BMR estimation equations can miss your individual metabolism

Does adding body composition make an equation personal?

People with similar height and weight need not have the same relationship between predicted and measured metabolism. A study of 3,001 adults attending nutrition clinics in Israel compared common RMR equations and a new model after measuring metabolism by indirect calorimetry and assessing body composition. Body composition describes the body in terms such as fat mass and fat-free mass. Here, fat-free mass means body mass excluding fat.[17]

The participants had a mean age of 41 years and a mean BMI of 28.5; 48% were men. Every existing equation compared in this population underestimated RMR. Mifflin-St Jeor’s mean bias was −12.6%.[17] That unfavorable result matters alongside the general adult review’s favorable assessment.[15] The name of a well-regarded equation alone cannot establish either the direction or the size of error in another group.

The new model included sex, age, fat mass, and fat-free mass. The researchers divided participants into a development group of 2,251 and a validation group of 750. The model’s mean bias was 0.7%, with 95% limits of agreement from −18.6 to 19.7%.[17] Body composition contributed to the model, but the individual range remained important. Better average agreement did not make each participant’s prediction identical to their measurement.

The model explained 73.5% of variation.[17] This does not mean it was exactly right for 73.5% of people. Explained variation describes how much of the variability the model accounted for; an individual accuracy rate counts people falling within a defined distance from their measured result. If a description uses only the word “accuracy,” the next question is what its percentage actually measures. Similar-looking percentages can answer substantially different questions.

The study assessed body composition using dual-energy X-ray absorptiometry.[17] Applying its findings to a device at home therefore requires attention to how the input measurements were obtained. Fat mass and fat-free mass are not simply extra boxes that guarantee the same performance wherever their values come from. Their measurement method is part of the model’s validation conditions, just as the participant group is part of those conditions.

Development and validation also need to be distinguished. Performance on the data used to construct an equation does not by itself establish performance in new people. The Chinese adult study’s separate small validation sample and call for more research show why the breadth of testing matters.[1] A new equation may sound more tailored to you, yet its label cannot replace evidence about where, how, and in whom it has been evaluated.

Age, living circumstances, and athletic background matter

An international dataset of 1,686 adults aged 65 and over was used to develop a new RMR equation using height, weight, age, and sex. It improved overall performance only marginally compared with existing equations. The population mean bias was about 50 kJ/day, or about 1%, while individual limits of agreement were approximately ±25%.[6] The favorable average and the less favorable individual performance are both necessary to understand the result.

Among adults aged 80 and over, the mean error was about 100 kJ/day, or about 2%, and accuracy was lower than in younger older adults. The authors nevertheless judged this population-level accuracy to remain within the clinically acceptable range.[6] This paper reports energy in kJ, whereas several other papers report kcal. Comparing the numerical values without their units would be misleading. Equally, an equation’s performance in older adults does not provide an automatic benchmark for the metabolism of someone in their working years.

Living circumstances add another distinction. A systematic review of older adults in long-term care facilities included 4 studies and 332 participants. Most equations tended to overestimate RMR at group level, and individual accuracy did not exceed 41%.[7] “Older adults” is therefore too broad a label to settle applicability. Evidence from people living in the community cannot simply be substituted for evidence from residents of care facilities.

This review does not prove that living in a facility causes prediction error. It brought together observational comparisons between measured and estimated RMR.[7] The finding supports examining equation performance in groups with different characteristics, not a causal explanation of the difference. When thinking about an older relative’s meals, the accuracy rate from your own calculator’s adult validation study cannot automatically be transferred to that relative’s circumstances.

A separate cross-sectional study of 182 residents of long-term care facilities tested 22 equations. Individual accuracy ranged from 27 to 45%, and mean bias ranged from −6 to +18%. The authors proposed treating predicted expenditure as an estimate rather than a fixed prescription, checking it against short-term body weight, intake, and clinical status.[14] That is a recommendation in the context of care-facility residents, not evidence from a trial of dietary adjustment in general adults.

Athletes present a different population problem. A systematic review and meta-analysis covering 29 studies and 1,430 adult athletes found that Ten-Haaf predictions were within ±10% of measured RMR in 80.2%. Other equations whose individual accuracy could be pooled ranged from 40.7 to 63.7%.[4] The authors nevertheless concluded that no single equation was guaranteed to be superior in every situation. The best pooled result still needs its population and conditions attached.

Several equations had average predicted values that were not significantly different from measurements, but many showed large differences across studies. Sex, athlete characteristics, fasting before measurement, and measurement methods influenced some equations’ performance.[4] An equation described as “for active people” leaves much unsaid. A person attending exercise sessions after work should consider whether the evidence actually concerns people with similar physical characteristics, sports backgrounds, and testing conditions.

For an adult who exercises after work, the athlete results need to be read with the difference in participants in mind. Switching a general calculator to an athlete equation simply because it had a high accuracy rate would skip that comparison.[4] Your everyday exercise routine and the validation of an equation in athletes are different questions. Keeping them separate helps you choose which evidence is relevant to your own decisions.

Common misconceptions

“No significant difference means my result is correct.” Statistical comparisons of averages do not establish agreement for every individual. In the Chinese study, predicted and measured BMR were not significantly different in the validation sample, while the development analysis showed wide limits of agreement.[1] The questions are separate: whether group averages differ detectably, and how far a particular prediction might be from a measurement. A calculator’s evaluation needs more than the first answer.

“If most people were within 10%, everyone else was probably only slightly outside.” The community-living adult study reported 82% within that range for Mifflin-St Jeor, but the reported accuracy rate does not describe how large errors were among those outside it.[16] Read the boundary used to define an accurate prediction, then look for measures of the remaining errors. The threshold is a rule for evaluating the equation, not evidence that every result clustered just beside it.

“Averaging several calculators will give my true metabolism.” In the Israeli clinic population, all existing equations tested underestimated RMR; in the Belgian clinic, the direction of bias differed between equations.[2][17] Combining predictions cannot, by itself, establish agreement with your measured metabolism. Before opening more calculator pages, check what comparison measurement and population support the estimates already on screen. Agreement among calculators is not the same test as agreement with indirect calorimetry.

“If the number does not explain my weight, my metabolism is broken.” These papers examine agreement between equation predictions and measurements.[15][16] They do not diagnose the reason for a mismatch between your weight trajectory and a displayed estimate. Understanding that mismatch also requires information about food intake and health circumstances. If you feel unwell or have health concerns, discuss them with a doctor rather than infer a diagnosis from a calculator’s performance.

What you can do today

• When deciding lunch or dinner portions, do not use the displayed BMR as your whole-day food ceiling. In the older-adult paper, RMR accounted for 50–70% of total energy requirements.[6] Keep energy at rest separate from the needs of your whole day, and check the population before applying a percentage. The figure on the screen does not represent all the energy needed for daily life.
• Do not cut staples or side dishes just because another calculator gives a lower number. Mifflin-St Jeor predictions were within 10% of measured RMR in 82% of the community-living adults studied.[16] That 10% was the study’s accuracy threshold, rather than a guarantee about your own error. Before changing a meal, distinguish a difference between equations from a difference established by measurement.
• When considering an older family member’s portions, check whether the evidence covers their age and living circumstances. Individual accuracy did not exceed 41% in the review of long-term care residents.[7] The separate study of 182 residents proposed treating predictions as estimates and checking them against short-term weight, intake, and clinical status.[14] A calculator result alone is insufficient grounds for fixing that person’s food allowance.
• Save the equation’s name, height, weight, and age inputs, including dates when the inputs change, alongside food intake and weight over time. In the Belgian clinic’s 731 participants, the highest accuracy was 73%.[2] This is a group result, rather than your personal probability of a correct value. Bring these records to discussions about a family member’s meals in a care facility, where individual prediction accuracy was low.[7]
• If your weight trajectory and the estimate do not fit and you have health concerns, show a doctor the calculation, inputs, and course of food intake and weight. Ask whether measurement would be useful. The general adult review identifies valid indirect calorimetry, with suitable procedures, as a way to avoid prediction errors when clinical judgment indicates that estimation is inadequate.[15]

Make room for an exercise habit at On the Shore Tachikawa

A BMR calculation is an estimate; putting exercise on your calendar is a separate decision. When reviewing your food and weight over time, On the Shore Tachikawa can be a place to keep an exercise habit within your work and household schedule.

The lava-stone hot yoga studio is open every day 8:00–23:30. Its address is 3F Etoile Bldg, 2-14-10 Akebonocho, Tachikawa, Tokyo, a 1-minute walk from JR Tachikawa Station’s North Exit. Compare the studio information with the prices when choosing a time and activity.

Yoga includes more than 25 kinds of lessons, ranging from hot yoga on lava-stone plates to room-temperature yoga. Alongside yoga, the choices include HIIT, personal training, Pilates, boxercise, and women-only kickboxercise. For a yoga visit, the trial costs ¥1,980 including tax: a 60-minute lesson with mat rental, 2 bath towels, and 1 face towel. Once you have chosen a time, use the trial booking page.

Participation is unavailable to anyone whose doctor has told them not to exercise. During pregnancy, guests cannot take part in lava-stone hot yoga; the studio offers room-temperature maternity yoga. Many of our yoga instructors speak English, and the studio’s owner often helps guests from the US bases in English herself.

References

1. Predictive Equation for Basal Metabolic Rate in Normal-Weight Chinese Adults — Xiaojing Wang et al., 2023, Nutrients. DOI: 10.3390/nu15194185
2. Basal metabolic rate using indirect calorimetry among individuals living with overweight or obesity: The accuracy of predictive equations for basal metabolic rate. — Kristof Van Dessel et al., 2024, Clinical nutrition ESPEN. DOI: 10.1016/j.clnesp.2023.12.024
3. Adequacy of basal metabolic rate prediction equations in individuals with severe obesity: A systematic review and meta‐analysis — Virginia Gaissionok Mariz et al., 2024, Obesity Reviews. DOI: 10.1111/obr.13739
4. Accuracy of Resting Metabolic Rate Prediction Equations in Athletes: A Systematic Review with Meta-analysis — J. E. R. O’Neill et al., 2023, Sports Medicine (Auckland, N.z.). DOI: 10.1007/s40279-023-01896-z
5. Development and validation of new predictive equations for resting metabolic rate (RMR) of older adults, aged 65 years and over. — J. Porter et al., 2023, The American journal of clinical nutrition. DOI: 10.1016/j.ajcnut.2023.04.010
6. To measure or to estimate? A comparison between indirect calorimetry and predictive equations for resting metabolic rate in institutionalized older adults: a systematic review. — T. Piassa et al., 2026, Archives of gerontology and geriatrics. DOI: 10.1016/j.archger.2026.106408
7. Comparing predictive energy-expenditure equations with indirect calorimetry in older adults living in long-term care facilities. — Katarzyna Zadka et al., 2026, Experimental gerontology. DOI: 10.1016/j.exger.2026.113037
8. Comparison of predictive equations for resting metabolic rate in healthy nonobese and obese adults: a systematic review. — D. Frankenfield et al., 2005, Journal of the American Dietetic Association. DOI: 10.1016/j.jada.2005.02.005
9. Bias and accuracy of resting metabolic rate equations in non-obese and obese adults. — D. Frankenfield, 2013, Clinical nutrition. DOI: 10.1016/j.clnu.2013.03.022
10. Cross-Validation of a New General Population Resting Metabolic Rate Prediction Equation Based on Body Composition — Aviv Kfir et al., 2023, Nutrients. DOI: 10.3390/nu15040805

Cover photo: A digital bathroom scale with a display for showing body weight. (Photo: Qurren / CC BY-SA 4.0 / Wikimedia Commons)

Information as of October 2026. For your own health, consult a doctor.

More from Diet Science (English)

• Metabolic adaptation and why your weight loss can slow down
• Body composition and weight loss beyond the bathroom scale
• Food logging and weight loss: what matters beyond perfect records
• Protein and weight loss beyond fullness and the bathroom scale
• Dietary fiber and weight loss: what gut research really shows
• Low-carb vs low-fat diets: what long-term weight loss shows
• Time-restricted eating and weight loss: what trials reveal
• Breakfast and weight loss: what eating or skipping really changes
• Sleep deprivation and appetite: why nighttime snacks add up
• Stress and emotional eating beyond the cortisol explanation
• Aerobic vs resistance training for fat loss and lean mass
• HIIT and body fat loss: what shorter workouts can really offer
• Strength training, metabolism and the science of weight regain
• Yoga and weight loss: what body composition research shows
• Hot yoga calories and sweat: what a lighter scale really means
• Pilates and body composition: what posture and fat studies show
• NEAT and weight loss: what everyday movement can really do
• Daily steps and weight loss: what 8,000 steps really tells us
• Keeping weight off: what long-term maintainers do and how they recover
• Daily self-weighing and weight loss: how to use the numbers
• Water before meals and weight loss: what 500 ml trials show
• Sugary drinks and sweeteners: what replacement means for weight
• Alcohol and weight loss through calories, appetite and fat use
• Ultra-processed foods and overeating through the trial evidence
• Eating rate and weight loss: what slower chewing can change
• Menopause weight gain and why your waist can change first
• Postpartum weight loss through diet, exercise and support
• Sarcopenia and protein: protecting muscle as you grow older
• Visceral fat in men and what exercise changes beyond weight
• Supervised exercise and weight loss: building lasting habits
• GLP-1 drugs and exercise for muscle and weight maintenance
• Weight cycling and metabolism: what 23 studies actually found
• Low GI diets and weight loss: blood sugar is only part of the story
• Mediterranean diet trials separate weight loss from fat loss
• Plant-based diets and weight loss depend on what trials compare
• Snacking frequency alone cannot tell you how weight will change
• Late eating and weight gain: what meal timing trials show
• Portion size and weight loss: change the serving before the plate
• Eating out and weight loss: what meal frequency really tells us
• Menu calorie labeling changes some orders, but results vary
• Coffee and caffeine affect metabolism more clearly than weight
• Green tea catechins and weight loss: what the numbers mean
• Fat burner supplements can raise metabolism without fat loss
• Meal replacements for weight loss and the return to regular food
• Probiotics show small weight differences, not a uniform effect
• Genetics and obesity: what FTO studies say about weight loss
• Weight stigma and self-criticism can accompany daily barriers
• Exercise and appetite: why hunger and eating back calories differ
• Weekend warrior exercise and what it means for weight loss
• Exercise timing and weight loss: morning or evening workouts?
• Fat-burning zone measures fuel use, not future weight loss
• Interval walking for weight loss: what the trials really show
• Stair climbing for weight loss: what fitness studies really show
• Holiday weight gain and why it may still be there in spring
• Menstrual cycle and weight loss: reading hunger and the scale
• Energy density and weight loss without shrinking every meal
• Whole grains and weight loss depend on more than a staple swap
• Nuts and weight loss: why high fat does not mean weight gain
• Body fat measurement accuracy in home scales and DXA scans
• Wearables and weight loss: why more steps may not move the scale
• Alternate-day fasting and 5:2 compared with daily dieting
• Ketogenic diets and weight loss beyond the first few months
• Dairy and calcium have different results in weight-loss trials
• Legumes and weight loss: what fullness can and cannot tell us
• Fish and omega-3 offer no consistent shortcut to weight loss
• Fruit and vegetables for weight loss: what adding more changes
• Meal sequence and blood sugar: what vegetables first can change
• Distracted eating and why lunch can affect your next snack
• Social eating and weight loss: how friends shape meal size
• Home food environment connects shopping, meals and weight
• Mindful eating and weight loss: what changes beyond the scale
• Implementation intentions can help diet and exercise plans work
• Financial incentives for weight loss and what happens afterward
• Social support and weight loss depend on what a partner does
• Rapid versus gradual weight loss and what happens afterward
• Food diary underreporting and why calorie totals can mislead
• Sedentary breaks change blood sugar without proving weight loss
• Active commuting may mean smaller weight gains over the years
• Swimming and weight loss: what changes beyond the scales?
• Running vs walking changes the comparison for weight loss
• Resistance training volume and frequency for muscle and fat
• Cold exposure activates brown fat but may not reduce weight
• Sauna and bathing for weight loss: what the scale cannot tell
• Weight loss and bone density depend on diet and exercise
• Weight loss and mood: why mental health follows its own path
• Shift work and weight gain beyond the late-night snack story
• Remote work changes meals, but weight gain is not inevitable
• Weekly weight fluctuations and what a Monday gain really means
• High-protein diet types and what their weight-loss results mean
• GLP-1 and exercise: preserving strength beyond weight loss
• Weight cycling and your health: what regaining weight means
• Appetite hormones and weight loss: leptin, ghrelin and hunger
• Calorie labels and the energy your body actually receives
• Late-night eating and weight gain: what meal timing studies show
• Active commuting and weight loss: what walking and cycling change
• Glycemic index and weight loss: what the diet trials show
• Mediterranean diet and weight loss: what the trials show
• DASH diet and weight loss: what changes beyond blood pressure
• Plant-based diets and weight loss: what the trials really show
• Nuts and weight loss: what high-fat snack research really shows
• Yogurt and weight loss: what long-term dairy studies reveal
• Caffeine and green tea for weight loss: what the numbers mean
• Weight loss supplements and the evidence on benefits and risks
• Probiotics and weight loss: how much do the studies show?
• Brown fat and cold exposure: why burning energy is not weight loss
• Passive heating and weight loss: what baths and saunas show
• Exercise and appetite: when hunger becomes compensatory eating
• Energy compensation and why exercise calories may not add up
• Fasted exercise and fat loss: what breakfast timing really changes
• Exercise timing for weight loss: morning or evening workouts?
• Mindful eating and weight loss: what changes beyond the scale
• Portion size and weight loss: what larger servings change
• Food environment and weight loss: what your kitchen can change
• Eating out and weight loss: what to count and how to choose
• Holiday weight gain and the small changes that last into spring
• Body fat measurement accuracy and what your home scale can tell you
• Menstrual cycle and weight changes: what the research shows
• Social support and weight loss: when family help becomes pressure
• Weight loss goal setting and the science of if-then plans

All Diet Science articles (English)