Calorie Prediction Equations

Estimating the number of calories an individual requires each dayoften called total daily energy expenditure (TDEE)is a cornerstone of nutrition planning, weight management, and clinical assessment. Because direct measurement (e.g., indirect calorimetry) is expensive and impractical for most people, a variety of predictive equations have been developed. This page surveys the most widely used formulas, explains their origin, outlines when each is appropriate, and highlights their strengths and limitations.

1. The Foundations: Basal Metabolic Rate (BMR)

All calorieprediction equations begin with an estimate of basal metabolic rate, the energy needed to keep the body functioning at rest (breathing, circulation, cellular processes). BMR is expressed in kilocalories per day (kcal/d) and is influenced by age, sex, body mass, height, and body composition.

1.1 Classic HarrisBenedict Equation (1919)

One of the earliest and most cited formulas, the HarrisBenedict equation, was derived from data on 239 healthy adults. The original equations are:

Sex	Equation
Male	BMR = 66.5 + (13.75weightkg) + (5.003heightcm) (6.775ageyr)
Female	BMR = 655.1 + (9.563weightkg) + (1.850heightcm) (4.676ageyr)

Although widely used, the original values tend to overestimate BMR for modern, more sedentary populations.

1.2 Revised HarrisBenedict (1984)

Roza and Shizgal adjusted the coefficients based on newer data, improving accuracy:

Sex	Equation
Male	BMR = 88.362 + (13.397weightkg) + (4.799heightcm) (5.677ageyr)
Female	BMR = 447.593 + (9.247weightkg) + (3.098heightcm) (4.330ageyr)

1.3 MifflinSt Jeor Equation (1990)

Developed from a study of 498 subjects, this equation is currently considered the most accurate for nonobese adults:

Sex	Equation
Male	BMR = (10weightkg) + (6.25heightcm) (5ageyr) + 5
Female	BMR = (10weightkg) + (6.25heightcm) (5ageyr) 161

1.4 Owen Equation (1988)

Based on a large dataset of measured resting metabolic rates, Owens simple weightonly formula can be useful when height is unavailable:

Male: BMR = 879+10.2weightkg
Female: BMR = 795+7.18weightkg

2. From BMR to Total Daily Energy Expenditure (TDEE)

To move from basal metabolism to the total calories burned in a day, an activity factor (also called a Physical Activity Level, PAL) is applied. The factor represents the multiplier for the energy cost of all activities beyond resting.

Activity Level	PAL
Sedentary (little or no exercise)	1.21.3
Lightly active (light exercise 13days/week)	1.41.5
Moderately active (moderate exercise 35days/week)	1.61.7
Very active (hard exercise 67days/week)	1.81.9
Extra active (very hard physical job or training twice/day)	2.02.4

Thus, TDEE = BMRPAL. The choice of PAL should reflect the individual's typical routine; many calculators ask for a description of weekly activity to select the most appropriate multiplier.

3. Equations Tailored for Specific Populations

3.1 Cunningham Equation (1980)

Emphasizes lean body mass (LBM), making it valuable for athletes and those with atypical body composition.

Cunningham: BMR = 500 + 22LBM (kg)

LBM can be estimated with skinfolds, bioelectrical impedance, or DEXA scans.

3.2 Schofield Equation (1985)

Developed for the World Health Organization, the Schofield equations are agespecific and are often used in publichealth settings.

Examples:

• Men 1830yr: BMR = 15.057weightkg + 692.
• Women 3160yr: BMR = 8.126weightkg + 845.

3.3 WHO/FAO/UNU Equations (2004)

The Food and Agriculture Organization updated predictive formulas for children and adolescents, reflecting growth-related energy needs. They are beyond the adult focus of this page but worth mentioning for completeness.

4. Choosing the Right Equation

There is no universally best formula. The decision depends on the target group, the required precision, and the data available:

1. General adult population MifflinSt Jeor (or revised HarrisBenedict) is recommended.
2. Obese individuals Some clinicians adjust weight input (e.g., using ideal body weight or adjusted body weight) before applying MifflinSt Jeor to avoid overestimation.
3. Athletes or people with high muscle mass Cunningham, which incorporates lean mass, gives a closer estimate.
4. Elderly Agespecific equations like Schofield or the revised HarrisBenedict may be more appropriate.
5. When only weight is known Owens equation provides a quick, though less precise, estimate.

5. Limitations and Sources of Error

• Population bias: Most formulas are derived from Western, middleclass samples; they may misrepresent other ethnicities or socioeconomic groups.
• Body composition: Two individuals with identical height, weight, and age can have markedly different BMRs if one carries more muscle and the other more fat.
• Temperature and climate: Cold environments increase thermogenesis, while warm climates may reduce it.
• Health status: Hyperthyroidism, fever, and certain medications raise metabolic rate; chronic disease can lower it.
• Selfreported data: Inaccurate weight or height inputs dramatically affect the output.

Because of these variables, predicted calories should be treated as starting points. Monitoring actual weight change and adjusting intake accordingly remains the gold standard.

6. Practical Example

Consider a 30yearold woman weighing 68kg, 165cm tall, moderately active.

1. Calculate BMR using MifflinSt Jeor:
   BMR = (1068) + (6.25165) (530) 161 = 680 + 1031.25 150 161 = 1,400.25kcal/day
2. Select PAL for moderate activity (1.65).
   TDEE = 1,4001.65 2,310kcal/day

If she wishes to lose weight, a common approach is to create a 500kcal/day deficit, aiming for roughly 0.5kg weight loss per week.

7. Using Online Calculators Wisely

Many websites embed these equations in simple forms. When using them:

• Check which formula is employed and whether it matches your needs.
• Enter precise measurements (preferably measured, not estimated).
• Remember that the activity factor is a rough estimate; adjust based on personal experience.

For clinicians, integrating predictive equations into electronic health records can streamline dietary counseling, but they should still verify calculations against clinical judgment.

8. Future Directions

Advances in wearable technology, machine learning, and metabolomics are paving the way for personalized energyexpenditure models that incorporate realtime heartrate variability, sleep patterns, and even genetic markers. Until such tools become universally accessible, the classic prediction equations will remain valuable, especially when used with an understanding of their context and limitations.

For further reading, see the original publications by Harris & Benedict (1919), Mifflin etal. (1990), and Cunningham (1980). Professional societies such as the Academy of Nutrition and Dietetics regularly update guidelines based on emerging evidence.