covering 432,179 people¹¹. Compared with moderate intake, low carb intake was tied to 20% higher death risk and high intake to 23% higher¹¹. The NIH funded the work¹¹.

The source of the replacement mattered most. Swapping carbs for animal fat and protein was linked to 18% higher risk. Swapping them for plant fat and protein was linked to 18% lower risk¹¹.

Harvard researchers saw a similar split in 2020, using national survey data on 37,233 U.S. adults¹³. Scored without regard to food quality, low-carb eating showed no link to death rates. Once quality counted, the picture changed¹³.

Compared with people who didn’t eat low-carb, those with the best-quality versions had a 27% lower risk of dying early, and those with the poorest had a 16% higher risk¹³. The better versions leaned on nuts, whole fruit, nonstarchy vegetables and whole grains. The worse ones were heavy in low-quality carbs like white bread, animal protein and saturated fat, with red meat and butter among the examples¹³.

These are observational studies. They find links in large groups but can’t prove that diet caused the deaths. Commenting to the cardiology news site TCTMD, Andrew Mente of the Population Health Research Institute in Ontario said that nutrient-level analyses blur good and poor carbohydrate sources, and he urged caution on extreme diets¹².

Planning a pregnancy? Folic acid travels with bread and pasta

In the United States, folic acid is added to staples such as bread and pasta¹⁴. That matters for neural tube defects, birth defects such as anencephaly and spina bifida¹⁴.

A University of North Carolina team looked for a link in 2018, using data gathered from 1998 to 2011 in nine states by the National Birth Defects Prevention Study¹⁴. The team set 1,740 women whose pregnancies were affected by anencephaly or spina bifida against 9,545 women whose babies had no birth defects¹⁴.

Restricting carbs went with a 30% higher chance of a neural tube defect¹⁴. The carb-restricting women also got under half as much folic acid from food as the rest, a likely reason for the link, the researchers suggested¹⁴. The CDC paid for the work, and the team called for studies in other populations to confirm it¹⁴.

If you are trying to conceive while cutting carbs, ask your clinician about folic acid.

The low-carb spectrum, from cutting soda to near-keto

This table is our own, built from the sources in this article. Gram ranges assume about 2,000 calories a day. Sample days are examples, not prescriptions.

Level	Carbs a day	A sample day	What the evidence says	Cautions
Quality swap	Over 225 g, over 45% of calories2	Oatmeal with fruit; bean and grain bowl; fish, vegetables and brown rice; water instead of soda	The very highest carb intakes were tied to higher death risk in pooled cohorts11	Few for most people
Moderate (Feinman) or “low-carb” (ADA)	130 to 225 g, 26% to 45%23	Eggs and greens with one slice of whole-grain toast; lentil soup and salad; tofu or chicken with roasted vegetables and a small serving of grain; berries	The ADA lists it for lowering A1C, weight, blood pressure and triglycerides3	Diabetes medicines may need adjusting2
Low (Feinman) or “very low” (ADA)	Under 130 g, under 26%123	Plain yogurt with nuts and berries; big salad with salmon or tempeh and olive oil; stir-fried tofu and vegetables with a few beans	Remission gains in type 2 diabetes at six months, fading by a year56	Medicines; folic acid before pregnancy214
Very low (ketogenic)	20 to 50 g23	Mostly nonstarchy vegetables, protein foods and fats	See our keto article	Clinician check first for anyone on diabetes medicine2

A plant-forward low-carb grocery list

This list leans on what the healthier low-carb eaters in long-term studies chose: plant fats and proteins, nonstarchy vegetables, whole fruit and nuts¹¹¹³.

• Nonstarchy vegetables: leafy greens, broccoli, cauliflower, cabbage, zucchini, peppers, mushrooms and green beans. The ADA highlights these in low-carb plans³.
• Plant proteins: tofu, tempeh and edamame, plus lentils and beans in measured portions on moderate plans.
• Nuts and seeds: almonds, walnuts, peanuts, pumpkin seeds and chia, and nut butters with no added sugar.
• Healthy fats: extra-virgin olive oil, avocado and olives.
• Animal foods, if you eat them: fish, eggs, plain yogurt and poultry. Go easy on red meat and butter, the saturated-fat sources Harvard’s researchers flagged¹³.
• Whole fruit: berries first, then apples, citrus and other whole fruits as your carb budget allows.
• Whole grains in small portions: oats, barley, quinoa and whole-grain bread, especially on moderate plans¹³.
• Skip or limit: soda, sweets, white bread, sugary cereal and packaged “low-carb” snack bars.

This list fits values many natural-health readers already hold: whole foods, plants and home cooking. In the long-term data, low-carb eating looked best when it pushed out refined grains and sugary foods, not when red meat and butter filled the gap¹¹¹³.

Bring the same scrutiny to low-carb businesses that you’d bring to big food. The company behind the AHRQ request wanted a federal endorsement¹, and a “low-carb” snack is still a packaged snack.

How strong is the evidence?

Claim	Evidence level	What the research shows
Cutting carbs improves blood sugar in type 2 diabetes	Strong evidence	ADA consensus; remission gains at six months256
Low-carb puts diabetes into lasting remission without medicine	Not enough evidence	No significant gain off medicine; little data past a year56
Low-carb diets make you burn extra calories	Promising but early	One trial found it; a reanalysis using the original plan didn’t78
Carbs and insulin are the main cause of obesity	Not enough evidence	Leading research groups still disagree910
What replaces carbs shapes long-term health	Promising but early	Plant swaps linked to lower death rates, animal swaps to higher, in observational data1113
Very low carb intake is fine when planning a pregnancy	Not enough evidence	One large study linked low intake to neural tube defects14

Talk with a clinician first if any of these apply

• You take insulin, a sulfonylurea or other diabetes medicine. The ADA advises expert input for people on glucose-lowering drugs, with regular monitoring and dose changes as needed². It also urges working with a knowledgeable clinician from the start to prevent dehydration and to lower insulin and other medicines safely¹. Our article on medicines and wellness goals explains what that involves.
• You take blood pressure medicine. The ADA counts lower blood pressure among the goals of low-carb eating³. Ask how to track your readings as your diet changes.
• You’re pregnant or trying to conceive. The ADA consensus said more evidence is needed before recommending low-carb eating in pregnancy². Ask about folic acid¹⁴.
• You have kidney disease. More evidence is needed here, too, the ADA said².
• You’ve had an eating disorder. The ADA flagged disordered eating as another gap in the evidence². If restriction or weight loss is becoming a struggle, call the National Alliance for Eating Disorders helpline at 1-866-662-1235. If you’re in crisis, call or text 988.

Questions readers ask

The bottom line

“Low-carb” is a range, not a rule, and the grams matter. For blood sugar, moderate carb cutting has the deepest support, though early remission gains tend to fade and stricter plans are harder to keep.

The carbs-and-insulin debate is still open among serious scientists. What decades of data suggest is simpler: when you cut carbs, fill the space with vegetables, nuts, olive oil, beans and fish rather than red meat and butter. If you take diabetes or blood pressure medicine, or are planning a pregnancy, make the plan with your clinician.

Related reading

• Keto is real medicine for epilepsy, but its edge for weight loss is small
• Some healthy habits clash with your medicines, so check these goals first
• Fiber, form and a short walk shape how carbs affect your blood sugar
• Your response to food is personal, but the payoff from diet algorithms is modest
• Why observational studies got smoking right and hormone therapy wrong

Sources

1. Agency for Healthcare Research and Quality. “Topic Brief: Low and Very Low Carbohydrate Diets for People with Diabetes Type 2 and Pre-diabetes.” AHRQ Effective Health Care Program, September 6, 2022. effectivehealthcare.ahrq.gov
2. Jones K, McArdle P. “The ADA nutrition therapy consensus report: A quick guide.” Journal of Diabetes Nursing, 2019. diabetesonthenet.com
3. American Diabetes Association. “Eating for Diabetes Management.” diabetes.org, accessed October 2026. diabetes.org
4. American Diabetes Association. “American Diabetes Association Releases Standards of Care in Diabetes—2026.” Press release, December 8, 2025. diabetes.org
5. Brown P. “Diabetes Distilled: Efficacy and safety of low and very low carbohydrate diets for type 2 diabetes remission” (on Goldenberg JZ, et al., BMJ, 2021). Diabetes & Primary Care, February 16, 2021. diabetesonthenet.com
6. Jeremias S. “Low-Carb Diet May Help Achieve Type 2 Diabetes Remission, Study Finds” (on Goldenberg JZ, et al., BMJ, 2021). AJMC, April 1, 2021. ajmc.com
7. Ebbeling CB, Feldman HA, Klein GL, et al. “Effects of a low carbohydrate diet on energy expenditure during weight loss maintenance: randomized trial.” BMJ, 2018. bmj.com
8. Hall KD, Guo J. “No Significant Effect of Dietary Carbohydrate versus Fat on the Reduction in Total Energy Expenditure During Maintenance of Lost Weight: A Secondary Analysis.” bioRxiv preprint, November 28, 2018. biorxiv.org
9. Ludwig DS, Aronne LJ, Astrup A, et al. “The carbohydrate-insulin model: a physiological perspective on the obesity pandemic.” American Journal of Clinical Nutrition, 2021. ovid.com
10. Hall KD, Farooqi IS, Friedman JM, et al. “The energy balance model of obesity: beyond calories in, calories out.” American Journal of Clinical Nutrition, 2022 (Washington University record). profiles.wustl.edu
11. Seidelmann SB, Claggett B, Cheng S, et al. “Dietary carbohydrate intake and mortality: a prospective cohort study and meta-analysis.” Lancet Public Health, 2018 (DOAJ record). pubmed.ncbi.nlm.nih.gov
12. TCTMD. “Both High- and Low-Carb Diets Increase Long-term Risk of Mortality: ARIC Analysis.” August 30, 2018. tctmd.com
13. Harvard T.H. Chan School of Public Health. “Healthy low-carbohydrate and low-fat diets may reduce risk of premature death” (on Shan Z, et al., JAMA Internal Medicine, 2020). January 21, 2020. hsph.harvard.edu
14. UNC Gillings School of Global Public Health. “UNC study links low carbohydrate intake to increased risk of birth defects” (on Desrosiers TA, et al., Birth Defects Research, 2018). January 25, 2018. sph.unc.edu