Steeper calorie cuts cost more muscle — exercise protects it

Not all weight lost is fat: a portion of any large weight loss comes from fat-free mass — muscle and other lean tissue. How big that portion is matters, because lean mass underpins strength, function and resting metabolism. Yet it has been unclear what actually determines the split: is it how aggressively calories are cut, whether exercise is included, how much weight is lost overall, or the starting body mass index? This review set out to quantify the proportion of weight lost as fat-free mass across different interventions, from diet and behaviour change through to bariatric surgery.

The authors systematically searched Medline and Embase for studies with reliable measurements of fat-free mass before and after weight loss of more than 10 kg, then pooled eligible data. They included 26 cohorts treated with dietary and behavioural interventions and 29 cohorts of bariatric surgery patients. Using a fixed-effect model of the percentage of fat-free mass loss per weight loss, linear regression tested the influence of degree of caloric restriction, exercise, magnitude of weight loss, initial BMI and type of surgery.

The degree of caloric restriction was positively associated with the percentage of weight lost as fat-free mass (r² = 0.31, P = 0.006): the harder the calorie cut, the more of the loss came from lean tissue. In three randomised controlled trials, exercise was shown to decrease the proportion of fat-free mass lost. Among surgical procedures, biliopancreatic diversion and Roux-en-Y gastric bypass caused greater fat-free mass loss than laparoscopic adjustable gastric banding (r² = 0.453, P < 0.001), and these differences were independent of initial BMI and the magnitude of weight loss.

The findings put numbers on a principle widely held in body-composition research: the severity of the energy deficit and the presence of exercise shape not just how much weight is lost but what that weight is made of. The evidence base has limits, though — it is drawn largely from cohorts rather than randomised trials, with only three RCTs directly informing the exercise effect, and it concerns large losses of more than 10 kg. How these relationships play out with smaller deficits, or how other factors interact with them, is not answered here.

For anyone undertaking substantial weight loss, this is a reason to think the composition of what you lose is not fixed: gentler caloric restriction was associated with a smaller share of the loss coming from lean tissue, and exercise during weight loss reduced that share in the trials that tested it. It is also a reminder that the scale alone tells an incomplete story, since two people losing the same weight can lose very different amounts of muscle along the way. This describes the pattern in the evidence rather than prescribing any particular deficit.