Do calories matter more than carbs or fat? The big trials say the mix barely counts

Mostly, yes — for weight, the total amount you eat matters far more than the ratio of carbohydrate to fat, and the largest head-to-head trials keep finding that competing diets produce nearly identical results once calories are accounted for [s1] [s2]. That does not make food quality irrelevant, but it undercuts the idea that one macronutrient is uniquely fattening.

The low-fat versus low-carb showdown

The cleanest test is the DIETFITS trial at Stanford, which randomised 609 adults aged 18 to 50 with a BMI of 28 to 40 to a healthy low-fat or a healthy low-carbohydrate diet for 12 months [s1]. Both groups were coached in 22 diet-specific small-group sessions to cut fat or carbohydrate as far as they sustainably could while emphasising food quality — vegetables, whole foods, and as little added sugar and refined grain as possible [s1].

The diets genuinely diverged. Over the 12 months the low-fat group ate 48% of energy as carbohydrate and 29% as fat; the low-carb group ate 30% carbohydrate and 45% fat [s1]. Yet the weight change was −5.3 kg on low-fat and −6.0 kg on low-carb, a between-group difference of just 0.7 kg (95% CI, −0.2 to 1.6) — not statistically significant [s1]. The trial also tested two popular escape hatches: whether people with a "low-fat" or "low-carb" genotype pattern, or with higher insulin secretion, did better on the matching diet [s1]. Neither held — there was no significant diet–genotype interaction (P = .20) and no diet–insulin interaction (P = .47) [s1].

An even larger trial found the same

A decade earlier, the POUNDS LOST trial randomised 811 overweight adults to one of four diets that varied fat (20% or 40% of energy), protein (15% or 25%) and carbohydrate (35% to 65%) [s2]. At six months every group had lost about 6 kg — around 7% of body weight — then began regaining after a year [s2].

By two years the differences between the macronutrient targets were trivial: 3.0 versus 3.6 kg for lower versus higher protein, 3.3 kg for both fat levels, and 2.9 versus 3.4 kg for the highest versus lowest carbohydrate diets, with P greater than 0.20 for every comparison [s2]. What did track with success was showing up: attendance at the counselling sessions was strongly associated with weight loss, at about 0.2 kg per session attended [s2]. The trial's conclusion was that reduced-calorie diets produce clinically meaningful weight loss regardless of which macronutrients they emphasise [s2].

Why "a calorie is a calorie" is roughly right — but not the whole story

These trials do not say food choice is irrelevant. They say that once total energy is similar, shifting the carbohydrate-to-fat ratio does little to weight [s1] [s2]. Food quality still matters, through a different door: it changes how full you feel, and therefore how many calories you eat in the first place. In both trials people were steered toward minimally processed foods, and both diets worked partly by making a spontaneous calorie deficit easier to sustain [s1] [s2].

There are honest limits. Both trials relied on people following advice in the real world, where adherence fades — which is exactly why attendance predicted results [s2]. Neither ran indefinitely, and both saw regain begin within a year or two [s2]. And "calories in" is not a dial you set consciously; it is the output of appetite, habit and environment, which is what makes sustained restriction hard.

The result also does not license the reverse claim — that because macros barely move the average, what you eat never matters. For a person with diabetes, carbohydrate load changes blood glucose regardless of weight; for someone with high triglycerides, fat and sugar quality shift the lipid panel [s1]. The trials measured weight, not every outcome, and both diets in DIETFITS were the healthy versions of their kind, not junk-food low-carb or junk-food low-fat [s1]. That distinction matters: an "it's all calories" reading is right about the scale and wrong if it is used to defend a diet of ultra-processed food that happens to hit a calorie target, because the same total energy from different foods can leave you hungrier and worse off on other markers [s1] [s2].

The takeaway is practical. There is no metabolic magic in cutting carbs versus cutting fat; pick the pattern of minimally processed food you can actually stick to, because adherence, not macronutrient ideology, is what moves the scale [s1] [s2].

This article is informational and is not medical advice.

Sources

• Effect of Low-Fat vs Low-Carbohydrate Diet on 12-Month Weight Loss in Overweight Adults and the Association With Genotype Pattern or Insulin Secretion — JAMA, 2018-02-20
• Comparison of Weight-Loss Diets with Different Compositions of Fat, Protein, and Carbohydrates — New England Journal of Medicine, 2009-02-25