If you train while you diet, how much lean mass do you keep?

Does Exercise Stop You Losing Muscle on a Diet? 34 Trials, 1,455 Adults

2026-11-07 · By Anthony Falcone

Verdict: Keep it. Train while you diet. Across 34 randomised trials in adults with obesity, adding structured exercise to a calorie deficit kept 0.87 kg more fat-free mass than the same diet alone, and on average prevented nearly half of the loss. Strength training, or strength and cardio combined, did best; cardio alone fell just short of a clear effect. The authors warn the pooled effect may be overstated, and fat-free mass is not the same as muscle.

What the researchers did

When you eat less to lose weight, you lose some fat-free mass along with the fat, and part of that is muscle. Muscle handles most of the sugar you absorb after a meal and helps keep you strong as you age, so the usual advice is to train while you diet. This review from the German Sport University Cologne asked how much training actually helps, and whether the type of training matters.

In June 2025 the team searched two databases for randomised trials in adults with an average BMI of at least 30. Every trial had to compare a calorie-restricted diet on its own with the same diet plus a structured strength, endurance or mixed programme (mixed meaning strength and cardio combined, or functional training for both), lasting at least 4 weeks. The diet-only group could not be given extra activity advice, and body composition had to be measured before and after.

34 trials with 1,455 people made the cut. They ranged from 10 to 100 participants, with average ages from 29 to 70. Fifteen trials were in women only, two in men only and seventeen in both. None used weight-loss drugs. The authors treat fat-free mass and lean body mass as the same thing, and so does this article.

They ran two analyses: a standard one comparing each exercise group with its diet-only group, and a network meta-analysis that keeps trials with several exercise arms intact and ranks the three training types against each other.

What they found

Adding exercise kept more fat-free mass: 0.87 kg more than the diet alone (95% confidence interval 0.59 to 1.16 kg). Taking out any single trial barely moved the result, which stayed between 0.81 and 0.92 kg. The exercise groups also lost 0.88 kg more body weight.

Put another way: averaged across the study arms, exercise prevented 45.7% of the fat-free mass loss, nearly half. The authors present this as an illustration from weighted averages, not a formal test. Across all the trials, the weighted average loss of fat-free mass was 1.9 kg. In 10 of the 44 exercise groups, fat-free mass held steady or even rose while the diet-only group lost some.

The type of training seemed to matter, though the standard analysis couldn't prove it (the difference between types had p = 0.092). In the network analysis, mixed training kept the most fat-free mass (1.20 kg more than diet alone), then strength training (0.83 kg). Endurance training alone kept 0.51 kg, and its confidence interval crossed zero (−0.04 to 1.05 kg), so it fell just short of significance. On the illustrative averages, mixed training prevented 66.4% of the loss, strength 60.6% and endurance 25.8%.

Neither the total number of training sessions nor the amount of weight lost explained why some trials found bigger effects than others.

Only two trials measured skeletal muscle itself rather than fat-free mass, and both pointed the same way. In one, men on the diet alone lost 2.6 kg of muscle, against 0.9 kg with added endurance training and 1.0 kg with added resistance training; for women the figures were 1.2 kg, 0.5 kg and 0.7 kg.

How sure can we be?

The direction is consistent; the size is less certain. The trials were of fair quality on average (5 on the PEDro scale, where the included trials ranged from 4 to 7), and on the Cochrane risk-of-bias tool 14 of the 34 were rated high risk and the other 20 had some concerns. The funnel plot and Egger's test showed signs that small trials with weaker results may be missing, so the authors warn that the 0.87 kg may be an overestimate.

Fat-free mass is not muscle. It also includes water, organs and other non-fat tissue, and the trials measured it five different ways, from DXA scans (14 trials) to bioelectrical impedance (9 trials). During weight loss, shifts in body water can change these readings without any change in muscle.

About half the trials also raised protein in the diet, and the review didn't separate the effect of protein from the effect of training. Two trials using high-intensity interval training stood out with less favourable results, which the authors think may reflect too little recovery; another HIIT trial found the opposite, so that question is open.

What this means for you

If you are cutting calories, add structured training. In these trials it prevented nearly half of the lean-mass loss that the diet alone caused.
Make strength training part of it. Strength work, alone or combined with cardio, did best; cardio on its own showed a smaller effect that fell just short of significance.
In these trials, more total sessions were not linked to keeping more lean mass, so the review can't give a minimum dose. A programme you can keep up for the whole diet is the sensible target.
Read smart-scale lean-mass numbers with care while you diet: water shifts can move them without any change in muscle.
If you have a medical condition, take regular medication (including weight-loss injections) or are older and new to exercise, talk to your doctor before starting a diet and training plan.

How much to trust it

The authors' own: the exercise programmes and diets varied a lot, which drove substantial heterogeneity; medication use and chronic disease weren't accounted for; fat-free mass was measured with five different methods; there were signs of publication bias, so the pooled effect may be overestimated; and deviations from the planned interventions were often a high risk of bias.
Fat-free mass is a stand-in for muscle, not a direct measure. Only two of the 34 trials measured skeletal muscle itself, too few for firm conclusions about muscle.
Trial quality was fair at best: 14 of 34 trials at high risk of bias and the other 20 with some concerns. Many trials were small, some with as few as 10 participants.
Everyone had obesity (an average BMI of 30 or more). The results may differ for leaner people dieting for a smaller loss, and no trial used weight-loss drugs, so the review can't say how exercise works alongside them.
The "nearly half" figure (45.7%) comes from illustrative weighted averages, and the paper's discussion quotes 54% for the same comparison. Both point the same way; we use the figure from the results and conclusions.
The authors report no funding. The senior author declares lecture and consultation fees from Abbott, Lilly and Novo Nordisk, two of which make weight-loss drugs; the other authors declare no conflicts.

The study

Deller M, Weiershaus J, Held S, Brinkmann C. Effects of Calorie Restriction With and Without Strength, Endurance or Mixed Training on Fat-Free and Skeletal Muscle Mass in Overweight or Obese Individuals: A Systematic Review With Pairwise Meta-Analysis and Network Meta-Analysis of Randomized Controlled Studies. Diabetes, Obesity and Metabolism, 2026. doi.org/10.1111/dom.70873

Our Fat Loss Protocol sets your calorie, protein and weekly-loss targets, shows how to protect your muscle with protein and training, builds meals that keep you full and plans your first four weeks, plateaus included. See the protocol

Want to change how you eat first? The Keto Blueprint is a 14-day, research-checked plan.