BMR Formula Explained: Mifflin-St Jeor, Harris-Benedict and the Rest Compared

5 November 2026 · 15 min read · AI Basics

Introduction

Search for a BMR formula and you will find several, all claiming to estimate the same thing, all giving different answers. Mifflin-St Jeor says one number, Harris-Benedict says another, and a lean-mass formula says a third. The confusion is understandable, because the formulas really do disagree, sometimes by 100 calories or more, and none of them is the true value, which can only be measured in a laboratory.

This guide explains every major BMR formula in plain language: what it asks for, what it assumes, where it came from, and when it is the right choice. By the end you will know why Mifflin-St Jeor is the default, when Harris-Benedict still appears, why body fat based formulas exist, and how to pick the formula that fits your situation instead of the first one a search result shows you.

Why there is more than one BMR formula

A BMR formula is a statistical shortcut. Researchers measure resting energy expenditure in a group of people, usually by indirect calorimetry, which analyses the oxygen consumed and carbon dioxide produced while a person rests, and then they fit a simple equation to those measurements. The equation uses easy inputs, weight, height, age, sex, and sometimes body fat, to predict the measurement for a new person who has never been near a metabolic cart. Different research groups, different decades, and different study populations produce different equations, and each one is best for people resembling its original study group.

That history explains the spread. Harris-Benedict came from measurements taken in 1919, on a small, mostly young, mostly lean population, at a time when average body composition was very different. Mifflin-St Jeor came from 498 adults measured in the late 1980s, a broader group including people with obesity. Katch-McArdle came from a desire to remove the guesswork about body composition by using lean mass directly. None of them is wrong in its own terms; each is a best fit for a different population, and the skill is matching the formula to the person.

Mifflin-St Jeor: the modern default

The Mifflin-St Jeor equation, published in 1990, is the formula behind most modern calculators, including ours. For men: BMR equals 10 times weight in kilograms plus 6.25 times height in centimetres minus 5 times age plus 5. For women: the same, with minus 161 instead of plus 5. It was derived from resting energy expenditure measured by indirect calorimetry in 498 healthy adults aged 19 to 78, both normal weight and obese, which is a far broader base than its predecessor. The Mifflin-St Jeor equation explained sets out its derivation, and the Mifflin-St Jeor equation on Wikipedia gives the original coefficients and the simplified form used here.

Its strength is accuracy for the general adult population. A systematic review by the American Dietetic Association found it the most reliable of the common predictive equations, typically within about 10 percent of measured values for non-obese adults, and better than Harris-Benedict for people with obesity as well. Its weakness is the same as every weight-based formula: it cannot see body composition. A muscular 90 kilogram person and a sedentary 90 kilogram person get the same answer, and the truth sits on opposite sides of it for each of them. For most users, though, that error is smaller than the error from dishonest activity factors, so Mifflin-St Jeor remains the right default.

Harris-Benedict: the original, and its 1984 revision

The Harris-Benedict equation dates to 1919, from a biometric study by James Arthur Harris and Francis Gano Benedict at the Carnegie Institution. The original equations are: for men, BMR equals 66.5 plus 13.75 times weight in kilograms plus 5.003 times height in centimetres minus 6.755 times age; for women, BMR equals 655.1 plus 9.563 times weight plus 1.85 times height minus 4.676 times age. They were based on 239 subjects, mostly young men, and they remained the standard for decades simply because nothing better existed.

The 1984 revision by Roza and Shizgal adjusted the coefficients for modern body composition: for men, BMR equals 88.362 plus 13.397 times weight plus 4.799 times height minus 5.677 times age; for women, BMR equals 447.593 plus 9.247 times weight plus 3.098 times height minus 4.330 times age. The revision is the version most calculators mean when they say Harris-Benedict today. Even revised, it tends to overestimate resting energy expenditure by around 5 percent compared with modern measurements, which is why Mifflin-St Jeor displaced it. Harris-Benedict still appears in textbooks, older apps, and clinical settings that have used it for decades, and its figures are usually 5 to 10 percent higher than Mifflin-St Jeor for the same person.

Worked comparison: our 35 year old man, 75 kilograms, 175 centimetres. Mifflin-St Jeor gives 1,674 calories. Revised Harris-Benedict gives 88.362 plus 13.397 times 75 plus 4.799 times 175 minus 5.677 times 35, which is 88.362 plus 1,004.775 plus 839.825 minus 198.695, about 1,734 calories, roughly 60 higher. Original Harris-Benedict gives about 1,713. The differences are modest for one day but add up over months, which is why the choice of equation matters for a sustained plan, and why you should stick to one equation when tracking changes rather than switching between calculators.

Katch-McArdle and Cunningham: formulas that use lean body mass

Both of these formulas start from lean body mass, the weight of everything that is not fat, instead of total weight. Katch-McArdle says BMR equals 370 plus 21.6 times lean body mass in kilograms. Cunningham says resting metabolic rate equals 500 plus 22 times lean body mass in kilograms. Lean body mass itself is total weight times one minus body fat fraction, so a 80 kilogram person at 20 percent body fat has a lean mass of 64 kilograms, giving a Katch-McArdle BMR of 370 plus 21.6 times 64, which is 1,752 calories, and a Cunningham figure of 500 plus 22 times 64, which is 1,908 calories. Cunningham estimates resting metabolic rate rather than strict BMR, which partly explains why it runs higher.

These formulas are better for people whose body composition is far from average: competitive athletes, bodybuilders, and very lean or very muscular people, for whom total weight misleads. Their weakness is practical: they need an accurate body fat percentage, and most home methods, bathroom scales with bioelectrical impedance, skinfold callipers used inexpertly, and visual estimates, are unreliable, often off by several percentage points. A 5 percentage point error in body fat moves lean mass by 4 kilograms for an 80 kilogram person, which moves the Katch-McArdle result by about 86 calories, wiping out its advantage. Use a lean-mass formula only when you have a trustworthy body fat measurement, from a DEXA scan or a skilled practitioner, not from a bathroom scale.

Other formulas you may meet

The Owen equation, from the 1980s, gives separate formulas for men and women based on weight alone: for men, resting metabolic rate equals 879 plus 10.2 times weight in kilograms; for women, 795 plus 7.18 times weight. It is simple but less accurate than Mifflin-St Jeor for most people, and it is mainly of historical interest now. The World Health Organisation, WHO, equations, also called Schofield equations, use weight and age bands, with different coefficients for age groups, and are widely used in nutrition surveys and in clinical settings outside the United States. They are reasonable for population work but offer no advantage for an individual over Mifflin-St Jeor.

The Henry equations, developed in the 2000s from a larger and more diverse dataset than Schofield, are sometimes recommended for specific populations, and the Mifflin-St Jeor equation itself has a version using the unrounded coefficients, 9.99 times weight plus 6.25 times height minus 4.92 times age plus 166 times sex minus 161, where sex is 1 for men and 0 for women. The difference between the rounded and unrounded forms is a few calories, not worth worrying about. If a calculator shows you a figure that differs from another by less than 50 calories, that is rounding and coefficient choice, not a real disagreement.

BMR, RMR, and TDEE: three different numbers

BMR is the strict laboratory figure, measured under the conditions described in our companion guide. RMR, resting metabolic rate, is measured under looser conditions, sitting rather than lying, not necessarily fasted for 12 hours, and it usually comes out a little higher, often by 10 percent or so, because the conditions are less strict. Cunningham and Owen estimate RMR, not BMR, which is one reason their figures run higher than Mifflin-St Jeor. TDEE, total daily energy expenditure, is BMR or RMR plus activity, exercise, digestion, and non-exercise movement, and it is the figure that actually sets how much to eat. Comparing a BMR figure from one calculator with a TDEE figure from another is the commonest source of apparent contradictions online.

Indirect calorimetry, the laboratory gold standard, measures oxygen consumption and carbon dioxide production while a person rests under a hood or in a room calorimeter, and converts gas exchange into energy expenditure. It is accurate but expensive, time-consuming, and unavailable to most people. Doubly labelled water, which tracks isotopes in body water over one to two weeks, measures total daily expenditure in free living people and is the gold standard for TDEE, but it is a research tool, not a consumer one. Every formula in this guide is an attempt to approximate those measurements from four easy numbers, and should be judged on that basis.

Which formula should you actually use

For most healthy adults, use Mifflin-St Jeor. It is the best validated for the general population, it is what our BMR Calculator uses, and it is the figure most dietitians will start from. If you are comparing your result with a clinician or an older app that uses Harris-Benedict, expect theirs to be a little higher and do not treat the difference as an error in either. If you are a competitive athlete or very muscular and you have a trustworthy body fat percentage, try Katch-McArdle as a cross-check: if the two agree within 100 calories, either is fine; if they differ by much more, the lean-mass figure is probably closer for you, provided the body fat figure is real.

Whatever formula you choose, use it consistently. Tracking change matters more than the starting absolute: if you lose 5 kilograms, recalculate with the same formula and watch the figure fall by about 50 calories from weight alone, plus the age effect over time. Switching formulas mid-plan creates an artificial jump that looks like a metabolic change but is only arithmetic. And whatever the formula says, validate it against two to three weeks of weight trend at a known intake, adjusting by 100 to 200 calories as the trend demands. The formula proposes, the scale disposes.

Mistakes in comparing formulas

The first mistake is comparing figures computed with different units, pounds in one calculator and kilograms in another, which produces gaps of hundreds of calories and looks like a formula disagreement. The second is comparing BMR with RMR or TDEE, which are different quantities. The third is entering different weights, a current weight in one and a goal weight in another. The fourth is rounding: calculators that round intermediate steps differ slightly from those that keep full precision. Before concluding that two formulas disagree wildly, recompute both from the same four inputs, in the same units, and check whether each figure is labelled BMR, RMR, or TDEE. Most wild disagreements dissolve at that step.

A note on units and rounding in every formula

Every formula in this guide uses kilograms and centimetres, and every comparison assumes the same units throughout. Mixing units is the fastest way to produce a figure that looks like a formula disagreement but is only a conversion error: a weight entered in pounds where kilograms were expected inflates the result by more than double, and a height entered in inches where centimetres were expected collapses it. Convert once, write the converted figures down, and use them for every formula you compare, so that any remaining difference is genuinely about the equation, not the arithmetic.

Rounding deserves the same care. The published coefficients are themselves rounded versions of fitted values, and calculators differ in whether they round intermediate terms or only the final result. Those choices move the answer by a few calories, never by hundreds, so a gap of less than 50 calories between two calculators using the same formula is not a disagreement worth investigating. A gap of 200 calories or more, on the other hand, almost always means different inputs, different units, or a different quantity being labelled, BMR in one case and TDEE in the other, and it is worth tracing before you trust either figure.

Frequently asked questions

The Bottom Line

Mifflin-St Jeor is the default for a reason: broad validation, simple inputs, and the best accuracy for most adults. Harris-Benedict is history that still lingers in older tools, and lean-mass formulas are specialist tools that only help when body fat is measured properly. Pick one formula, use it consistently, and let your weight trend have the final word.