Mifflin–St Jeor vs Harris–Benedict vs Katch–McArdle: which BMR formula should you use?

Type the same person into three calorie calculators and you can get three different numbers. The reason is almost always the formula underneath. All three common ones estimate resting metabolic rate, the energy your body burns at complete rest, from a handful of measurements, and they were fitted to different groups of people at different times. This guide shows the three equations, runs one person through all of them, summarises what the validation studies found, and ends with a recommendation. The TDEE and Macro Calculator on this site uses Mifflin–St Jeor; this is the reasoning behind that choice.

The three equations

In all formulas, W is body weight in kilograms, H is height in centimetres, A is age in years, and the result is kilocalories per day.

Mifflin–St Jeor (1990)

Derived from 498 healthy adults measured by indirect calorimetry in the late 1980s, with a deliberately wide age and weight range.

• Men: 10 × W + 6.25 × H − 5 × A + 5
• Women: 10 × W + 6.25 × H − 5 × A − 161

Harris–Benedict (1918, revised 1984)

The original was fitted to 239 subjects in 1918. Roza and Shizgal refitted it in 1984 on a larger sample; those revised coefficients are the ones most calculators labelled "Harris–Benedict" use today.

• Men: 88.362 + 13.397 × W + 4.799 × H − 5.677 × A
• Women: 447.593 + 9.247 × W + 3.098 × H − 4.330 × A

Katch–McArdle

Ignores sex, age and height and uses lean body mass instead, following Cunningham's finding that lean mass explains most of the variation in resting metabolism. Lean body mass is weight minus fat mass, so you need a body-fat percentage.

• All adults: 370 + 21.6 × LBM, where LBM = W × (1 − body fat ÷ 100)

One person, three answers

Take a 35-year-old man, 80 kg, 180 cm, with 20% body fat (lean body mass 64 kg), and a 30-year-old woman, 62 kg, 165 cm.

For the man, Mifflin–St Jeor and Katch–McArdle land within 3 kcal of each other, and Harris–Benedict sits about 70 kcal higher. That is a 4% spread at rest. Multiply by a moderate activity factor of 1.55 and the gap grows to roughly 110 kcal per day: 2,720 versus 2,829. Over a month that is the difference between a plan that holds weight and one that slowly adds it, which is why the choice of formula is not cosmetic.

What the validation studies found

The most cited comparison is the 2005 systematic review by Frankenfield, Roth-Yousey and Compher for the American Dietetic Association. It pooled studies that measured resting metabolic rate directly and compared the measurements with the predictions of several equations. Mifflin–St Jeor came within 10% of the measured value for 82% of non-obese adults and 70% of obese adults, the best result of the equations reviewed. Harris–Benedict hit that band less often and tended to overestimate, which matches the worked example above: it was fitted a century ago to a leaner, more physically active population, so it tends to run high for people today.

Two caveats apply to every formula. First, "within 10%" still means that roughly one person in five is further off than that; a 10% miss on a 1,750 kcal resting rate is 175 kcal a day. Second, the validation samples were mostly white North American and European adults aged 18 to 65; the errors are larger in older adults, in people with very high or very low muscle mass, and in populations the equations were not fitted to.

When Katch–McArdle is the better choice

Because it works from lean body mass, Katch–McArdle handles the two cases where weight-based formulas fail: very muscular people, whom Mifflin–St Jeor underestimates because it treats extra kilograms as average tissue, and people with high body fat, whom it overestimates for the same reason. The catch is the input. Consumer body-fat scales are commonly off by several percentage points. If our 80 kg man is actually at 25% body fat rather than 20%, his lean mass is 60 kg and the formula gives 1,666 kcal, 86 kcal below the earlier result. A body-fat error of five points moves the answer as much as switching formulas does. Katch–McArdle only earns its extra precision when the body-fat figure comes from a reliable method such as DEXA or a properly done skinfold measurement.

From resting rate to daily need

None of these numbers is what you eat. Resting metabolic rate is multiplied by an activity factor to get total daily energy expenditure. The usual multipliers, 1.2 for sedentary up to 1.9 for very active, are rounded versions of the physical activity level ranges in the FAO/WHO/UNU report. The multiplier is the larger source of error: choosing 1.55 when 1.375 is honest adds about 300 kcal to the man's estimate, four times the spread between the formulas. Most people overestimate their activity level, so pick the lower band when in doubt.

Recommendation

1. Use Mifflin–St Jeor as the default. It has the best validation record and needs only measurements you already know.
2. Use Katch–McArdle if you have a body-fat percentage from DEXA or skinfolds and you are noticeably more muscular or heavier than average.
3. Treat any result as ±10%. The formula gives a starting point, not a prescription.
4. Check it against reality: weigh yourself under the same conditions for two to three weeks. If weight drifts, adjust intake by 100 to 200 kcal and repeat. The measured trend beats every equation.

This is exactly what the TDEE and Macro Calculator does: Mifflin–St Jeor, a conservative activity multiplier, and a macro split you can adjust once you have real data.

Limitations

All three equations describe healthy adults. They are not validated for children, pregnancy, thyroid or other metabolic conditions, or for people on medication that changes energy expenditure. Resting metabolic rate also adapts to prolonged dieting, so an estimate made at the start of a diet drifts high over time.