How Much Carb Should You Cut? The 30–40% Evidence Zone

For weight loss, cutting carbohydrates appears to help most when it moves a typical diet into the middle—not when it drives carbohydrate intake toward zero. A 2024 dose-response meta-analysis of 110 randomized trials found the strongest longer-term weight results at roughly 30–40% of calories from carbohydrates in adults with overweight or obesity. Below that range, added restriction produced little reliable extra benefit after a year.

That does not make 30% a magic number. The evidence cannot support that kind of dietary astrology. It does offer a useful correction to the idea that every gram of carbohydrate removed creates another unit of benefit. The relationship looks more like a curve than a cliff.

Where the weight-loss curve starts to flatten

In their 2024 systematic review and dose-response meta-analysis, Sepideh Soltani and colleagues modeled carbohydrate intake as a percentage of total energy across 110 randomized controlled trials in adults with a BMI of at least 25. Rather than sorting diets into broad labels such as "low-carb" and "not low-carb," the authors used restricted cubic splines to estimate how weight changed across the carbohydrate range.

At six months, weight loss accelerated as carbohydrate fell from about 45% to 30% of calories. Below about 30%, the curve flattened and, in some portions of the model, turned upward. At 12 months, diets providing 30–40% of calories from carbohydrate had the best sustained weight results. Below roughly 20–25%, further restriction offered negligible or negative incremental return. Beyond 12 months, both lower-than-30% and higher-than-40% carbohydrate intakes performed worse than the middle range.

The practical reading of the 110-trial model is modest: moving toward 30–40% of calories from carbohydrate may improve weight loss; pushing far below that range has not shown a dependable long-term payoff.

The certainty matters. The review rated the six-month spline evidence as moderate certainty, but the confidence intervals widened substantially after 12 months because fewer trials contributed data. Treat 30% as an approximate zone where returns diminish, not as a prescription precise to the decimal point.

The authors proposed several reasons why very low carbohydrate intake may lose its edge over time: dietary fatigue and social friction can erode adherence; early ketosis-related changes in appetite or energy expenditure may fade; and people may replace carbohydrate calories with enough fat or protein to erase the energy deficit. Those explanations are plausible, but the review did not report a formal adherence sub-analysis. The curve shows the pattern; it does not prove which mechanism caused it.

When calories are matched, the advantage is small

The best test of whether carbohydrate restriction itself produces a large weight-loss advantage is a direct comparison with balanced-carbohydrate diets under similar energy restriction. A 2022 Cochrane review pooled 61 randomized trials involving 6,925 participants.

Among adults with overweight or obesity who did not have type 2 diabetes, low-carbohydrate diets produced about 1.07 kg more weight loss than balanced-carbohydrate diets at 3–8.5 months, with moderate-certainty evidence. At 12 months or longer, the difference was 0.93 kg in favor of low-carb diets (95% CI, 0.04 to 1.81 kg; moderate certainty). The estimate technically favored low-carb, but less than one kilogram is small beside the roughly 4–5 kg that equals 5% of body weight for an 80–90 kg person.

Results became less persuasive when maintenance was part of the intervention. In the weight-loss-plus-maintenance subgroup, the difference at 12 months or longer narrowed to 0.30 kg, with low-certainty evidence. The review also examined the chance of achieving at least 5% weight loss at 12 months or more: the risk ratio was 1.11 (95% CI, 0.94–1.31), a non-significant result rated very low certainty. Only two trials with 137 participants informed that analysis, so it cannot settle whether low-carb helps more people reach a clinically meaningful loss.

These findings do not mean carbohydrate restriction fails. Both dietary approaches can support weight loss. They mean that, when overall calories are similarly restricted, carbohydrate percentage alone does not explain a dramatic or durable gap in body weight.

The replacement macronutrient changes the answer

"Low-carb" describes a subtraction, not a full diet. The calories removed from carbohydrate have to go somewhere. A 2025 meta-analysis by Shuo Feng and colleagues found that protein-substituted carbohydrate-restricted diets outperformed fat-substituted versions for weight outcomes and lean-mass preservation. Fat substitution also carried a larger LDL-raising signal.

That distinction helps explain why trials grouped under the same low-carb label can produce different results. A 30% carbohydrate diet built around protein-rich foods, non-starchy vegetables, legumes where they fit, and unsaturated fats is not metabolically interchangeable with one that replaces most carbohydrate calories with saturated fat. The macronutrient mix may modify both adherence and risk markers.

Feng and colleagues also found that carbohydrate-restricted diets lowered triglycerides and raised HDL cholesterol, while increasing LDL cholesterol and total cholesterol. Those are opposing biomarker changes, not a net cardiovascular verdict. No included evidence establishes that the triglyceride and HDL changes offset higher LDL for an individual's cardiovascular risk.

Cardiovascular biomarkers do not establish superiority

At 12 months or longer, the Cochrane review found no statistically significant low-carb advantage for diastolic blood pressure or LDL cholesterol in non-diabetic participants. For LDL, the mean difference was +0.04 mmol/L (95% CI, −0.05 to +0.12), which includes no difference. These are surrogate outcomes, not heart attacks, strokes, or mortality. None of the major reviews was powered to determine whether low-carb diets prevent those events.

For people with established cardiovascular disease, the comparison with other dietary patterns is especially relevant. A 2024 network meta-analysis by Bonekamp and colleagues found Mediterranean and other balanced patterns ranked at least comparably with low-carbohydrate diets for cardiovascular risk-factor reduction in patients with cardiovascular disease. Low-carb therefore has no demonstrated cardiovascular superiority in secondary prevention.

Anyone with coronary artery disease, elevated LDL cholesterol, or high cardiovascular risk should discuss a major carbohydrate reduction with a clinician who can monitor lipids and medications. A favorable triglyceride result does not erase an LDL increase by force of optimism.

Type 2 diabetes has a different, narrower question

Weight loss and glucose control should not be treated as the same endpoint. In people with type 2 diabetes, a 2022 dose-response meta-analysis of 50 randomized trials found a more linear relationship between carbohydrate restriction and glycemic outcomes across roughly 65% to 10% of calories from carbohydrate. Deeper restriction appeared to produce progressively greater HbA1c benefit in that population.

That finding does not overturn the weight-loss curve in the 110-trial analysis. The studies examined different populations and primary outcomes. The diabetes evidence concerns glycemic control; the broader analysis concerns body weight in adults with overweight or obesity.

Short-term glucose improvements may also partly reflect weight loss rather than carbohydrate restriction alone. In a six-week randomized trial by Thomsen et al., 72 adults with type 2 diabetes and obesity followed either a 30% carbohydrate, 30% protein, 40% fat diet or a 50% carbohydrate conventional diabetes diet. Food provision and individually tailored energy intake produced the same 6% weight loss in both groups. The carbohydrate-reduced diet lowered HbA1c more, by 1.9 mmol/mol (95% CI, −3.5 to −0.3), but it did not improve glucose tolerance, beta-cell response to glucose, or insulin sensitivity during a four-hour oral glucose tolerance test.

Long-term durability remains uncertain. A 2025 systematic review found that HbA1c advantages of low-carb over low-fat diets in type 2 diabetes attenuated with longer follow-up. And in the small Keto-Med randomized crossover trial of 33 people with prediabetes or type 2 diabetes, Gardner and colleagues found no clear HbA1c superiority for a ketogenic diet over a Mediterranean-style diet. The ketogenic arm produced greater short-term weight loss, while the Mediterranean approach had better adherence and lipid outcomes in some participants.

What the evidence supports

If your goal is weight loss, a carbohydrate target around 30–40% of calories is a defensible starting range, provided the diet creates an energy deficit you can sustain. Replace some carbohydrate with protein rather than simply adding fat, especially if preserving lean mass matters. The best carbohydrate level is still one you can follow while meeting your nutritional needs and managing LDL cholesterol.

For type 2 diabetes, deeper carbohydrate restriction can improve HbA1c in the short term, but medication adjustments and glucose monitoring may be necessary. The longer-term evidence does not show that ketogenic intake reliably beats a well-designed Mediterranean-style or balanced-carbohydrate diet.

The strongest conclusion is also the least glamorous: carbohydrate restriction can be useful, but extreme restriction has not earned a permanent advantage over a high-quality, calorie-controlled diet that people can actually maintain.