Match two people on lean mass and you close about 83 percent of the gap#
Two adults of the same height and weight can burn hundreds of calories a day differently at rest, and the first thing to know is how much of that gap is not mysterious at all. When 130 adults from 54 families had resting metabolic rate measured by indirect calorimetry, three unglamorous covariates — fat-free mass, age and sex, with fat-free mass by far the most important — accounted for 83 percent of the variance between them1. "Same size" on a bathroom scale is not the same body. Two people who weigh 70 kg can differ by 8 kg of lean tissue, and that difference alone is worth more than any habit either of them has.
The remaining sixth is the interesting part, and it is not noise. It divides into things that have been measured — the size of your internal organs, and something that runs in families — and a residue nobody has named. This article is about that sixth: what is actually in it, how big it is in calories, and why it means the single number a calculator hands you is a population average wearing your name. TDEE explained covers what the number is made of; this is about why yours is not anyone else's.
The organs are where the next slice hides#
If lean mass explains most of the gap, the obvious next question is which lean mass. Fat-free mass is a bag containing brain, liver, heart, kidneys, gut, bone, blood and muscle, and those components do not run at remotely similar rates.
Researchers put that to the test directly: 87 adults had liver, kidney, spleen, heart and brain mass measured by MRI, with fat and fat-free mass by DXA, and resting expenditure measured. A model using fat-free mass, fat mass, age, sex and race explained 69.6 percent of the variance in REE. Adding the measured masses of heart, liver, kidneys and spleen pushed it to 72.3 percent. Adding brain mass took it to 74.5 percent — leaving 25 percent unexplained, which the authors attributed to measurement error, unmeasured organs such as the gut and lungs, and person-to-person differences in the organs' own metabolic rates2.
Five percentage points sounds small until you read what fell out of the model when the organs went in. Age, sex and race had all been significant predictors. After brain and trunk-organ mass were added, none of them was: age P = 0.143, sex P = 0.684, race P = 0.384. They had been standing in for organ size the whole time.
Sex is in every metabolic equation ever published. Measure people's organs and it stops predicting anything. It was never a metabolic variable — it was a proxy for one.
That reframes what a calculator is doing when it asks your sex and your age. It is estimating, from four cheap inputs, a quantity that only an MRI scanner can see. A quarter of the differences between people survive even the scanner — a residual close in size to the one a separate study found when it partitioned basal metabolic rate against every candidate variable it could measure, discussed in why every TDEE calculator is an estimate.
| What the model adjusts for | Share of REE variance explained |
|---|---|
| Fat-free mass, age, sex (130 adults, indirect calorimetry) | 83% |
| Fat-free mass, fat mass, age, sex, race (87 adults, MRI + DXA) | 69.6% |
| ...plus heart, liver, kidney and spleen mass | 72.3% |
| ...plus brain mass | 74.5% |
| Family membership, independent of fat-free mass, age and sex | +11 percentage points |
It runs in families, and the twins show how hard#
Bogardus's team went looking for one more variable, and found it in the pedigree. After fat-free mass, age and sex were accounted for, membership of a particular family explained a further 11 percent of the variance in resting metabolic rate1. Your resting rate resembles your relatives' beyond what your shared body composition can explain.
A separate cohort found the same clustering in whole-day expenditure: among 94 siblings from 36 families measured in a respiratory chamber, adjusted 24-hour energy expenditure aggregated within families at an intraclass correlation of 0.483.
The most vivid version is on the intake side. Twelve pairs of young adult male identical twins were overfed by 1,000 kcal a day, six days a week, for 84 days — an excess of 84,000 calories each. Mean weight gain was 8.1 kg. The range was 4.3 to 13.3 kg. Twins resembled their co-twin closely enough that there was about three times as much variance among pairs as within pairs, and about six times as much for changes in abdominal visceral fat4. Identical calories, identical supervision, a threefold spread in outcome, and the spread was organized by genome.
Two brackets. Twelve pairs is small, and "three times more variance among pairs" is a family-resemblance statistic rather than a heritability estimate. But the direction is not really contested, and it is the honest answer to the question people are actually asking when they say a friend eats more and stays lean.
What a 200-calorie shortfall predicted#
None of the above says a low metabolic rate causes weight gain. One study tried to test that prospectively, and its result is worth reading carefully because it gets quoted much more loosely than it was written.
Ninety-five adults had 24-hour energy expenditure measured in a respiratory chamber and were followed for two years. Adjusted expenditure correlated inversely with the rate of weight change (r = −0.39, P < 0.001), and the estimated risk of gaining more than 7.5 kg was fourfold higher in people running 200 kcal/day below their predicted value than in those running 200 kcal/day above it (P < 0.01). In a separate 126 people followed for four years, those who gained more than 10 kg had started with a lower resting rate — 1,694 versus 1,764 kcal/day, P < 0.023.
Hold that 70-calorie difference next to the outcome it preceded. It is smaller than the error bar on most calorie estimates, and it separated people who put on 10 kg from people who didn't over four years. Which is the useful lesson and also the reason to be careful: this was a single population with an unusually high obesity risk, the association is not proof that the low rate did the causing, and a fourfold relative risk in a cohort is not a verdict on any individual. What it does establish is that the residual is not a rounding error. Differences too small to feel, sustained over years, land somewhere.
Using a number that is yours and unmeasurable#
Put the pieces together and the practical shape is clear enough.
Most of the difference between you and someone your size is lean mass, which you can partly influence and which is why body composition, not scale weight, is the variable worth watching — though the resting-burn contribution of new muscle is smaller than the folklore claims. The next slice is organ proportion, which is fixed, invisible and worth roughly what an MRI would tell you. The slice after that is familial and equally fixed. What remains is a quarter of the variance that nobody has been able to name with a scanner in the room.
So the correct posture toward a calculated number is neither trust nor dismissal. It is: this is the average of people who share my four inputs, my own value sits somewhere in a band around it, and the only instrument that can locate me inside that band is my own weight trend over a few weeks — the method in how to find your maintenance calories. Your baseline, hour by hour, is described in how many calories you burn doing nothing; how much of the between-person gap is simply a matter of being bigger is the subject of why height and body size drive calorie burn.
And if the true answer to "why does she eat more than me and stay lean" is more lean tissue, a larger liver, a different family, and something a laboratory cannot name — that is a better place to stand than the alternatives, which are that she is lying or that you are broken. Neither is usually true, and the third explanation is the one with the measurements behind it. The commoner reasons a diet stalls are laid out in why am I not losing weight; a slow metabolism is rarely the front-runner.
FAQ#
Is a "slow metabolism" actually inherited?#
Partly, and it has been measured. In 130 adults from 54 families, family membership explained 11 percent of the variance in resting metabolic rate after fat-free mass, age and sex were accounted for, and in 94 siblings from 36 families, adjusted 24-hour expenditure clustered within families at an intraclass correlation of 0.48. That is a real familial signal — much smaller than lean mass, and much larger than nothing.
Does a low metabolic rate mean you will gain weight?#
It shifts the odds rather than deciding the outcome. Among 95 adults followed for two years, running 200 kcal/day below predicted expenditure came with a fourfold higher risk of gaining more than 7.5 kg compared with running 200 above. That was one high-risk population and an association rather than a demonstrated cause — but the differences involved were small enough to be invisible day to day, which is the point.
If my friend eats more than me and stays lean, what is actually different about us?#
Most likely lean tissue, first and foremost — fat-free mass, age and sex together account for 83 percent of the between-person variance in resting rate. After that, organ size: measuring heart, liver, kidney, spleen and brain mass by MRI explained about 5 percentage points of variance that height and weight cannot see, and it knocked age, sex and race out of the model entirely. Some of the rest runs in her family, and about a quarter of it nobody has managed to name.
Sources#
- Bogardus C, Lillioja S, Ravussin E, et al. Familial dependence of the resting metabolic rate. N Engl J Med. 1986;315(2):96-100.
- Javed F, He Q, Davidson LE, et al. Brain and high metabolic rate organ mass: contributions to resting energy expenditure beyond fat-free mass. Am J Clin Nutr. 2010;91(4):907-912.
- Ravussin E, Lillioja S, Knowler WC, et al. Reduced rate of energy expenditure as a risk factor for body-weight gain. N Engl J Med. 1988;318(8):467-472.
- Bouchard C, Tremblay A, Després JP, et al. The response to long-term overfeeding in identical twins. N Engl J Med. 1990;322(21):1477-1482.
