Mifflin-St Jeor, Harris-Benedict or Katch-McArdle: which BMR formula should you use?

The four equations side by side

Weight is in kilograms, height in centimetres, age in years and lean body mass (LBM) in kilograms. Each result is kilocalories a day at complete rest.

Mifflin-St Jeor (1990)
  Men:    (9.99 × weight) + (6.25 × height) − (4.92 × age) + 5
  Women:  (9.99 × weight) + (6.25 × height) − (4.92 × age) − 161

Harris-Benedict, revised by Roza and Shizgal (1984)
  Men:    88.362 + (13.397 × weight) + (4.799 × height) − (5.677 × age)
  Women:  447.593 + (9.247 × weight) + (3.098 × height) − (4.330 × age)

Katch-McArdle:  370 + (21.6 × LBM)
Cunningham:     500 + (22 × LBM)

LBM = weight × (1 − body fat % ÷ 100)

The original Harris-Benedict coefficients from 1918 and 1919 (constants of 66.5 for men and 655.1 for women) still turn up in older textbooks. For both clients below they land within 20 kcal of the 1984 revision. Katch-McArdle and Cunningham ignore age and sex completely. Research suggests that once lean body mass is known, age and sex add little to the picture of how much energy the body uses.

Same clients, different answers

Here are two clients run through every formula. The first is a 38-year-old woman, 72 kg, 165 cm, 32% body fat (about 49 kg of lean mass). The second is a 28-year-old man, 90 kg, 180 cm, 12% body fat (about 79 kg of lean mass) who lifts four times a week.

For the woman, four of the five sit within about 70 kcal of each other. That's well inside the error of any prediction equation, so the choice barely matters. For the man, the spread is 350 kcal before you multiply by an activity factor of 1.76, which turns it into more than 600 kcal a day. With a client like him, the formula decides whether his target is right.

Which one to pick

• Mifflin-St Jeor for healthy adults when you don't have a reliable body fat figure. A 2005 systematic review found it the most accurate of the common equations for both non-obese and obese adults.
• Katch-McArdle when you have body fat from DEXA, a good skinfold protocol or a validated bioimpedance device, and the client is lean or trains hard. Mifflin-St Jeor tends to underestimate these clients.
• Cunningham for the same clients as Katch-McArdle, but expect it to read about 150 kcal higher. It's popular in sport, where under-fuelling costs more than a small surplus.
• Harris-Benedict when you need continuity with older records, a referral letter or a hospital protocol that uses it. Otherwise it has no advantage over Mifflin-St Jeor.
• Measured resting metabolic rate from indirect calorimetry beats every equation. If a client brings one, use it.

One warning on lean-mass formulas. They're only as good as the body fat number you feed them, and a bathroom scale reading can be out by several percentage points. A bad body fat figure makes Katch-McArdle worse than Mifflin-St Jeor, not better.

When none of them fit

None of these equations were built for children, pregnancy, critical illness or clients with a condition or medication that changes metabolism. For children and teenagers, use Schofield. In hospital settings, follow your unit's protocol or measure.

Whichever formula you choose, treat it as a starting point. After two to four weeks of honest logging, the client's weight trend tells you more than the equation did. If weight holds steady on a target that should produce a loss, adjust by 100 to 200 kcal and review again. There's more on the default equation in the Mifflin-St Jeor equation.