Low-Carbohydrate Diet: Definitions in Grams, Trial Evidence and an Indian Lower-Carb Day

A low-carbohydrate diet reduces the starch and sugar you eat, mainly rice, roti, bread, potato, sweets and sweet drinks, and makes up part of the energy with protein and fat. There is no single agreed definition. The one used most in trials is under 130 g of carbohydrate a day, or under 26 per cent of a 2,000 kcal diet; ICMR-NIN's balanced Indian diet, by comparison, gets 50 to 55 per cent of its energy from carbohydrate. This entry sits in NNWA's handbook of diets and covers the moderate end of the range, not ketosis.

What counts as a low-carbohydrate diet?

It depends on who is counting, and that is the first thing to know before comparing one study or one diet chart with another. The American Diabetes Association's 2019 consensus report says plainly that there is no consistent definition of low carbohydrate, and sets its own bands for the purpose of its review. A 2021 meta-analysis in the BMJ and a 2022 Cochrane review each drew the lines somewhere else. The table sets them side by side.

To turn a percentage into grams, multiply your day's energy by the percentage and divide by 4, since carbohydrate gives 4 kcal a gram. The consensus report's own conversion table shows that on 2,000 kcal, 50 per cent is 250 g, 40 per cent is 200 g, 30 per cent is 150 g and 10 per cent is 50 g. On 1,500 kcal the same 40 per cent is only 150 g.

The Indian guideline: cereals at no more than 45 per cent of energy

ICMR-NIN's Dietary Guidelines for Indians, revised in 2024, do not recommend a low-carbohydrate diet. They recommend a balanced one in which no more than 45 per cent of energy comes from cereals and millets and up to 15 per cent from pulses, beans and meat, which works out to 50 to 55 per cent of energy from carbohydrate, 10 to 15 per cent from protein and 20 to 30 per cent from fat. At least half the cereal should be whole grain.

The same document records what Indians eat in practice: cereals supply 50 to 70 per cent of the day's energy, while pulses, meat, poultry and fish together supply 6 to 9 per cent against a recommended 14 per cent. So for many households the guideline is already a lower-carbohydrate plate than the one on the table. Its vegetarian model day of 2,000 kcal holds 250 g of raw cereals and about 275 g of carbohydrate in all. The carbohydrate table for Indian foods ranks the staples.

Two things stand out. Dal is about half carbohydrate by dry weight, so a vegetarian who replaces roti with extra dal has cut less than it seems. And the foods that carry almost none, such as eggs, fish, chicken, paneer, nuts and non-starchy vegetables, are the ones a lower-carb Indian day is built from.

Does a low-carbohydrate diet lead to more weight loss?

Over a year or more, about the same as a balanced diet with the same effort. The 2022 Cochrane review pooled 61 randomised trials with 6,925 adults who were overweight or living with obesity. In people without type 2 diabetes, low-carbohydrate diets gave 1.07 kg more weight loss at three to eight and a half months and 0.93 kg more at one to two years. The reviewers judged this as probably little to no difference, with moderate certainty.

The DIETFITS trial shows the same thing in one study. Its 609 adults without diabetes were assigned to a healthy low-fat or a healthy low-carbohydrate diet for 12 months, with both groups taught to choose good-quality food. The low-carbohydrate group lost 6.0 kg and the low-fat group 5.3 kg, a difference that was not statistically significant. By the end, the low-carbohydrate group was eating 30 per cent of its energy as carbohydrate and the low-fat group 48 per cent.

What do the trials show for blood glucose?

A clearer effect in the first six months, fading by twelve. The 2021 BMJ meta-analysis pooled 23 trials with 1,357 adults who had type 2 diabetes. At six months, 57 per cent of those on low-carbohydrate diets had an HbA1c under 6.5 per cent against 31 per cent on control diets. When that threshold had to be met without diabetes medicine, the effect was smaller and not significant, and it fell sharply in studies that included people using insulin. At 12 months the data were sparse.

The Cochrane review, looking at people with type 2 diabetes over one to two years, found probably little to no difference in HbA1c or LDL cholesterol between low-carbohydrate and balanced diets. The American Diabetes Association's report reads the evidence the same way: HbA1c benefits were more pronounced at three and six months but not at 12 and 24. It calls reduced-carbohydrate eating a viable approach for selected adults with type 2 diabetes, under the care of their treating team.

One Indian analysis points the same direction without going as far. Using diet and HbA1c data from 18,090 adults in the ICMR-INDIAB survey, researchers modelled the macronutrient mix linked to better glucose status. For newly diagnosed diabetes the model gave carbohydrate at 49 to 54 per cent of energy, protein at 19 to 20 and fat at 21 to 26. That is a modest cut with more protein, not a low-carbohydrate diet by any definition in the table above. The entry on insulin and insulin resistance explains the physiology.

How strong is the evidence?

Moderate for the short term and thin beyond two years. The Cochrane reviewers rated most outcomes as at high risk of bias, largely from participants dropping out, and found too little reporting of adverse effects to draw conclusions. The BMJ authors graded their evidence as moderate to low certainty. Every trial in the Cochrane review but one was run in a high-income country; none was Indian.

For longer horizons there are only observational studies, which cannot prove cause. A 2018 analysis in The Lancet Public Health followed 15,428 American adults for a median of 25 years and pooled seven other cohorts. Mortality was lowest at 50 to 55 per cent of energy from carbohydrate, and higher both below 40 per cent and above 70 per cent. The source of the replacement mattered: mortality rose when carbohydrate was exchanged for animal fat and protein and fell when the exchange was plant-based.

An Indian lower-carb day, in household measures

The vegetarian day below keeps one small cereal portion at each main meal and comes to about 121 g of carbohydrate, under the 130 g line and under half of the 275 g in ICMR-NIN's model day. It is an illustration, not a prescription. The carbohydrate figures are worked from raw weights in the Indian Food Composition Tables; the raw weight assumed for each portion is this page's approximation, and onion, tomato and cooking add a few grams more. Use the macro calculator to set your own split.

If you eat eggs, fish or chicken, two eggs or a palm-sized piece of fish can replace the paneer at dinner and take the day to roughly 109 g, since these foods carry no carbohydrate. Nothing in this day needs almond flour or a packet labelled keto. For ketogenic eating, which is far stricter, see NNWA's Indian keto diet plan.

A worked example: from a cereal-heavy plate to the guideline

This is an illustrative case, not a real person. Take a 44-year-old shop owner in Chennai who has type 2 diabetes, takes two glucose-lowering tablets, eats about 2,000 kcal a day and gets roughly 65 per cent of it from rice and other cereals, which is inside the range ICMR-NIN reports for Indian diets.

Common myths and mistakes

The second mistake is to treat low-carb as high-fat by default. In the cohort data the direction of the result turned on what replaced the carbohydrate, with plant sources of fat and protein on the favourable side. Choosing lower glycaemic index Indian foods for the carbohydrate that remains is the other half of the job.

Who should see a doctor first, and who should avoid it?

Anyone taking insulin or tablets that lower blood glucose must speak to the prescribing doctor before cutting carbohydrate. The American Diabetes Association's report states that a very low-carbohydrate plan can cause fluid loss and reduce blood glucose swiftly, so a knowledgeable practitioner is needed at the start to prevent dehydration and to adjust insulin and other glucose-lowering medicines against hypoglycaemia. A dose change is the doctor's decision.

The same report says that for people with chronic kidney disease, people with disordered eating and pregnant women, concerns about very low-carbohydrate plans are unresolved and no recommendation can be made. ICMR-NIN adds that a high protein intake may worsen kidney function in those who already have kidney disease. For medical nutrition in diabetes, start with NNWA's guide to a type 2 diabetes diet in India and your own treating team.

Watch LDL cholesterol too. The BMJ meta-analysis noted a worsening of LDL cholesterol at 12 months on low-carbohydrate diets that was clinically important though not statistically significant. If you follow a lower-carb plan for months, ask your doctor whether a lipid test is due.

Is low-carb the same as keto?

No. Keto is the far end of the range: the consensus report describes a goal of 20 to 50 g of non-fibre carbohydrate a day, with more than half of energy from fat, in order to produce nutritional ketosis. The day in this entry holds more than twice that. In the BMJ analysis, very low-carbohydrate diets were less effective for weight at six months than less restrictive ones, a result explained by how few people kept to them.

Can weight come down without counting carbohydrate at all?

Yes. The consensus report's reading of the weight-loss trials is that what matters is an eating plan you can sustain that leaves an energy deficit, whatever its macronutrient split. ICMR-NIN's advice is that weight reduction should be gradual, that a reducing diet should not go below 1,000 kcal a day and should supply all nutrients, and that half a kilogram a week is considered safe. NNWA's Weight Management course teaches that planning for healthy adults, and the calorie calculator gives a starting figure.

Learning this with NNWA

Carbohydrate planning for healthy adults is taught in NNWA's Diploma in Nutrition. The academy is a private skill institute in Kolkata, and its classes run online in English and in Hindi. Learners practise planning plates for healthy clients and learn to pass anyone with diabetes or another medical condition to a doctor. No NNWA course is a degree, and none is a dietitian qualification.