BMR Formula Selection: Which Equation Should You Use?
Use Mifflin-St Jeor unless you have an accurate body fat measurement — then use Katch-McArdle. Harris-Benedict overestimates BMR by 5–8%. Cunningham overestimates for women by 3–5%. The Deurenberg formula is only for estimating body fat from BMI, not BMR. This guide compares all 5 formulas with real calculations, shows when each is more accurate, and gives you a decision tree to pick the right one in 30 seconds.
The Five BMR Formulas Compared
| Formula | Year | Uses | Accuracy | Best For |
|---|---|---|---|---|
| Mifflin-St Jeor | 1990 | Weight, height, age, sex | ±5% | Most people (default choice) |
| Katch-McArdle | 1996 | Lean body mass | ±3–4% | If you know body fat % |
| Harris-Benedict | 1919 | Weight, height, age, sex | ±7–10% | Historical reference only |
| Cunningham | 1991 | Lean body mass | ±4–5% | Athletes with known LBM |
| Oxford (Henry) | 2005 | Weight, height, age, sex | ±5–6% | Children and elderly |
Side-by-Side Calculation: Same Person, All Formulas
Subject: Male, 30, 5'10" (178cm), 175 lbs (79.5 kg), 15% body fat
Lean Body Mass (LBM): 79.5 × (1 − 0.15) = 67.6 kg
| Formula | Equation | BMR Result | vs Mifflin |
|---|---|---|---|
| Mifflin-St Jeor | 10×79.5 + 6.25×178 − 5×30 + 5 | 1,725 cal | — |
| Katch-McArdle | 370 + 21.6×67.6 | 1,830 cal | +105 (+6.1%) |
| Harris-Benedict | 88.4 + 13.4×79.5 + 4.8×178 − 5.68×30 | 1,858 cal | +133 (+7.7%) |
| Cunningham | 500 + 22×67.6 | 1,987 cal | +262 (+15.2%) |
| Oxford (Henry) | 15.1×79.5 + 6.91×178 − 1.44×30 + 52 | 1,839 cal | +114 (+6.6%) |
Spread: 262 calories between lowest and highest. Over a year at "maintenance," that's 27.4 lbs of weight difference.
Same Subject at 25% Body Fat
LBM: 79.5 × (1 − 0.25) = 59.6 kg
| Formula | BMR Result | vs Mifflin |
|---|---|---|
| Mifflin-St Jeor | 1,725 cal | — |
| Katch-McArdle | 1,657 cal | −68 (−3.9%) |
| Harris-Benedict | 1,858 cal | +133 (+7.7%) |
| Cunningham | 1,811 cal | +86 (+5.0%) |
Key insight: At 15% body fat, Katch-McArdle gives a higher BMR than Mifflin. At 25%, it gives a lower BMR. This is because Katch-McArdle scales with lean mass, not total weight — so it rewards muscle and penalizes fat.
The Decision Tree: Which Formula to Use
Do you have a reliable body fat measurement (DEXA or experienced skinfold)?
│
├── YES → Use KATCH-McARDLE
│ ├── Most accurate when body fat % is known (±3–4%)
│ ├── Formula: BMR = 370 + (21.6 × LBM in kg)
│ └── LBM = weight × (1 − body fat %)
│
└── NO → Use MIFFLIN-ST JEOR
├── Best formula when body fat is unknown (±5%)
├── Formula: BMR = 10×weight(kg) + 6.25×height(cm) − 5×age + 5 (men) or −161 (women)
└── Recommended by the American Dietetic Association (2005 review)
NEVER USE:
├── Harris-Benedict (1919) — overestimates by 5–8%, outdated
├── Cunningham — overestimates for non-athletes and women
└── Any "proprietary" formula that doesn't disclose its equation
When Each Formula Is Best (and Worst)
Mifflin-St Jeor (1990)
Best for: General population, overweight/obese individuals, people who don't know body fat %
Worst for: Very lean athletes (<10% body fat) — underestimates by 2–5%
Why it's the default: The ADA's 2005 meta-analysis (Frankenfield et al.) compared 4 BMR formulas against indirect calorimetry in 309 adults. Mifflin-St Jeor had the smallest bias (+2.3% overestimate) and tightest error range (±5%).
Katch-McArdle (1996)
Best for: Anyone with an accurate body fat measurement, athletes, bodybuilders
Worst for: People using BMI-estimated or BIA-scale body fat % (garbage in, garbage out)
Critical dependency: If your body fat measurement is off by 5%, your BMR is off by ~50–80 calories. Use this formula ONLY if you have DEXA or 3-site skinfold measurements.
Harris-Benedict (1919, revised 1984)
Best for: Historical comparison, nothing else
Worst for: Everyone — it was developed on 239 young, active men and 136 women in early 1900s. Modern populations are heavier, older, and less active.
The problem: It systematically overestimates BMR by 5–8% because:
- Original subjects were younger (avg 27 for men, 33 for women)
- Original subjects were more active (many were athletes, students, and laborers)
- It uses total weight, not lean mass — so it gives "credit" for fat weight
Cunningham (1991)
Best for: Endurance athletes with known LBM
Worst for: Women, sedentary individuals, anyone without accurate LBM
Why: Cunningham uses a higher LBM coefficient (22 vs Katch-McArdle's 21.6) and a higher constant (500 vs 370). This makes it ~8–10% higher than Katch-McArdle for the same person. For female athletes, it overestimates by 3–5%.
Female-Specific Calculation Example
Subject: Female, 28, 5'6" (168cm), 145 lbs (65.8 kg), 28% body fat
LBM: 65.8 × (1 − 0.28) = 47.4 kg
| Formula | BMR Result | Likely Accuracy |
|---|---|---|
| Mifflin-St Jeor | 1,414 cal | ±70 cal (best default) |
| Katch-McArdle | 1,394 cal | ±42 cal (best if BF% is accurate) |
| Harris-Benedict | 1,520 cal | ±106 cal (overestimates) |
| Cunningham | 1,543 cal | ±77 cal (overestimates for women) |
Mifflin and Katch-McArdle agree within 20 calories — this gives high confidence. Harris-Benedict and Cunningham are 100–130 cal higher, which would create a significant calorie surplus if used for deficit planning.
Execution Checklist
-
Use Mifflin-St Jeor as your default. It's the formula validated on the largest, most diverse population and recommended by the ADA. Most good calculators use it.
-
Switch to Katch-McArdle if you have DEXA or skinfold data. It's more accurate but ONLY if your body fat measurement is reliable. If you're using a BIA scale's body fat %, stick with Mifflin.
-
Check which formula your calculator uses. Many "TDEE calculators" silently use Harris-Benedict. Look for "Mifflin-St Jeor" or "Mifflin" in the FAQ or methodology page. If it doesn't say, assume Harris-Benedict and subtract 5% from your result.
-
For women, avoid Cunningham. It overestimates BMR by 3–5% for female subjects in validation studies. Mifflin-St Jeor or Katch-McArdle are better choices.
-
For athletes under 10% body fat (men) or 16% (women): Katch-McArdle with DEXA-measured body fat is your best option. Mifflin underestimates BMR for very lean individuals because it doesn't account for the higher metabolic activity of muscle tissue per unit weight.
Common Mistakes (What Competitors Get Wrong)
❌ Using Harris-Benedict without telling users
Competitors say: Nothing — they silently use Harris-Benedict (or a "proprietary" formula that's actually Harris-Benedict with minor tweaks)
Reality: Harris-Benedict overestimates BMR by 5–8%. If your calculator gives 1,858 when Mifflin gives 1,725, you'll eat 133 extra calories per day "at maintenance" — gaining ~1 lb per month without understanding why. Always check which formula your calculator uses.
❌ "All BMR formulas are basically the same"
Competitors say: "The difference between formulas is negligible" (some calculator FAQ pages)
Reality: The spread between formulas can be 250+ calories (see Cunningham vs Mifflin above). Over a year, that's 26 lbs. The difference is not "negligible" — it's the difference between maintaining, gaining, or losing weight.
❌ Using Katch-McArdle with inaccurate body fat data
Competitors say: "Katch-McArdle is more accurate, so use it whenever possible"
Reality: Katch-McArdle is only more accurate IF your body fat measurement is accurate. If you plug in a BIA-scale body fat of 18% when your real body fat is 22%, your Katch-McArdle BMR will be off by ~80 calories — worse than Mifflin-St Jeor, which doesn't need body fat at all. Garbage in, garbage out.
❌ Not adjusting for age
Competitors say: Some calculators don't include age in the formula
Reality: BMR decreases ~1–2% per decade after age 20, primarily due to lean mass loss. Mifflin-St Jeor accounts for this (−5 cal per year of age). A calculator that ignores age will overestimate BMR by ~50 cal for a 40-year-old and ~100 cal for a 60-year-old.
❌ Applying activity multipliers to the wrong base
Competitors say: Some sites calculate "TDEE" directly without showing BMR, making it impossible to verify
Reality: TDEE = BMR × activity multiplier. If the BMR is wrong (wrong formula), the TDEE is wrong by the same percentage. Always verify your BMR first, then apply the multiplier yourself if needed.
Related Tools
- BMR Calculator — Uses Mifflin-St Jeor with option to switch formulas
- Maintenance Calories Calculator — BMR × activity multiplier
- TDEE from History — Calculate TDEE from your actual weight data
- Body Fat Estimator — Get body fat % for Katch-McArdle
- Lean Body Mass Calculator — Calculate LBM for Katch-McArdle
- Macro Calculator — Convert BMR into macro targets