Entire grocery runs reorganize around this debate. Bread vanishes from one kitchen counter while butter returns to the next. Online arguments stretch for hundreds of replies, each side armed with a success story they fully believe and a study they half-read.
The largest Cochrane synthesis ever assembled on this exact question pooled sixty-one randomized trials and nearly seven thousand participants to settle it.
Low-Carb vs. Low-Fat: Which Actually Works Better Long Term?
Roughly one kilogram. At three to eight months, low-carb edged ahead by about a kilogram. By the one-to-two-year mark, the margin shrank to less than a kilogram and landed inside the range of a normal day's water fluctuation. Not clinically important. Smaller than the shift your bathroom scale registers between waking up and going to bed.
Neither low-carb nor low-fat produces meaningfully more weight loss. Across sixty-one trials, the difference was roughly one kilogram, shrinking to under a kilogram by two years. Individual variation within each diet, spanning up to forty kilograms, dwarfed the between-diet difference. Adherence and food quality predicted outcomes far better than the macronutrient ratio.
— Naude et al. 2022 · Cochrane Database Syst Rev · 61 RCTs, n=6,925
A year-long trial of 609 adults showed what the meta-analysis could not. Inside each diet group, the spread from biggest loser to biggest gainer was forty kilograms. Some people lost thirty kilograms eating low-fat. Others gained ten on low-carb. Same instructions, same coaching, same twelve months. Whether someone was assigned to low-carb or low-fat told you almost nothing about where they would land. The full trial puts numbers on every corner of that range.
40 kg
The spread within each diet group — from biggest loser to biggest gainer — in the same trial where the two diets differed by one kilogram
Low-carb does produce faster results on the scale, and the reason is visible in the first week. Glycogen stores deplete when carbohydrates drop. Water bound to those stores leaves with them. Two to three kilograms disappear in days. Eat carbohydrates again and every gram returns. None of it is fat. When body composition was measured directly, the group eating low-fat lost three times more actual body fat per day: fifty-one grams versus sixteen. What the scale reported and what the tissue underneath showed were two separate measurements.
Neither diet came with a calorie target. Both groups ate real food and stopped when they were full. Both spontaneously cut roughly five hundred to six hundred calories a day without counting a single one. Labels diverged, caloric reduction was identical. What made both diets work was food quality: processed food dropped, whole food increased, and hunger followed without being forced. It was the same mechanism 90s diet culture buried under fat-free cookies and sugar-loaded snack bars.
Under tightly controlled conditions where every meal was prepared and delivered, cutting carbohydrates did increase daily energy expenditure by about two hundred extra calories. Under those conditions, it existed. In every free-living trial where people chose their own food, it vanished. A metabolic edge that cannot survive a trip to the grocery store is not the edge the debate thinks it is.
Most of the sixty-one trials lasted fewer than two years. Whether one approach pulls ahead across a decade remains an open question, and the evidence to answer it does not exist yet.
One kilogram separated two diets. Forty kilograms separated two people on the same one. The debate picked the wrong scoreboard. What separated the two people had nothing to do with whether they ate bread or butter, and everything to do with why most diets fail in the first place.