Mifflin-St Jeor vs Harris-Benedict vs Katch-McArdle: Formula Accuracy Compared

You plug your stats into a TDEE calculator, get a number like 2,450 calories, and start planning meals around it. But behind that single output sits a formula — and not every calculator uses the same one. Mifflin-St Jeor, Harris-Benedict, Katch-McArdle: three equations, three different assumptions about your body, three slightly different answers.

The gap between them can be 200-300 calories per day. Over a week, that is a full day's worth of eating. If you are cutting weight for a competition or trying to gain lean mass on a tight surplus, the formula you choose matters more than you think.

This post breaks down exactly how each formula works, runs the same person through all three with real numbers, and shows you which one to trust based on your situation and the data behind each equation.

What BMR Actually Means (and Why Formulas Differ)

Basal metabolic rate is the energy your body burns at complete rest — lying still, awake, in a temperature-neutral room, having fasted for 12 hours. It covers heartbeat, breathing, cell repair, brain activity, and organ function. For most people, BMR accounts for 60-75% of total daily energy expenditure.

Every BMR formula is a regression equation. Researchers measured actual metabolic rates in a group of people using indirect calorimetry (a sealed hood that captures oxygen consumption and CO2 output), then built a math model that best predicted those measurements from easy-to-collect variables like height, weight, age, and sex.

The formulas differ because they were built from different populations, in different decades, using different statistical methods. Harris-Benedict came from data collected in the 1910s and 1920s. Mifflin-St Jeor used subjects from 1990. Katch-McArdle threw out the demographic variables entirely and used lean body mass as the sole predictor. Each approach carries its own strengths and blind spots.

The Mifflin-St Jeor Equation

Published in 1990 by M.D. Mifflin and S.T. St Jeor, this formula was derived from 498 healthy subjects (247 women, 251 men) ranging from normal weight to obese. It uses four variables: weight in kilograms, height in centimeters, age in years, and sex.

For men: BMR = (10 × weight in kg) + (6.25 × height in cm) - (5 × age in years) + 5

For women: BMR = (10 × weight in kg) + (6.25 × height in cm) - (5 × age in years) - 161

The American Dietetic Association (now the Academy of Nutrition and Dietetics) endorsed Mifflin-St Jeor in 2005 as the most accurate for estimating BMR in both normal-weight and obese individuals. A 2005 validation study by Frankenfield et al. found it predicted BMR within 10% of measured values for 82% of non-obese subjects — the highest accuracy rate among the formulas tested.

Its main limitation: it knows nothing about your body composition. Two men who are both 180 lbs, 5'10", and 30 years old get the same BMR, even if one is 12% body fat and the other is 30%. Muscle burns roughly 6 calories per pound per day at rest; fat burns about 2. That difference adds up.

The Harris-Benedict Equation

James Arthur Harris and Francis Gano Benedict published this formula in 1919, making it the oldest of the three. The original study measured 239 subjects using a respiration calorimeter at the Carnegie Nutrition Laboratory in Boston. In 1984, Roza and Shizgal revised the coefficients using updated data, and most modern calculators use the revised version.

Revised (1984) for men: BMR = 88.362 + (13.397 × weight in kg) + (4.799 × height in cm) - (5.677 × age in years)

Revised (1984) for women: BMR = 447.593 + (9.247 × weight in kg) + (3.098 × height in cm) - (4.330 × age in years)

Harris-Benedict tends to overestimate BMR by 5-15% compared to measured values, especially in obese individuals. The Frankenfield review found it predicted within 10% for only 69% of non-obese subjects and 64% of obese subjects. The 1919 data set skewed toward younger, leaner, more active subjects than today's general population — a demographic mismatch that the 1984 revision only partially corrected.

Despite its lower accuracy, Harris-Benedict remains the default in many clinical nutrition textbooks and older software systems. If a calculator does not specify which formula it uses, there is a reasonable chance it is Harris-Benedict.

The Katch-McArdle Equation

Frank Katch and William McArdle took a fundamentally different approach. Instead of using height, weight, age, and sex, their formula uses a single input: lean body mass (LBM) in kilograms. LBM is your total weight minus your fat mass.

BMR = 370 + (21.6 × lean body mass in kg)

Because it uses LBM, this formula is sex-agnostic — it does not need a separate equation for men and women. The logic is straightforward: metabolically active tissue (muscle, organs, bone) drives resting energy expenditure, so measure that directly instead of guessing from proxy variables.

The catch is obvious: you need to know your body fat percentage to calculate LBM. And body fat measurement itself introduces error. DEXA scans are accurate to within 1-2%, but a consumer bioelectrical impedance scale can be off by 5-8% depending on hydration. Skinfold calipers depend entirely on the skill of the person doing the measurement. Use ProCalc's body fat calculator to get a reasonable estimate before plugging into Katch-McArdle.

For lean, muscular individuals, Katch-McArdle typically outperforms both Mifflin-St Jeor and Harris-Benedict. A 200-lb man at 10% body fat has 180 lbs of lean mass, and Katch-McArdle captures the metabolic impact of that muscle. For people at average or higher body fat levels, though, the formula loses its edge because the LBM estimate itself becomes less reliable.

Worked Example: Same Person, Three Formulas

Meet our test subject: a 30-year-old male, 5'10" (177.8 cm), 180 lbs (81.6 kg), 18% body fat. He is moderately active — lifts weights three times a week and walks daily.

Mifflin-St Jeor

BMR = (10 × 81.6) + (6.25 × 177.8) - (5 × 30) + 5

BMR = 816 + 1,111.25 - 150 + 5 = 1,782 calories/day

Harris-Benedict (Revised 1984)

BMR = 88.362 + (13.397 × 81.6) + (4.799 × 177.8) - (5.677 × 30)

BMR = 88.362 + 1,093.20 + 853.26 - 170.31 = 1,864 calories/day

Katch-McArdle

First, calculate lean body mass: 81.6 kg × (1 - 0.18) = 66.9 kg LBM

BMR = 370 + (21.6 × 66.9) = 370 + 1,445.04 = 1,815 calories/day

The Spread

Mifflin-St Jeor gives 1,782. Harris-Benedict gives 1,864. Katch-McArdle gives 1,815. The range is 82 calories — about 4.6% from lowest to highest. For this particular person (male, healthy weight, moderate body fat), the formulas are reasonably close. The gap widens with more extreme body compositions.

To convert any of these to TDEE, multiply by an activity factor. For moderately active (exercise 3-5 days/week), the standard multiplier is 1.55. That gives a TDEE range of 2,762 to 2,889 calories. Run your own numbers through the TDEE calculator to see where you land.

Head-to-Head Accuracy Comparison

The most cited comparison study is Frankenfield et al. (2005), which tested four BMR equations against indirect calorimetry in 337 healthy adults. Here is how the three formulas performed:

Metric	Mifflin-St Jeor	Harris-Benedict (Revised)	Katch-McArdle
Predicted within 10% of measured (non-obese)	82%	69%	73%*
Predicted within 10% of measured (obese)	70%	64%	68%*
Mean bias (non-obese)	-1% (slight underestimate)	+5% (overestimate)	+2% (slight overestimate)
Mean bias (obese)	-3% (underestimate)	+9% (overestimate)	Varies with BF% accuracy
Year developed	1990	1919 / revised 1984	1975
Inputs required	Weight, height, age, sex	Weight, height, age, sex	Lean body mass only
Best population fit	General adults, both sexes	Younger, leaner populations	Lean/muscular individuals

*Katch-McArdle accuracy depends heavily on the precision of the body fat measurement used to derive LBM. These figures assume DEXA-level accuracy.

A 2013 meta-analysis by Sabounchi, Rahmandad, and Ammerman reviewed 248 studies and confirmed that Mifflin-St Jeor had the lowest overall prediction error for healthy adults across the widest range of body types. Harris-Benedict consistently overestimated, particularly as BMI increased.

Which Formula Should You Use?

The right choice depends on what you know about your body and what you are trying to do.

Use Mifflin-St Jeor if: You do not know your body fat percentage, you are at a roughly average body composition, or you just want the most reliable general-purpose estimate. This is the safe default. If you only run one formula, make it this one.

Use Katch-McArdle if: You know your body fat percentage from a reliable source (DEXA scan, hydrostatic weighing, or skilled caliper measurement), and you are notably lean or notably muscular. If you are a competitive athlete, bodybuilder, or have been resistance training seriously for years, Katch-McArdle will likely give you a more accurate number than Mifflin-St Jeor.

Be cautious with Harris-Benedict if: You are overweight or obese. The formula will likely overestimate your BMR by 100-300 calories, which compounds into real dietary error over weeks. If a clinical setting uses Harris-Benedict (many still do), ask if they have applied an adjustment factor for higher body weights.

Run all three and average if: You want a sanity check. If all three formulas land within 50-100 calories of each other, you have a high-confidence estimate. If one formula diverges sharply from the other two, investigate why — it usually points to body composition being unusual for your height and weight.

How Body Composition Changes Everything

To see where the formulas really diverge, compare two people with identical scale weight but different body compositions.

Variable	Person A (Lean)	Person B (Higher BF%)
Sex	Male	Male
Age	30	30
Height	5'10" (177.8 cm)	5'10" (177.8 cm)
Weight	180 lbs (81.6 kg)	180 lbs (81.6 kg)
Body fat %	10%	28%
Lean body mass	73.4 kg	58.8 kg
Mifflin-St Jeor BMR	1,782 cal	1,782 cal
Harris-Benedict BMR	1,864 cal	1,864 cal
Katch-McArdle BMR	1,955 cal	1,640 cal

Mifflin-St Jeor and Harris-Benedict produce identical results for both people because they only see height, weight, age, and sex. Katch-McArdle sees the 14.6 kg difference in lean mass and produces a 315-calorie gap. For Person A, the lean athlete, Katch-McArdle is almost certainly closer to reality. For Person B, the same formula correctly predicts a lower BMR that the other two equations miss entirely.

This is Katch-McArdle's superpower — and its justification for requiring body fat data. Check your BMI first to see where you fall on the weight-for-height spectrum, then decide if a body fat measurement is worth the effort.

Common Misconceptions About BMR Formulas

"The higher number is safer." Not if you are trying to lose fat. Eating at a Harris-Benedict estimate when Mifflin-St Jeor is more accurate means your deficit is 100+ calories smaller than you think. Over 12 weeks, that is roughly 2.5 lbs of fat loss you expected but did not get.

"These formulas account for exercise." No. All three produce BMR — the energy cost of doing absolutely nothing. To get TDEE, you multiply BMR by an activity factor (typically 1.2 for sedentary up to 1.9 for very active). The activity multiplier is a separate estimation layer with its own error margin. Use a calorie calculator that applies the multiplier for you.

"Body fat percentage does not matter much." At average body compositions (15-25% for men, 20-30% for women), the gap between formulas is modest. Below 12% or above 35% body fat, the formulas diverge significantly, and using one that ignores body composition introduces meaningful error.

"Online calculators all use the same formula." They do not. Some use Harris-Benedict without specifying which revision. Some use Mifflin-St Jeor. A few use Katch-McArdle but default to a guessed body fat percentage if you do not provide one. Always check which equation your calculator uses, and note whether it asks for body fat as an input.

Practical Tips for Getting a Better Estimate

Run Mifflin-St Jeor as your baseline. If you have a reliable body fat number, also run Katch-McArdle. If the two results are within 50 calories, use the average. If they diverge by more than 100 calories, trust whichever formula has better input data for your situation.

Treat any formula output as a starting point, not a prescription. Track your weight and intake for 2-3 weeks, then adjust. If you are losing 1 lb/week at an estimated 500-calorie deficit, your TDEE estimate was roughly correct. If you are not losing, your true TDEE is lower than calculated — reduce by 100-200 calories and reassess.

Recalculate every 10-15 lbs of weight change. BMR drops as you lose weight (fewer cells to fuel) and rises as you gain lean mass. A formula result from 6 months and 20 lbs ago is stale data.

Do not apply two adjustment factors to the same number. If you use Katch-McArdle (which already accounts for body composition via LBM), adding a "lean mass bonus" on top double-counts the effect. Pick one formula and apply the activity multiplier cleanly.

Frequently Asked Questions

Which BMR formula do most doctors and dietitians use?

The Academy of Nutrition and Dietetics recommends Mifflin-St Jeor as the preferred equation for estimating BMR in healthy adults. Many clinical settings still default to Harris-Benedict because it has been embedded in nutrition software for decades, but the evidence favors Mifflin-St Jeor for general accuracy.

Can I use Katch-McArdle if I only know my body fat from a bathroom scale?

You can, but understand the margin of error. Consumer bioelectrical impedance scales can be off by 5-8% on body fat readings depending on hydration, time of day, and recent meals. If your scale says 20% but reality is anywhere from 14% to 26%, the Katch-McArdle output inherits that uncertainty. For a rough estimate, it still beats ignoring body composition entirely. For precision, get a DEXA scan.

Does BMR decrease with age, and do the formulas capture that?

Yes. Both Mifflin-St Jeor and Harris-Benedict include an age term that subtracts calories as you get older — roughly 5 calories per year for Mifflin-St Jeor. This reflects the well-documented decline in metabolic rate with aging, driven largely by loss of lean tissue. Katch-McArdle captures this indirectly: if you lose muscle mass as you age, your LBM drops, and the formula outputs a lower BMR.

Are any of these formulas accurate for children or teenagers?

No. All three were derived from adult populations. Pediatric BMR estimation requires age-specific equations like Schofield (1985) or Henry (2005), which account for the metabolic demands of growth. Do not apply adult formulas to anyone under 18.

How much does the activity multiplier matter compared to the BMR formula choice?

The activity multiplier introduces far more potential error. The difference between "lightly active" (1.375) and "moderately active" (1.55) on a BMR of 1,800 calories is 315 calories/day. The typical gap between Mifflin-St Jeor and Harris-Benedict is 50-100 calories. Choosing the right activity level matters more than choosing the right formula — though ideally you get both right.

Should I use a different formula when bulking versus cutting?

The formula itself does not change based on your goal. What changes is how you use the output. For cutting, many coaches recommend using the lower estimate (typically Mifflin-St Jeor) to ensure the deficit is real. For bulking, using a slightly higher estimate reduces the risk of excessive fat gain from an oversized surplus. Either way, real-world tracking over 2-3 weeks gives you better data than any formula alone.

What about the Cunningham equation? Is it different from Katch-McArdle?

The Cunningham equation (BMR = 500 + 22 × LBM in kg) is similar to Katch-McArdle but uses a higher intercept and coefficient, developed specifically for athletic populations. Some sports nutrition practitioners prefer Cunningham for competitive athletes, but Katch-McArdle remains more widely cited. The difference between them is typically 50-130 calories.

Sources: Mifflin MD, St Jeor ST, et al. "A new predictive equation for resting energy expenditure in healthy individuals." American Journal of Clinical Nutrition, 1990;51(2):241-247. Frankenfield D, Roth-Yousey L, Compher C. "Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults." Journal of the American Dietetic Association, 2005;105(5):775-789. Sabounchi NS, Rahmandad H, Ammerman A. "Best-fitting prediction equations for basal metabolic rate: informing obesity interventions in diverse populations." International Journal of Obesity, 2013;37(10):1364-1370. Harris JA, Benedict FG. "A Biometric Study of Human Basal Metabolism." Proceedings of the National Academy of Sciences, 1918;4(12):370-373.