Jul 27, 2026
The Uncomfortable Truth About Metabolic Testing
For someone who's built a career around metabolic testing, this probably isn’t something people expect me to write.
I’ve spent thousands of hours looking at metabolic data, and I genuinely believe it’s one of the most valuable tools we have for understanding human physiology. Done well, it can transform how athletes fuel and train, and how coaches make decisions. It’s the foundation of everything we do at Metabolix.
But I’m getting increasingly pissed off with where the industry is heading.
Metabolic testing has become more accessible than ever. Portable analyzers have entered the market, software has become more sophisticated, and testing is now being offered in gyms, wellness clinics, and mobile facilities that would never have considered buying a traditional metabolic cart.
On the surface, that sounds like progress. In many ways, it is.
The problem is that accessibility has outpaced understanding.
People are putting on a mask, exercising on a treadmill or bike, and leaving with a polished report containing their VO₂ Max, thresholds, training zones, fuel utilization, fitness classifications and personalized recommendations.
From the customer’s perspective, it looks like a metabolic test.
But was it the metabolic test they thought they were paying for?
That’s the question more people need to ask.
The issue isn’t that every new or portable analyzer is inaccurate. That would be an easy argument to make, but it wouldn’t be an honest one. Several portable systems have legitimate evidence supporting particular measurements under particular conditions.
The problem is the assumption that every device producing a VO₂ Max number is providing an equivalent test.
It isn’t.
A mask, a treadmill and a colorful report can make two assessments look almost identical from the outside. Underneath, the equipment may be measuring different variables, using different sensors, relying on different assumptions and producing outputs supported by very different levels of evidence.
The person paying for the test may have no idea.
Whenever someone shows me a metabolic testing report, I ask one question before looking at anything else:
What was actually measured?
It’s amazing how often that catches people off guard.
Most people understandably assume a metabolic analyzer measures everything shown in the report. It doesn’t. Every system begins with a relatively small amount of raw information.
Depending on the analyzer, it may directly measure airflow, oxygen concentration and carbon dioxide concentration. Heart rate, speed and workload usually come from connected equipment.
That’s the raw data. Everything else is built from there.
Some values are calculated from those measurements. Some are estimated. Some require interpretation by the person conducting the test. Others are generated by software using algorithms, reference data and assumptions the customer may never see.
And some are little more than educated guesses wrapped in attractive graphics.
The problem isn’t that metabolic testing involves calculations, estimates or interpretation. Physiology requires all three. The problem is that reports often present them with exactly the same confidence as the measurements underneath them.
Those aren’t the same thing. Not even close.
Imagine I hand you a bathroom scale. The scale tells you that you weigh 170 pounds.
That’s a measurement.
Now imagine I tell you that because you weigh 170 pounds, you’re carrying precisely 31.2 pounds of body fat, your metabolic age is 31, your mitochondria are underperforming and your cortisol levels are elevated.
Hopefully, you’d ask how I worked that out. What else did I measure? What assumptions did I make? What evidence supports those conclusions? How accurate are the estimates?
Strangely, we don’t always ask the same questions when it comes to metabolic testing. We accept every number on the report as though it was directly measured by the analyzer.
It wasn’t.
Modern software is exceptionally good at making uncertainty disappear. If a report tells you that you were obtaining 72.4% of your energy from fat at 136 beats per minute, it looks authoritative. It feels scientific.
But precision and accuracy are not the same thing.
Understanding the difference between what was measured, what was calculated and what was inferred may be the single most important thing you can learn before paying for a metabolic assessment.
Let’s take VO2 Master.
The current product is sold as the VO2 Master Analyzer and marketed as providing “lab-grade” VO₂ analysis, resting metabolic rate testing, ventilatory thresholds and automated reports. VO2 Master’s own FAQ confirms that its current analyzer does not include carbon dioxide measurement, with a CO₂ version still under development. Its published output list includes VO₂ Max, ventilation, breathing metrics, VT1, VT2, training zones and resting metabolic rate outputs.
That doesn’t make it useless.
A 2020 study compared two versions of the VO2 Master Pro with a Parvo Medics TrueOne 2400 during stationary cycling. It found acceptable validity and test-retest reliability for oxygen consumption and ventilation across most of the intensities tested, although the differences were not zero and reliability was somewhat poorer than the comparison system.
That is legitimate evidence for VO₂ and ventilation during the cycling protocols that were studied.
It is not evidence that the device measures carbon dioxide. It does not validate RER, because RER requires both VCO₂ and VO₂. It does not validate exact fat and carbohydrate oxidation, because the established equations used to calculate substrate oxidation require measurements of both gases.
VO2 Master itself explains that its resting energy expenditure calculation assumes an average respiratory quotient of 0.85 because VCO₂ is not measured. The company acknowledges that an individual’s actual value can vary and that this assumption introduces error. It also states that the principal added value of VCO₂ for a resting report is determining fat-versus-carbohydrate utilization.
That is an estimate. It may still be useful, but it is not the same as measuring both respiratory gases.
The importance of those distinctions became even clearer in July 2026, when a study compared a VO2 Master Pro Model 1.6.1 with a Parvo Medics TrueOne 2400 in 14 endurance-trained runners. The VO2 Master Pro reported an average relative VO₂ Max of 65.4 ml/kg/min, compared with 53.5 ml/kg/min from the laboratory system. It overestimated VO₂ Max in every participant, with a mean bias of 11.9 ml/kg/min and poor agreement between the systems.
That study is small, and it isn’t a universal verdict on every device in every situation. It also doesn’t erase the more favorable cycling or simulator research.
It demonstrates something more important.
Validation is specific.
A device can perform acceptably during controlled cycling and perform very differently during maximal running. Results from one model, software version, exercise mode, protocol and population cannot simply be stretched across every use of the product.
VO2 Master’s current site nevertheless says it delivers lab-grade analysis without compromising data quality. Its own support information also acknowledges that the analyzer may sometimes fail to capture sufficient data for automatic ventilatory threshold identification.
Again, that doesn’t mean the device has no value. If someone wants a highly portable measurement of oxygen consumption and ventilation during a suitable activity, it may be an appropriate tool.
But if they believe they’re receiving complete respiratory gas analysis, direct RER measurement, individualized substrate oxidation or an assessment equivalent to a full dual-gas metabolic cart, that belief needs correcting.
The test may still be useful.
It just may not be the test they think they’re getting.
PNOĒ presents a different problem.
Unlike the current VO2 Master Analyzer, PNOĒ measures both oxygen and carbon dioxide. It is capable of producing VO₂, VCO₂, ventilation and respiratory quotient from breath-by-breath gas analysis.
It also has legitimate published validation evidence.
A 2019 study tested PNOĒ against a COSMED Quark CPET in 22 recreationally active adults during a four-stage cycling protocol. The researchers reported close agreement for VO₂, VCO₂, ventilation and respiratory quotient and concluded that PNOĒ could accurately determine those respiratory variables across the intensities studied in healthy people under controlled laboratory conditions. One author disclosed a previous scientific-advisor relationship with the company that developed the device.
A 2025 study in 21 healthy young adults compared PNOĒ with a COSMED K5 during treadmill walking. It assessed VO₂, VCO₂, RER, metabolic equivalents, tidal volume and energy expenditure. The correlations between systems were moderate, with generally favorable agreement and some variability between outcomes and stages.
That is meaningful evidence. PNOĒ has a substantially stronger measurement case than an oxygen-only analyzer when the goal is full respiratory gas analysis.
But here’s where things become slightly ridiculous.
PNOĒ currently says its test produces “clinical-grade data” on 23 metabolic biomarkers, including biological age and nervous-system activation. Its reports claim to identify whether the lungs, heart, cells or nervous system are underperforming, while generating personalized training, nutrition, calorie and macronutrient recommendations. Its marketing also describes fat-versus-carbohydrate use at each exercise intensity and promotes the test as assessing biological age, cardiometabolic risk and overall physiological function.
Those are much larger claims than accurately measuring VO₂, VCO₂, ventilation and respiratory quotient.
The validation studies do not establish that a breath test accurately determines every biological-age score, nervous-system classification, cellular-performance conclusion or health-risk inference generated by the current platform. They also do not validate every automated training, nutrition, recovery or treatment recommendation placed downstream of those measurements.
Validating a gas analyzer does not validate an entire software ecosystem.
PNOĒ’s own FAQ goes further, saying its data can be used to recommend IV therapy, red-light treatment, hormone-optimization programs and other high-value services. It describes metabolic testing as a “diagnostic funnel” that can justify those offerings and increase their uptake.
That should make people pause.
A valid VO₂ measurement doesn’t prove someone needs IV therapy. A measured RER doesn’t validate hormone optimization. A software-generated biological age doesn’t establish that red light, supplements or an anti-aging protocol will improve the inferred problem.
Every step away from the original measurement introduces another assumption.
Eventually, you’re no longer looking at what the analyzer measured. You’re looking at what the software believes those measurements might mean, followed by what a business might be able to sell you because of them.
That can still produce useful hypotheses.
But it needs to be described honestly.
Unfortunately, the analyzer is not the only problem.
Even a genuinely well-validated metabolic cart can produce questionable data when it’s calibrated incorrectly, paired with the wrong protocol or operated by someone who doesn’t understand what they’re looking at.
A metabolic analyzer is not a vending machine. You don’t put an athlete in one end, press a button and receive unquestionable physiological truth from the other.
The mask has to fit. Leaks have to be identified. The equipment has to be calibrated. The protocol has to suit the person and the question being asked. The raw data have to be inspected, and the thresholds have to be interpreted within the context of the entire test.
That last part matters more than the industry would sometimes like to admit.
Almost anyone can buy one of these systems. That doesn’t mean they understand gas exchange, VO₂ kinetics, protocol design, ventilatory thresholds, exercise physiology or the limitations of the equipment.
PNOĒ, for example, describes its certification as a six-to-eight-hour, self-paced course and says certification is recommended rather than required. That may teach someone how to operate the platform and read its reports. It does not turn them into an exercise physiologist.
Pressing “Generate Report” does not make someone an exercise physiologist.
The easier the equipment becomes to buy and operate, the easier it becomes to confuse operating software with understanding the test.
When questionable equipment, an inappropriate protocol or an inexperienced operator produces poor data, the software may still attach confident conclusions to it. The operator may not know enough to recognize that anything is wrong.
The result still looks professional.
That’s the dangerous part.
Athletes use these results to assess their fitness. Coaches use them to prescribe training. People compare themselves with population percentiles and may leave believing they’re average, exceptional or in the top one percent.
If the VO₂ value is wrong, the percentile attached to it is wrong too.
Thresholds and training zones may be misplaced. Fuel-use estimates may be unreliable. Changes between tests may reflect equipment error, inconsistent protocols or normal measurement variation rather than genuine physiological adaptation.
Someone may be falsely reassured by an inflated score or unnecessarily discouraged by an underestimated one.
Bad data creates bad decisions. Once a questionable number is placed inside an impressive-looking report, it becomes remarkably convincing.
That isn’t better information.
It’s false confidence.
Before paying for a metabolic test, ask which exact analyzer will be used. Ask what it measures directly and whether it measures both oxygen and carbon dioxide. Ask whether that specific model has been independently tested for the exercise mode, intensity and population relevant to your assessment.
Ask how it’s calibrated. Ask what protocol will be used and why. Ask how thresholds are identified. Ask which results were measured, which were calculated and which were inferred by software.
Most importantly, ask what decisions the test can genuinely support.
A reputable facility should be able to answer those questions clearly and without becoming defensive.
At Metabolix, we shouldn’t be exempt from that standard. Using professional dual-gas mixing-chamber technology doesn’t magically make every number correct or every interpretation beyond question.
Our equipment, protocols, interpretation and communication should be subjected to exactly the same scrutiny.
If a result carries uncertainty, we should explain it. If our equipment cannot support a particular conclusion, we shouldn’t make it. If a protocol cannot answer the question being asked, we shouldn’t pretend that it can.
And if better evidence changes our understanding, we should be willing to change how we work.
That doesn’t weaken metabolic testing. It strengthens it.
Metabolic testing is too valuable to be diluted by equipment, operators and software that create more certainty than the underlying data can support.
People deserve to know what they’re actually being measured with. They deserve to know who is interpreting the data, how much confidence they should place in the result, and whether the conclusions being sold to them are genuinely supported.
Because putting a mask on someone and handing them a VO₂ Max number doesn’t make every metabolic test equivalent.
Sometimes the honest answer is, “We don’t know.”
That’s infinitely more useful than pretending a bad measurement is the truth.
References: