Metabolic Suppression: Measuring for the RMRratio

Updated October 2025: With a newer paper published by Heikura et al. (2025) that measured eleven predictive resting metabolic equations and how their RMRratio results compared against the International Olympic Committee's Relative Energy Deficiency in Sport Clinical Assessment Tool Version 2 (CAT2), I've updated this post to include their results.

As a pre-reading to the below article, I would recommend familiarizing yourself with the CAT2 and its multiple diagnostic categories: green, yellow, orange, and red. Here's a quick snapshot:

IOC REDs CAT2 diagnostic categories (15).

---

I’ve begun teaching athletes more about their resting metabolic rate (RMR), or the calories burned to keep the athlete alive if they were to lay in bed all day doing nothing.

Even when on the couch binge-watching Netflix, muscles of the heart and lungs work 24/7 to keep blood and oxygen pumping throughout the body. New red blood cells are being made. Damaged tissues and muscle cells are being replaced. Your RMR should be optimized and prioritized, whether it’s an off day or training day.

But what happens when in a state of chronic low energy? The body cannot continue to operate at 100% and is forced to pull back. This is metabolic suppression and it can be measured and then compared to one’s own data set.

In this article, I’m going to review:

• The RMRratio.

• How to measure and predict a person’s RMR.

• How to apply and use the RMRratio in practice.

The RMRratio: What Is It? Why Calculate It?

Like any ratio, the RMRratio aims to evaluate how similar (or dissimilar) measured and predicted RMR are to one another and to act as a surrogate for energy deficiency and/or a suppressed metabolism. The equation is: (1)

RMRratio = measured RMR kcal (mRMR) / predicted RMR kcal (pRMR)

In theory, a normal RMRratio would range between 0.9-1.1. Cunningham (1980) wrote that the confidence limit for the slope of the Cunningham equation (see the below section on predicted RMR) is ±1.9, meaning a 10% prediction error for one’s RMR is expected—hence the 0.9-1.1 or 90-110% range. (2) However, if the RMRratio is below 0.9 (or is normal and continues to trend lower over time), metabolic suppression is assumed. (3) An important caveat is that the 0.9 is considered a surrogate marker of low energy availability--it is not a diagnostic marker on its own. (14)

For instance, let’s say an athlete’s predicted-by-an-equation's RMR is 1,800 kcal. When measured with an indirect calorimeter, the athlete’s RMR is returned as 1,850 kcal:

RMRratio = 1.03 = (1,850 kcal mRMR) / (1,800 kcal pRMR)

This is within the 0.9-1.1 range and would be interpreted as normal.

Three studies evaluating “clinical models of starvation, such as anorexia nervosa, during periods of low body weight and prior to feeding” found a RMRratio in the 0.6-0.8 range. (4) Of those studies, two included females with and without anorexia nervosa (5,6), whereas the Melchior et al. (1989) study I could only find the abstract for (gender was unspecified for the 11 adults with anorexia nervosa). (7) This means the more extreme and chronic the deficit becomes, the more suppression occurs and the lower the ratio becomes.

How-to: Predicted or Estimated RMR (pRMR)

The pRMR is what a practitioner would calculate based on a handful of metrics. Depending on the predictive equation used, inputs include lean body mass, fat-free mass, age, sex, height, and/or weight.

Staal et al. (2018) evaluated three predictive RMR equations against a direct RMR measurement in a study of adult professional ballerinas, including both males and females. (3) The equations used included the Cunningham, Harris-Benedict, and a DXA-derived equation from Koehler et al. (2016). (8) The researchers concluded that the Cunningham equation should be used to calculate RMR in athletes when calculating RMRratio. (3) This equation, though, was created using non-active individuals. Since the Staal et al. (2018) study, others have found that "no single equation consistently outperforms another" when estimating RMR and that "RMR predictive equations should be chosen, or developed, based on a close match between source and target population characteristics." (14)

For reference, the Cunningham equation (C80) is: (2)

BMR (calories/day) = 500 + 22(LBM)

BMR is the basal metabolic rate and LBM is lean body mass. The latter is what Cunningham (1980) wrote as being the “single predictor of BMR” and accounts for 70% of BMR variability. (2)

Heikura et al. (2025) further calculated the RMRratio using 11 RMR predictive equations, including Cunningham. Here's what they found:

• When using the Cunningham equation to evaluate REDs, the "RMRratio was only lower in extreme cases of REDs (only red vs. green light athletes)--demonstrating that RMR may have limited utility in REDs detection outside of extreme cases."

• The van Hooren and Jagim equations had the highest sensitivities in RMRratio and diagnosing the state of REDs per the IOC CAT2 (i.e., green, yellow, orange, or red category). Both van Hooren and Jagim were 90% with Cunningham being 70% sensitive.*

• Both the van Hooren and Jagim equations were designed in athletic populations:

• Across the eleven equations tested, "prevalence of low RMR ranged from 1% to 68%." (14)

The van Hooren equation was developed in professional cyclists (21 males and four females): (14,17)

([0.963 - (0.186 x sex) + (0.106 x body mass in kilograms)] x 1,000) / 4.184

The Jagim equation was developed in NCAA Division III athletes (68 male athletes: football, distance running, and baseball; 48 female athletes: sprints, distance running, throws, swimming, soccer, and tennis): (14,18)

Females: (21.1 x body mass in kilograms) + 288.6

Males: (19.46 x body mass in kilograms) + 775.33

*Sensitivity = (True positive cases) / (True positive cases + False negative cases) = Those who truly have X diagnosis test positive for the diagnosis. (16)

How-to: Measured RMR

Before the how-to, let’s be mindful when considering an athlete’s background when measuring RMR, as it could affect the outcome and interpretation of the RMRratio.

The Biggest Loser studies showed that the severe, chronic calorie restriction undertaken by contestants (a ~65% calorie restriction compared to baseline) suppressed their metabolism even after the intentional calorie deficit stopped, and often when weight gain occurred. The study evaluated contestants’ metabolic rate using both a metabolic cart and doubly labeled water (read more about DLW here) multiple times when contestants were cast for the show, six and 13 weeks into the show, at week 30 for the show’s finale, and then six years later. Ultimately, compared to control subjects of similar body weights without having experienced a severe, chronic calorie restriction (i.e., non-Biggest Loser contestants), contestants still had measured lower-than-expected metabolisms that could not be explained otherwise. This was termed persistent metabolic adaptation by the researchers and “metabolic damage” or “breaking one’s metabolism” by media outlets. (9,10)

Even though The Biggest Loser doesn't feel like reality, it is. For instance, others who intentionally reduce their calorie intake over long periods of times include those with calorie restrictive eating disorders. Sterling et al. (2009) conducted a retrospective chart review on adolescent females aged 14-19 years with normal body weights (90-120% of their ideal bodyweight), but with a history of anorexia nervosa (AN), bulimia nervosa (BN), or an eating disorder not otherwise specified (EDNOS). Sixty of the females were amenorrheic and 121 were menstruating normally. Using the Harris-Benedict equation when calculating the RMRratio, the amenorrheic females averaged 0.792 ± 0.1 and those menstruating normally averaged 0.852 ± 0.97. Using the 0.9-1.1 RMRratio range, both groups showed averages consistent with metabolic adaptation. One of the authors’ limitations was that neither the duration of the eating disorder nor time elapsed since weight restoration were evaluated. (11) I would add the degree and timeline of calorie restriction would have been interesting to know, and that AN, BN, and EDNOS should all affect metabolism in different ways. Keep in mind that the Sterling et al. (2009) study included weight restored patients whereas The Biggest Loser contestants had regained on average two-thirds of what they had originally lost. (9-11)

Measuring one’s RMR is done via indirect calorimetry (IC), which measures pulmonary gas exchange: how much oxygen is inhaled and how much carbon dioxide is produced. IC is commonly used in the clinical setting where predictive equations can greatly under- or overestimate a medical condition or emergency, and therefore affect the amount of nutrition necessary to support a patient’s healing, recovery, and often survival. For instance, burns, sepsis, and inflammation all increase RMR whereas paralysis, coma, and sarcopenia reduce RMR. (12) Weir’s equation is used to estimate RMR:

Weir’s equation (kcal/day) = [(VO2 x 3.941) + (VCO2 x 1.11)] x 1440

In the non-clinical world, standardizing an IC measurement is recommended. Vigorous exercise, eating, drinking (all fluid, but especially coffee and alcohol), and stimulants (e.g., pre-workout powders) are restricted for a set number of hours before the IC measurement. Participants are encouraged to be in a low-stress state and arrive early in the morning by car to avoid excessive movement (e.g., avoiding transportation via bike or foot). IC measurements are taken in a calm environment where the patient is resting 10-15 minutes beforehand. A measurement is successful upon a steady-state period of gas exchange, indicating a 3-to-5-minute interval of VO2 and VCO2 varying by less than 10%. (13)

When purchasing an IC, ensure you’re reviewing validation data to ensure your machine’s output is accurate (never assume!). Once you have a machine, ensure you calibrate it before each use.

However, one of the issues with RMR measurement relates to methodological limitations. For instance:

• What machine was used: For instance, a ventilated hood versus a handheld IC.

• How long the test lasted: Like any data collection, the more data points you collect, the more confident you become in what's the average value, possibly even throwing out extreme values. This is identical for time spent collecting one's inhalation and exhalation data. When using a ventilated hood, the first few minutes of data collection is thrown out. Then, the next ~15 minutes accounts for usable data, with five of those minutes being chosen for data showing the least amount of variation.

• How LBM was measured: For instance, skin fold, BOD POD, BIA, or DXA.

• How the participant was prepared: RMR is collected early in the morning in a fasted state. If using a machine to collect an athlete's RMR after a workout in the middle of the day, the data isn't going to be sound. (14)

Key Takeaways and Applying the RMRratio

Like any data, RMRratio is only one piece of the puzzle when evaluating someone’s health. It’s not the ultimate data point directing one’s care. Heikura et al. (2025) wrote that evidence "suggests that the utility of RMR as a one-off indicator of LEA, Triad, or REDs is likely poor." (14)

For instance, and in practice when using the Cunningham equation only, I’ve measured athletes with values on the low-normal end of the 0.9-1.1 range with only some falling below that value. Taking an aerial perspective of these athletes, I’ve noticed cycles of intentional weight manipulation, extreme calorie restriction, eating disorders, and/or secondary amenorrhea. Often, athletes with these signs and symptoms end up in the normal RMRratio range when using the Cunningham equation, yet as a practitioner working within the medical team to provide treatment, we knew the RMRratio wasn't validating our other data points--but that's OK.

Given The Biggest Loser and eating disorder data showed long-term suppressive effects on one’s RMR—even when considered healthy or recovered—a measured RMR may come back lower than expected based on someone’s history with chronic and severe calorie restriction. However, in the presence of restriction and a low RMRratio, it’s reasonable to work to improve one’s nutritional and caloric intake, measuring RMR again to evaluate any improvements in the RMRratio. For some athletes, optimizing their metabolism with nutrition may not increase their ratio of 0.88 up to a 1.0, but some improvements can still occur. Have a realistic approach to what your goal as the practitioner is and continue to collect data as necessary.

If you lack accurate body composition data or an IC, you cannot calculate the RMRratio. However, there will be other data you and the medical team can collect if underfueling and/or metabolic suppression are of concern. For instance, diet logs, menstrual or sex drive perturbations, and labs (e.g., T3, estradiol in women not using hormonal contraceptives), among other findings. These values work together to tell you a story about the athlete's health beyond a RMRratio. Whether it's integrating the RMRratio or not, one data point does not diagnose an athlete as REDs. Always work closely with your medical team to share pertinent data.

Given humans fall on continuums (0.9-1.1) and some are outliers (<0.9 and >1.1), focusing on monitoring trends and not absolute values is helpful when communicating findings to athletes, and how this piece of data fits into their overall nutrition program and planning. If the athlete presents as an appropriate candidate, measuring their RMR and RMRratio is another available tool the practitioner has to educate an athlete about their health (and sometimes how destructive their behaviors are). And in the presence of other concerning signs and symptoms, a "normal" RMRratio does not mean the athlete is OK.

References

(1) Sterringer, T., & Larson-Meyer, D.E. (2022). RMR ratio as a surrogate marker for low energy availability. Curr Nutr Rep,11(2):263-72. https://pubmed.ncbi.nlm.nih.gov/35080753/

(2) Cunningham, J.J. (1980). A reanalysis of the factors influencing basal metabolic rate in normal adults. Am J Clin Nutr,33(11):2372-4. https://pubmed.ncbi.nlm.nih.gov/7435418/  

(3) Staal, S., Sjödin, A., Fahrenholtz, I., Bonnesen, K., & Melin, A.K. (2018). Low RMRratio as a surrogate marker for energy deficiency, the choice of predictive equation vital for correctly identifying male and female ballet dancers at risk. Int J Sport Nutr Exerc Metabl,28(4):412-418. https://pubmed.ncbi.nlm.nih.gov/29405782/

(4) De Souza, M.J., Hontscharuk, R., Olmsted, M., Kerr, G., & Williams, N.I. (2007). Drive for thinness score is a proxy indicator of energy deficiency in exercising women. Appetite,48(3):359-67. https://pubmed.ncbi.nlm.nih.gov/17184880/

(5) Marra, M., Polito, A., De Filippo, E., Cuzzolaro, M., Ciarapica, D., … & Scalfi, L. (2002). Are the general equations to predict BMR applicable to patients with anorexia nervosa? Eat Weight Disord,7(1):53-9. https://pubmed.ncbi.nlm.nih.gov/11933912/

(6) Polito, A., Fabbri, A., Ferro-Luzzi, A., Cuzzolaro, M., Censi, L., … & Giannini, D. (2000). Basal metabolic rate in anorexia nervosa: relation to body composition and leptin concentrations. Am J Clin Nutr,71(6):1495-502. https://pubmed.ncbi.nlm.nih.gov/10837290/

(7) Melchior, J.C., Rigaud, D., Rozen, R., Malon, D., & Apfelbaum, M. (1989). Energy expenditure economy induced by decrease in lean body mass in anorexia nervosa. Eur J Clin Nutr,43(11):793-9. https://pubmed.ncbi.nlm.nih.gov/2627927/

(8) Koehler, K., Williams, N.I., Mallinson, R.J., Southmayd, E.A., Allaway, H.C.M., & De Souza, M.J. (2016). Low resting metabolic rate in exercise-associated amenorrhea is not due to a reduced proportion of highly active metabolic tissue compartments. Am J Physiol Endocrinol Metab,311(2):E480-7. https://pubmed.ncbi.nlm.nih.gov/27382033/

(9) Hall, K. (2022). Energy compensation and metabolic adaptation: “The Biggest Loser” study reinterpreted. Obesity,30(1):11-13. https://pubmed.ncbi.nlm.nih.gov/34816627/

(10) Fothergill, E., Guo, J., Howard, L., Kerns, J.C., Knuth, N.D., … & Hall, K.D. (2016). Persistent metabolic adaptation 6 years after “The Biggest Loser” competition. Obesity,24(8):1612-9. https://pubmed.ncbi.nlm.nih.gov/27136388/

(11) Sterling, W.M., Golden, N.H., Jacobson, M.S., Ornstein, R.M., & Hertz, S.M. (2009). Metabolic assessment of menstruating and nonmenstruating normal weight adolescents. Int J Eat Disord,42(7):658-63. https://pubmed.ncbi.nlm.nih.gov/19247996/

(12) Delsoglio, M., Achamrah, N., Berger, M.M., & Pichard, C. (2019). Indirect calorimetry in clinical practice. J Clin Med,8(9):1387. https://pmc.ncbi.nlm.nih.gov/articles/PMC6780066/

(13) Das Gupta, R., Ramachandran, R., Venkatesan, P., Anoop, S., Joseph, M., & Thomas, N. (2017). Indirect calorimetry: from bench to bedside. Indian J Endocrinol Metab,21(4):594-9 https://pmc.ncbi.nlm.nih.gov/articles/PMC5477450/

(14) Heikura, I.A., Tsai, M.-C., Sesbreno, E., McCluskey, W.T.P., Johnson, L., … & Stellingwerff, T. (2025). Current resting metabolic rate prediction equations lack sensitivity and specificity to indicate Relative Energy Deficiency in Sport: a large cohort study in elite athletes. Int J Sport Nutr Exerc Metab,35(4):324-36. https://pubmed.ncbi.nlm.nih.gov/40262739/

(15) Stellingwerff, T., Mountjoy, M., McCluskey, W.T.P., Ackerman, K.E., Verhagen, E., & Heikura, I.A. (2023). Review of the scientific rationale, development and validation of the International Olympic Committee Relative Energy Deficiency in Sport Clinical Assessment Tool: V.2 (IOC REDs CAT2)—by a subgroup of the IOC consensus on REDs. B J Sports Med,57(17):1109-21. https://bjsm.bmj.com/content/57/17/1109

(16) Sensitivity and specificity. (n.d.). Osmosis from Elsevier. Retrieved October 21, 2025, from, https://www.osmosis.org/learn/Sensitivity_and_specificity

(17) van Hooren, B., Cox, M., Rietjens, G., & Plasqui, G. (2023). Determination of energy expenditure in professional cyclists using power data: validation against doubly labeled water. Scand J Med Sci Sports,33(4):407-19. https://pubmed.ncbi.nlm.nih.gov/36404133/

(18) Jagim, A.R., Camic, C.L., Askow, A., Luedke, J., Erickson, J., ... & Oliver, J.M. (2019). Sex differences in resting metabolic rate among athletes. J Strength Cond Res,33(11):3008-14. https://journals.lww.com/nsca-jscr/fulltext/2019/11000/sex_differences_in_resting_metabolic_rate_among.16.aspx