Every blood test report comes with a reference range that looks authoritative, a clear boundary between healthy and sick. But where do these numbers actually come from? And why do they fail so predictably for athletes?
The answer lies in statistics, not medicine. Reference ranges are built using methods developed in the 1960s for general populations. They were never designed for people who deliberately alter their physiology through training, diet, or pharmacology. Understanding the construction process reveals exactly why they break down for athletes, and what to do about it.
📐A reference range is not a medical diagnosis. It is a statistical observation: 95% of the people we tested fell between these two numbers. The other 5% were healthy too — they were just statistical outliers.
The Reference Range Construction Process
When a laboratory establishes a reference range for a new marker, they follow a standardized process. It looks scientific, and it is, but the assumptions baked into that process break down for athletic populations:
Recruit a Reference Population
The lab pulls its reference group from the local patient population, typically 120 to 500 people. These are folks who came in for routine screening, not athletes chasing performance gains. The sample is a mixed bag: elderly patients, sedentary types, people on various medications, and plenty with subclinical conditions that haven't been flagged yet.
The problem is immediate: if 90% of your reference population does not train, markers affected by training, creatinine, CK, and AST, will be calibrated to sedentary values. A muscular athlete automatically becomes an outlier.
Consider creatinine. A sedentary reference population typically shows creatinine between 0.7 and 1.1 mg/dL. A 200 lb bodybuilder with 15% body fat and 180 mg of dietary protein per day will often run 1.3 to 1.5 mg/dL. That is not kidney disease; it is muscle mass and meat intake. The standard range flags him as abnormal, and a licensed medical professional unfamiliar with training physiology may order unnecessary renal workup. The same logic applies to CK, which can be 500 to 1000 U/L after a heavy leg day, while the lab's upper limit sits at 200. These are expected deviations, not pathology.
Measure and Plot the Distribution
The lab measures the marker in every reference subject and plots the distribution. Most biological markers follow a bell curve, the normal distribution, but some are skewed. Creatinine, for example, skews right because a small number of people have unusually high values.
When distributions are skewed, labs will often apply a log transformation or fall back on non-parametric methods. But that doesn't change the core issue: the reference population dictates the shape of the distribution. If that population is made up of sedentary people, the "normal" curve has no room for athletic physiology.
Apply the 2.5th to 97.5th Percentile Rule
The standard methodology goes back to 1987, when the International Federation of Clinical Chemistry set the reference interval as the range between the 2.5th and 97.5th percentiles of the reference population. That captures the middle 95% of values, nothing more, nothing less.
By definition, 2.5% of healthy people fall below the lower limit and 2.5% fall above the upper limit. Those are false positives, healthy individuals flagged as abnormal purely because they sit at the edges of the distribution. For athletes, who are outliers on multiple markers, the probability of being flagged is far higher than 5%.
Consider a 28 year old powerlifter on 500 mg testosterone weekly. His panel shows HDL at 28 mg/dL (low), creatinine at 1.4 mg/dL (high), AST at 60 U/L (high), and hematocrit at 50% (high). A standard report flags all four. But each is a direct consequence of his protocol: AAS suppress HDL, muscle mass raises creatinine, training raises AST, and testosterone stimulates erythropoiesis. None of these indicate disease. The false positive rate compounds because these markers are not independent; they are all shifted by the same underlying physiology. Without athletic context, the licensed medical professional sees a patient with liver, kidney, and lipid problems, and may recommend stopping TRT or starting a statin.
Publish and Forget
Once established, reference ranges rarely change. Labs may update them every 5-10 years, but the methodology stays the same. The ranges on your 2026 blood work were likely validated using population data from the early 2010s, before the explosion of strength training popularity and before most physicians had ever heard of AAS use in non-competitive athletes.
Why 95% Is Not Enough
The 95% rule sounds conservative, given that it covers almost everyone. But probability compounds across markers. With 20 markers on a blood panel, each carrying a 5% chance of flagging a healthy person, the probability of at least one false positive becomes surprisingly high.
The Math
Run 20 independent markers, each with a 5% false positive rate, and the chance of at least one flag jumps to 1 - (0.95)^20 = 64%. Nearly two-thirds of healthy people will get an "abnormal" result purely by chance.
The 64% figure assumes independent markers. In practice, markers are correlated. If you train, your CK, AST, and ALT all rise together. If you use AAS, your HDL, SHBG, and LH all fall together. This correlation means the probability of at least one flag is even higher than 64%, because the shifts are not random. A 20 marker panel with 5 correlated markers that all move in the same direction will almost certainly produce multiple flags. The real question is not whether you will get a false positive, but how many. Knowing this prevents the panic that comes from seeing three red arrows on a printout.
The Athlete Multiplier
Athletes are not randomly distributed. They are systematically shifted on multiple markers. Your probability of a false positive is not 64%. It is higher, because you are an outlier by design, not by chance.
Take hematocrit. A sedentary reference population averages 42% for men. A trained athlete on TRT often runs 48 to 52%. That is a systematic shift, not a random outlier. The standard range upper limit is 50%, so a value of 51% flags as high. But for an athlete, that value is expected and often benign, provided it is stable. The same applies to eGFR calculated from creatinine. A muscular athlete with creatinine 1.4 will have an eGFR around 60, which is below the 90 threshold for chronic kidney disease. Yet his actual kidney function, measured by Cystatin C, is normal. The standard range misclassifies him as stage 3 CKD.
The Multiple Testing Problem
Systematic Failure Modes
Standard ranges fail athletes in predictable ways, and once you understand how they fail, you can spot the problem before it triggers panic or a misdiagnosis.
A common mistake is to treat every flagged value as a separate problem. A licensed medical professional might see elevated AST and ALT and order a liver ultrasound, while the real cause is a recent heavy squat session. Another mistake is to ignore trends. A stable creatinine of 1.4 over three years is different from a creatinine that climbs from 1.2 to 1.6 in six months. The first is expected physiology; the second demands investigation. Athletes should always compare current values to their own baseline, not to a population average. If a marker is outside the standard range but matches your historical values and you feel well, it is likely physiological. If it changes rapidly or is accompanied by symptoms, that is the real signal.
False Positives: Expected Physiology Flagged as Disease
False Negatives: Real Problems Hidden in the Noise
Range Inflation: Labs Using Different Ranges
A Different Methodology
GearCheck's athletic reference bands aren't simply wider versions of standard ranges. They're built using a fundamentally different approach, one that preserves sensitivity to real pathology while eliminating false positives from expected physiology:
Reference Bands Built for This Population
GearCheck doesn't use the general population for its reference ranges. Each marker's band is defined separately for a natural context and an on-cycle context, and separately again by sex, rather than applying one general-population interval to everyone. That shifts the baseline to match the physiology actually being measured, instead of widening a range that was never built for it.
This isn't a simple widening of the standard interval; it's a different baseline entirely, drawn from the clinical and exercise-physiology literature on how these compounds and this level of training activity actually shift each marker, refined against real reports over time. Take HDL: the on-cycle band's lower bound sits meaningfully below the natural band's, because AAS users legitimately run lower HDL as an expected pharmacological effect, while the upper bound doesn't move, because a high HDL is still a good thing regardless of context.
An Optimal Window Inside a Wider Normal Band
GearCheck doesn't draw a single line between normal and abnormal. Each marker carries two nested bands: a stricter "optimal" window, and a wider "normal" band around it.
- Inside the optimal window, a marker contributes full points and needs no explanation.
- Between optimal and the edge of the normal band, a marker reads as a mild drift — worth noting, not alarming.
- Outside the normal band entirely, deviation is graded on a curve tuned to that specific marker, and a small number of markers carry a hard safety ceiling regardless of context — a severely reduced eGFR or a critically high hematocrit caps the picture no matter how well-explained everything else is.
Confirmatory Marker Cross-Checks
Certain markers get skewed by training itself, so GearCheck won't act on them alone. If your creatinine-based eGFR reads low, it has to be confirmed with Cystatin C. Elevated AST or ALT gets cross-checked against GGT and CK. That catches the most common false positive patterns before they ever trigger an alert.
Trend-Aware Interpretation
A single out-of-range value tells you less than the direction it's moving. GearCheck weights trends heavily, so a stable "abnormal" reading gets interpreted differently than one that's rapidly worsening. That's how experienced clinicians actually think, not how statistical software flags outliers.
Consider a 35 year old on 200 mg testosterone cypionate weekly. His hematocrit has been stable at 48% for a year. A new blood test shows 52%. That is a 4 point jump, not a gradual drift. The trend matters more than the absolute value. A stable 52% in a lifelong athlete with high altitude training may be acceptable, but a rapid rise from 48 to 52 in three months suggests the dose is too high or he added an erythropoiesis stimulator. GearCheck's trend weighting would flag this as a warning, while a single value of 52% might not trigger an alert if it is within the athletic band. The direction of change is the first clue to intervention.
When to Trust the Standard Range
Standard ranges aren't universally wrong. For markers that training or AAS don't affect, they remain accurate and should be trusted. These include:
GGT
Because GGT is liver-specific and not elevated by training, a standard-range elevation is a genuine signal of hepatobiliary stress.
Cystatin C
Cystatin C is a muscle-independent kidney marker, and standard ranges hold up because neither training nor AAS systematically shift it.
Platelets
Training doesn't meaningfully shift this value, so the standard thresholds for thrombocytosis still hold up for athletes.
TSH, Free T4, Free T3
Thyroid markers have well-established normal ranges that hold across populations, and athletic activity does not significantly alter them.
Thyroid markers are a good example of where standard ranges work. Free T3 and free T4 are tightly regulated and not significantly altered by training or AAS use. However, total T3 and total T4 can be affected by changes in binding proteins, which some AAS influence. If you are using an aromatase inhibitor or an oral AAS, your SHBG may drop, which can lower total hormone levels while free hormone remains normal. Always check free T3 and free T4, not just total. A standard range for TSH is reliable, but if you are on levothyroxine, your target TSH may be lower than the population range. The key is to use standard ranges for thyroid markers unless you have a known thyroid condition, and to interpret them in the context of your medication and AAS use.