Interpreting RMR Tests in Sports Dietetics: When to Trust the Number, When to Ignore It

A sports dietitian's office at 8:14 in the morning looks the same everywhere. The athlete is sitting up on a treatment table. The metabolic cart is humming. The hood is over their head. Ten minutes later the printout slides out and the screen says 1,742 kcal/day. The athlete asks "is that good?" The dietitian writes the number in the chart. The conversation moves on.

This is the moment most RMR tests get interpreted incorrectly.

A resting metabolic rate measurement from indirect calorimetry is the closest thing a sports RD has to a ground-truth read on the largest single component of total daily energy expenditure. Done right, it produces an estimate that is more accurate than any predictive equation by a factor of 5–15%. Done wrong, it produces a number that is somehow less trustworthy than a Mifflin-St Jeor estimate would have been — because nobody questions a number that came off a metabolic cart.

This is the framework I use to read RMR reports in clinical practice. It will not replace your training in indirect calorimetry, but it will sharpen how you turn the report into a defensible chart entry.

What the Test Actually Measures

Indirect calorimetry estimates oxygen consumption (VO2) and carbon dioxide production (VCO2) at rest, then applies the Weir equation to compute kcal/day. The math is mechanically sound. The error is not in the math. The error is in the assumption that the athlete in the chair was, in fact, at metabolic rest.

A 10-minute steady-state RMR test on a properly calibrated machine, with a properly prepared athlete, returns a value with within-day reproducibility around ±3–5%. A test on the same machine the same week with an athlete who walked from the parking lot, drank coffee that morning, slept poorly, or trained the prior afternoon can return a value 200–400 kcal off the same athlete's true RMR. The machine reports both values to two decimal places and the report looks identical.

The implication: the printout is one observation under one set of conditions. The clinical question is not "what does the report say?" — it is "is this report trustworthy enough to drive the plan?"

Pre-Test Conditions That Invalidate a Result

Six conditions reliably distort indirect calorimetry. If any of them are present, the result is not a clean RMR — it is a rumor.

Caffeine in the prior 4 hours. Caffeine elevates RMR by 5–10% acutely. An athlete who had a morning cold brew and tested at 9am is reading 100–150 kcal too high. Document caffeine timing on the intake form.

Food in the prior 5 hours. The thermic effect of food contaminates the measurement for 4–6 hours after a meal, depending on size and macronutrient composition. Same direction as caffeine — overestimates RMR.

Exercise in the prior 14 hours. Excess post-exercise oxygen consumption (EPOC) is real, measurable, and lasts longer than most clinicians assume. A hard session the prior evening will inflate a 7am RMR by 3–8%.

Cold or warm room. Thermal stress increases metabolic rate via thermogenesis. A test in a 64°F room reads higher than the same athlete in a 72°F room. ACSM-aligned protocols specify 20–25°C (68–77°F), neutral relative humidity.

Recent illness or inflammation. Subclinical viral illness, upper respiratory infection, recent vaccination, or active injury inflammation all raise RMR. This is biologically real, not artifact — but it is not the athlete's baseline RMR.

Active dieting. Adaptive thermogenesis depresses RMR by 5–15% during sustained energy deficit. The number is real for that period of time, but it is not the athlete's "set point" — it is the depressed RMR of an underfed athlete. The clinical action that follows from a 1,420 kcal RMR reading depends entirely on whether the athlete is in maintenance, deficit, or rebound.

A test that violates one of these conditions is still useful as a one-off snapshot. It is not useful as a baseline RMR for prescription. The chart entry should reflect that distinction.

Reading the Report

A standard RMR printout from a modern metabolic cart (TrueOne, ParvoMedics, COSMED Q-NRG, etc.) contains six fields you should be reading, not just the headline kcal/day:

1. RQ (Respiratory Quotient). The ratio of VCO2 to VO2. Resting RQ in a fasted, healthy individual sits between 0.78 and 0.85. RQ above 0.90 means the athlete is in the postprandial window or has done recent intense work. RQ below 0.75 in a non-fasted adult is unusual and may indicate measurement error, hyperventilation, or a leaky hood seal. RQ below 0.70 is mechanistically improbable and should trigger a re-test.

2. Steady-state achievement. The cart's software flags whether the test reached steady state — typically defined as a 5-minute window where minute-to-minute VO2 varied by less than ±10%. If the printout flags "did not reach steady state," the kcal value is an extrapolation, not a measurement.

3. Coefficient of variation across the steady-state window. A CV under 5% is solid. Above 10% indicates the athlete was restless, talking, or breathing irregularly — read the kcal number with skepticism.

4. Test duration. Modern protocols require 10–20 minutes of total recording with at least 5 minutes of valid steady-state data. A 7-minute test that was cut short is not equivalent to a clean 15-minute test.

5. Predicted RMR (per equation). Most carts auto-compute Mifflin-St Jeor and Harris-Benedict for comparison. The measured/predicted ratio is the single most useful field on the printout. An athlete reading 88% of predicted is in adaptive territory. An athlete reading 115% of predicted is an outlier — investigate before trusting it.

6. Calibration date and gas reference timestamp. A cart that has not been calibrated in two weeks, or that reads significantly off the most recent reference gas, is producing numbers you cannot trust. Most centers will not surface this on the patient-facing printout — ask the technician.

If the report does not surface these fields, the report is incomplete. Request the raw data export or order a different lab.

Reconciling Measured RMR with Predictive Equations

A measured RMR is one estimate. A Mifflin-St Jeor or Cunningham equation estimate from the same morning is another. Neither is gospel. The job is to look at both, identify when they disagree, and weight the difference.

The predictive equations have known biases:

• Mifflin-St Jeor under-predicts in athletes with higher fat-free mass — typically by 5–10% in muscular populations because it uses total body mass. A measured RMR above predicted in a lean, muscular athlete is expected and usually trustworthy.
• Cunningham uses fat-free mass directly and tracks better in athlete populations, but requires a recent FFM estimate from DEXA, BIA, or skinfolds. If the FFM estimate is stale or wrong, Cunningham will be wrong by the same proportion.
• Harris-Benedict is older, less accurate in modern populations, and consistently over-predicts in younger adults. Use it only for back-compatibility with older charts.

For a typical sports RD chart, Cunningham is the better predictive comparison if FFM is recent and reliable. The interpretation rule:

• Measured/Cunningham within 5%: high-confidence baseline. Use the measured value.
• Measured/Cunningham 5–15% lower: possible adaptive depression — investigate intake history, weight trend, training load.
• Measured/Cunningham 15%+ lower: strong signal of adaptive thermogenesis or test-condition violation. Re-test under controlled conditions before acting.
• Measured/Cunningham 5–15% higher: usually fine in athletes; consider whether test conditions were violated upward (caffeine, EPOC, cold room).
• Measured/Cunningham 15%+ higher: investigate. Likely test-condition contamination or rare hypermetabolic state. Re-test before using the number.

This is the same Bayesian framing that applies to any noisy clinical instrument: use the prior (predictive equation, FFM, training and intake history) to gate when the new measurement should update the plan.

A Worked Example

Male collegiate distance runner, 21 years old, 65 kg, 11% body fat by DXA (FFM 57.9 kg). Cunningham predicts ~1,820 kcal/day RMR. Mifflin-St Jeor predicts ~1,640 kcal/day.

He arrives for an indirect calorimetry test at the campus performance lab. Fasted 11 hours, no caffeine, no training in 18 hours, room temperature 70°F. Test runs 15 minutes, steady state achieved in minutes 6–11, CV 3.4%, RQ 0.81. Reported RMR: 1,540 kcal/day.

Measured/Cunningham = 0.85 — 15% under predicted. The test conditions are clean. The number is a real reflection of his current metabolic state, not an artifact.

The clinical interpretation: 1,540 kcal/day is the depressed RMR of an underfed endurance athlete. The chart entry should not document "RMR 1,540 kcal/day" as a baseline. It should document "measured RMR 1,540 kcal/day at 85% of Cunningham predicted; consistent with adaptive thermogenesis from sustained energy deficit; not appropriate as baseline RMR for prescription. Re-measure after 6 weeks of restored intake to establish maintenance baseline."

That is the difference between charting a number and charting a clinical interpretation.

The Documentation Pattern

For every RMR test, the Objective section of the SOAP should capture eight fields:

• Date and time of test
• Cart make/model and last calibration date
• Pre-test conditions: hours fasted, caffeine timing, hours since last training, room temperature
• Test parameters: total duration, steady-state window, CV, RQ
• Measured RMR (kcal/day)
• Predicted RMR (Cunningham preferred; Mifflin-St Jeor secondary) with measured/predicted ratio
• Interpretation: clean baseline / suspect / invalidated by test conditions / depressed (adaptive) / elevated (post-exercise or caffeine)
• Plan impact: use as TDEE input / discard / re-test in N weeks

The "Interpretation" line is where clinical judgment goes on the page. A reader picking up the chart in 12 weeks should know whether the RMR number is a tight estimate the team should fuel against, or a noisy reading flagged for re-measurement.

Common Mistakes

Treating measured RMR as the athlete's metabolic ceiling. A 1,420 kcal RMR in a depleted athlete is not the number to multiply by an activity factor and prescribe against. It is the number to flag.

Skipping the predictive equation comparison. The single most useful field on the printout is the measured/predicted ratio, and most chart entries leave it out.

Ignoring RQ. An RQ of 0.92 on a fasted RMR test is a tell — either the athlete was not actually fasted or the steady state was never reached. The kcal number that follows is suspect.

Using one-off measurements as longitudinal data. A single RMR test under unknown calibration conditions does not compare cleanly to one taken 6 months later on a different machine. For trend tracking, anchor to the same lab, same machine, same protocol.

Not documenting the conditions. A naked "RMR: 1,742 kcal/day" entry is not a clinical record. It is a fact-shaped object that future-you will not be able to interpret.

The Bottom Line

A measured RMR is the most accurate single estimate of an athlete's resting metabolism — when the test conditions are clean, the cart is calibrated, and the report is read with the right pre-test context. None of those are guaranteed by the existence of a printout.

Read the conditions. Read the RQ. Read the steady-state flag. Compare to Cunningham. Interpret the gap. Document the interpretation, not just the number.

If your RMR entries currently look like "RMR 1,742 kcal — Mifflin agrees" with no further context, the entries are not clinical documentation — they are screenshots in prose. Tighten the framework. The next reader of the chart, including future-you, will be able to act on the number rather than guess at it.

[Calsanova's Dietitian plan](/signup?role=dietitian) integrates RMR test results, food log data, and wearable expenditure into a single energy-balance dashboard, with built-in measured-vs-predicted reconciliation and SOAP templates that bake the conditions checklist into every Objective entry. Start your 30-day free trial and stop charting numbers without context.