Adding exercise to a diet: 21 randomised trials pooled, about 4.5 kg more weight lost on average, and no proof it spares muscle

Eating less and moving more is the default advice for anyone carrying extra weight, and most weight programmes bundle the two together. That makes it surprisingly hard to say what each half does. A trial that compares diet plus exercise with doing nothing tells you about the package; only a trial that compares it with the same diet alone tells you what the exercise added. A review published in Frontiers in Nutrition on 28 October, by researchers in China and at the University of Georgia, kept those questions apart.

The team searched four large databases up to 28 September 2026 and sifted 16,531 records down to 23 reports from 21 randomised trials in adults with overweight or obesity. Every trial tested structured exercise combined with a diet: ordinary calorie cutting, intermittent fasting, time-restricted eating, alternate-day fasting, Ramadan fasting or a low-carbohydrate diet. Instead of averaging everything into one number, they grouped the trials by what the combined programme was up against: the diet alone, the exercise alone, a different diet and exercise plan, or minimal guidance. They looked at end-of-trial weight, BMI, fat-free mass and body fat percentage.

Against diet alone, the comparison that isolates what exercise adds, 16 reports showed people on the combined programme finished 4.49 kg lighter on average (95% confidence interval 0.92 to 8.05 kg) and 1.50 BMI units lower. Against exercise alone, the gap was 3.42 kg, but the five reports were too few and too scattered to rule out no difference. Each of the other two comparisons rested on a single trial.

The headline average hides a lot. The trials disagreed with each other strongly, and the prediction interval, which estimates what a new trial in a new setting might find, ran from 16.8 kg more weight lost to 7.9 kg less. In plain terms: on average, adding exercise to a diet helped the scales, but the effect varied so much between programmes that some could add nothing. Rating the evidence with the GRADE system, the authors judged its certainty very low for every outcome.

The body composition results are the part many people will find surprising. Exercise is often sold as the way to lose fat while keeping muscle, yet fat-free mass, which includes muscle, organs, bone and body water, came out 0.53 kg lower in the combined groups than with diet alone, with a range that included no difference. Body fat percentage didn't shift reliably either. The authors say plainly that this does not show exercise protects lean mass, and that the data couldn't tell whether calorie cutting or fasting works better alongside exercise. They also stress that exercise improves fitness, strength and keeping weight off even when it adds little on the scales, benefits this review didn't measure.