Low Carbohydrate Nutrition

Banting and the 1863 Letter on Corpulence

William Banting's pamphlet Letter on Corpulence, Addressed to the Public is dated May 1863, and the first edition ran to 1,000 copies. A fifth edition published in New York in 1865 includes remarks by the surgeon William Harvey, who advised Banting. Banting reported avoiding bread, butter, milk, sugar, beer and potatoes and eating meat, poultry, fish and most non-starchy vegetables. He recorded a fall from 202 lb in August 1862 to 156 lb in September 1863. These figures are one man's own account and are not trial evidence.

A 2026 article in Medical Humanities reads Banting's 1863 memoir as a remission narrative, placing it beside a 2020 memoir in a tradition of redemptive storytelling about type 2 diabetes and obesity. It is a literary and cultural analysis and does not assess the diet itself.

Atkins and the debate of the 1970s

In 1973 the American Medical Association's Council on Foods and Nutrition published a review of Dr. Atkins' Diet Revolution, a 1972 book. The review calls the low carbohydrate approach to weight reduction neither new nor innovative, traces it to the diet devised by Harvey and recommended in Banting's Letter, and notes a cyclical return of similar diets over the preceding twenty years. Only the abstract of the review was open to read for this page, so its conclusions are not summarized here.

Trial-era interest and popular labels

A 2003 trial report states that, despite the popularity of the low carbohydrate, high-protein, high-fat Atkins diet, no randomized controlled trial had evaluated it before that study. Later trials compared Atkins and other named diets directly. The sources reviewed for this page do not describe the origins of South Beach or of low-carb high-fat as popular labels, so no history is given for them here.

No single definition

A 2021 narrative review states that there is no universally agreed definition of a low carbohydrate diet. It uses a threshold of under 130 g per day, also expressed as 26 percent of total energy, and treats very low carbohydrate approaches as a stricter category. The term describes a range of eating patterns, not one standard diet, so results from one study do not automatically apply to another.

The American Diabetes Association's professional pages describe a low carbohydrate pattern in the 2019 consensus as 26 to 45 percent of total calories and a very low carbohydrate pattern as under 26 percent. The 2026 Standards of Care note that the wide range of definitions is an important challenge in interpreting carbohydrate-restricted research.

Definitions used in two major reviews

A 2021 BMJ meta-analysis defined low carbohydrate diets as under 130 g per day or under 26 percent of a 2,000 kcal diet, and very low carbohydrate diets as under 10 percent of calories. A 2022 Cochrane review compared against balanced-carbohydrate diets of 45 to 65 percent of energy, and classed low carbohydrate as above 50 g to 150 g per day or under 45 percent, and very low as 50 g per day or less, or under 10 percent. The same word carries different cut points in these reviews and in the American Diabetes Association pages, and these figures describe study categories, not recommendations.

Related Machipedia pages

This page belongs to the Machipedia family page Dietary Approaches. It covers low carbohydrate patterns that are not ketogenic by definition. The Ketogenic Diet page covers patterns built to induce ketosis, and the 2021 narrative review notes that the term ketogenic is sometimes used for very low carbohydrate approaches. The Paleo Diet and Carnivore Diet have their own pages and are not covered here. Dietetics and Nutritional Therapy are separate pages on the professions that may advise on eating patterns.

How trials delivered low carbohydrate interventions

Trials differ in how much support they give. A 2003 one-year trial of 63 obese adults used minimal professional contact to mimic how most people diet. In the 2007 A TO Z trial, 311 overweight premenopausal women received weekly instruction for two months and then ten months of follow-up. In the 2018 DIETFITS trial, health educators delivered 22 small-group sessions over 12 months to adults assigned to a low carbohydrate or low-fat diet. Reported carbohydrate intake at 12 months averaged 30 percent of energy in the low carbohydrate arm.

This page does not give meal plans, targets or step-by-step instructions. Adherence is a recurring issue in the sources: the 2003 trial reported poor adherence and high attrition in both diet groups, and the 2026 Standards of Care say studies generally indicate challenges with long-term sustainability.

Who advises on it

The American Diabetes Association states that health care professionals should maintain consistent medical oversight of people following very low carbohydrate eating plans. In the United States, Registered Dietitian Nutritionists are credentialed through the Commission on Dietetic Registration, the credentialing agency of the Academy of Nutrition and Dietetics, and the Academy describes a minimum of a graduate degree, supervised practice and a national examination as requirements. The sources reviewed do not describe training or registers for other jurisdictions or for non-dietitian nutrition practitioners.

Weight management

Low carbohydrate diets continue to be widely promoted and marketed as more effective for weight loss, and healthier, than balanced-carbohydrate diets. The Cochrane review treated this as a claim to test and stated that debates about effective and safe diets for managing obesity in adults are ongoing.

Type 2 diabetes and related measures

In the American Diabetes Association's 2019 consensus summary, the listed expected effects of low carbohydrate and very low carbohydrate patterns include lower A1C, weight loss, lower blood pressure, higher HDL cholesterol and lower triglycerides. Diabetes UK states that low carbohydrate diets can help with weight loss, glucose management and cardiovascular risk factors in some adults with type 2 diabetes in the short term. These are statements of expected effects drawn from trials and reviews, and the sections below report what the trials found.

Weight loss: Cochrane review

A 2022 Cochrane review included 61 parallel randomized trials with 6,925 participants, searched to June 2021. In adults without type 2 diabetes, low carbohydrate diets probably made little to no difference compared with balanced-carbohydrate diets: a mean difference of -1.07 kg at three to 8.5 months (37 trials, 3,286 participants) and -0.93 kg at one to two years (14 trials, 1,805 participants), both moderate certainty.

In adults with type 2 diabetes the differences were -1.26 kg at three to six months (14 trials, 1,114 participants) and -0.33 kg at one to two years (7 trials, 813 participants), also moderate certainty. Risk of bias was mostly high, and adverse-effect evidence was too limited for conclusions.

Individual trials

In the 2003 trial, participants on a low carbohydrate diet lost more weight than those on a conventional diet at three months (6.8 versus 2.7 percent of body weight) and six months (7.0 versus 3.2 percent), but not significantly more at 12 months (4.4 versus 2.5 percent). HDL cholesterol rose and triglycerides fell more with the low carbohydrate diet. Adherence was poor and attrition high in both groups.

The 2007 A TO Z trial randomized 311 premenopausal women to four diets; mean 12-month loss was 4.7 kg for Atkins, 1.6 kg for Zone, 2.6 kg for LEARN and 2.2 kg for Ornish, and only the Atkins versus Zone difference was significant. In the 2018 DIETFITS trial of 609 adults without diabetes, 12-month change was -5.3 kg on the low-fat diet and -6.0 kg on the low carbohydrate diet (difference 0.7 kg, 95% CI -0.2 to 1.6).

Type 2 diabetes

A 2021 BMJ meta-analysis of 23 randomized trials (1,357 participants) found that at six months low carbohydrate diets gave higher rates of remission defined as HbA1c under 6.5 percent (57 percent versus 31 percent; 8 studies, 264 participants). The effect was smaller and not significant when remission excluded medication use, and it fell in studies that included people using insulin. Data at 12 months were sparse, and weight and triglyceride gains at six months diminished. The authors rated the evidence moderate to low certainty.

The 2026 American Diabetes Association Standards of Care report that carbohydrate-restricted patterns reduced A1C in the short term (under six months), with less difference beyond one year. They cite a meta-analysis of six trials of at least 12 months (524 participants) in which a low carbohydrate pattern was beneficial for lipids but not for glycemic management.

Long-term cohort data

A 2018 analysis followed 15,428 US adults for a median of 25 years and pooled it with seven multinational cohorts (432,179 participants). It found a U-shaped association between carbohydrate share of energy and all-cause mortality, lowest at 50 to 55 percent; the pooled hazard ratio for intake under 40 percent was 1.20 (95% CI 1.09 to 1.32). Substituting animal-derived fat or protein for carbohydrate was associated with higher mortality and plant-derived with lower. The data are observational and questionnaire-based, so they cannot show cause.

Diabetes UK and the British Dietetic Association

Diabetes UK's May 2021 position statement, prepared with the Scientific Advisory Committee on Nutrition, recommends lower carbohydrate diets as a short-term effective option for adults with type 2 diabetes who have overweight or obesity, including diets providing 50 to 130 g a day. It states there is no consistent evidence that they are more effective than other approaches in the long term, so they should not be seen as the diet for everyone.

A 2021 review reports that a 2018 British Dietetic Association policy statement supports carbohydrate restriction as a viable option for adults with type 2 diabetes, with caveats about long-term effects. That is a second-hand description, and the statement itself was not opened. The same review discloses that several of its authors work for a charity that provides diabetes education including information on low carbohydrate diets.

American Diabetes Association and Dietary Guidelines

The American Diabetes Association's 2026 Standards of Care include low carbohydrate and very low carbohydrate patterns within medical nutrition therapy for diabetes. They advise regular reassessment and individualization for people who choose this approach and emphasize high-quality, minimally processed, high-fiber carbohydrate sources whatever the amount.

The Dietary Guidelines for Americans, 2020 to 2025 give an Acceptable Macronutrient Distribution Range for carbohydrate of 45 to 65 percent of calories for adults and do not mention low carbohydrate diets by name. A newer 2025 to 2030 edition exists and was not reviewed. This page does not report NICE guidance, which could not be opened in a readable form.

Medicines and medical oversight

The 2026 Standards of Care state that insulin and other diabetes medications may need adjusting on very low carbohydrate plans to prevent hypoglycemia, and that blood pressure needs monitoring. They advise that such plans be avoided in people taking SGLT2 inhibitors because of the risk of ketoacidosis, noting numerous published case reports of diabetic ketoacidosis in people with type 1 or type 2 diabetes using SGLT2 inhibitors with very low carbohydrate or ketogenic patterns.

A narrative review adds that blood pressure medicines may also need adjustment and that sulfonylurea and insulin doses are typically reviewed at the start. This should be arranged with a clinician and not attempted alone.

Groups for whom caution is advised

The American Diabetes Association states that very low carbohydrate plans are not currently recommended for people who are pregnant or lactating, children, people with kidney disease, or people with or at risk for disordered eating. Diabetes UK does not recommend low carbohydrate diets for people with type 1 diabetes, citing no strong evidence of safety or effectiveness, or for children with diabetes, citing reports of poor growth, greater cardiovascular risk and psychological problems.

Lipids and unresolved questions

Effects on cholesterol vary by outcome. The 2003 trial reported higher HDL cholesterol and lower triglycerides, while the 2021 BMJ meta-analysis reported a clinically important but not statistically significant worsening of LDL cholesterol at 12 months. The narrative review says LDL may rise in some people. Long-term safety remains uncertain: the Cochrane review found adverse-effect data too limited for conclusions, and the cohort analysis found higher mortality associated with low carbohydrate intake when animal sources replaced carbohydrate.

This page is general information and not personal advice. A change in eating pattern should not delay or replace assessment or prescribed treatment, and anyone with a medical condition or taking medicines can ask their treating clinician or a registered dietitian before changing how they eat.