Reference Ranges vs Optimal Ranges

"Your results are normal" feels like a clean bill of health, but the phrase means something narrower: your value sits inside the band that 95% of a reference population produces. Normal is a statistic, not a verdict — and for a handful of markers, the population average is nowhere near where outcomes are best. This page explains where lab ranges come from, where "optimal" is real, and where it is marketing.

Where "Normal" Comes From

A reference range is manufactured, not discovered. Laboratories recruit a reference population of apparently healthy people, measure the marker, and declare the central 95% of the distribution "normal" — the rest get flagged high or low. The standard is CLSI EP28-A3c, the laboratory-standards document that governs how the sausage is made (Clinical and Laboratory Standards Institute, 2010). Three consequences follow immediately. First, 5% of healthy people are outside the range by construction — being flagged is a mathematical consequence of the method, not evidence of disease. Second, "normal" describes where people are, not where they should be. Third, the range is only as healthy as the reference population: build it from a modern, mostly overweight cohort and you normalize overweight.

• 🏥 Ranges vary by lab. Different analyzers and different reference populations produce different cutoffs — a value "normal" at one lab can be flagged at another. This is why trends should be read from the same lab, the theme of the trends page.
• 📈 Ranges describe, they don't prescribe. A range is a statistical summary. Whether a given value predicts good outcomes is a separate question, answered only by outcome studies — and mostly unanswered.
• 🕰️ Ranges drift. When a population gets heavier and sicker, the "normal" envelope widens and shifts. Ranges are periodically re-derived — which is the system working, and also the system following the crowd.

The Population Has Drifted

This is the quiet scandal of the reference range: the reference population tracks the average, and the average has moved. A cholesterol value that sat at the 50th percentile in 1960 sits far lower than the 50th percentile today in many Western cohorts, because the population now carries more obesity, more prediabetes, and more early atherosclerosis into the "healthy" reference sample. The consequence is practical: a value described as "normal for your age and sex" can mean "normal for a population that is quietly metabolically unwell." For markers where outcomes improve continuously as the value falls — ApoB, glucose, blood pressure — the honest question is never "am I normal?" but "is my number where the risk data say it should be?"

When Optimal Is Real: The Big Four

For four markers, the outcome data are strong enough that a defensible "optimal" exists below the population average. These are the ones where the phrase earns its keep.

• 🧪 ApoB / LDL. Genetic evidence shows lifetime exposure drives atherosclerosis: people born with lower LDL accumulate less plaque at every age (Ference et al., European Heart Journal, 2017). "Lower is better" is as close to settled as cardiovascular medicine gets — the lipid-panel topic has the particle math.
• 🩸 Glucose. In a multicenter study of healthy non-diabetic adults wearing continuous glucose monitors, mean glucose was about 99 mg/dL with ~96% of time spent between 70 and 140 mg/dL (Shah et al., JCEM, 2019). HbA1c risk does not wait for the 5.7% diagnostic line — it rises within the "normal" range. Glucose 101 has the mortality curve.
• 🫀 Blood pressure. Cardiovascular risk rises continuously from roughly 115/75 mmHg — there is no threshold, just a slope. The blood-pressure topic owns the categories and the 130/80 debate.
• 📏 Waist-to-height. A ratio below 0.5 — waist less than half your height — is the free, home-owned target that tracks visceral fat better than BMI. The visceral-fat topic explains why.

When "Optimal" Is Marketing

The wellness-testing industry borrows the credibility of the Big Four and applies it to markers where no comparable outcome data exist. The tell is always the same: a confident, narrow "optimal" band printed on a lab report, with no study behind it.

• ☀️ Vitamin D. The genuine disagreement: the Institute of Medicine set sufficiency at 20 ng/mL for bone health (2011), while the Endocrine Society's guideline preferred 30 ng/mL (2011) — two serious bodies, two numbers, same evidence. Meanwhile randomized supplementation trials found no cardiovascular or cancer benefit (VITAL, NEJM, 2019), which drains most of the urgency from "optimizing" the value. The honest read: correct real deficiency; don't chase a lab-defined optimum.
• 👨 Testosterone. A single morning value in the 400s ng/dL may be "low-normal" on a wellness panel and "normal" to an endocrinologist. The hormone pulses and declines through the day, so one draw cannot define your level — which is why the men's-health pillar anchors treatment to symptoms plus repeated draws, not to a band.
• 🧘 Cortisol and "adrenal fatigue." Cortisol follows a steep daily rhythm and reacts to stress by design. "Adrenal fatigue" panels claim to find subclinical adrenal dysfunction, but endocrine societies do not recognize the condition — the stress pillar covers what cortisol testing can and cannot say.
• 🧬 "Biological age" bands. A clock score of "52" versus your chronological 54 is a measurement with real science behind it and a lot of noise in front of it. The biomarkers-that-mislead page dissects the commercial version.

The TSH Problem

Thyroid-stimulating hormone is the best case study of the reference-range trap, because it looks like the Big Four and is not. The typical upper limit of normal runs around 4.0–4.5 mIU/L, and a value above it labels you "subclinical hypothyroid" — inviting treatment. Three facts complicate the picture. First, TSH rises with age in healthy populations: NHANES data show the distribution shifting upward decade by decade, so many older adults sit mildly above the standard cutoff without any thyroid problem (Surks & Hollowell, JCEM, 2007). Second, a large randomized trial of levothyroxine in older adults with subclinical hypothyroidism found no improvement in symptoms or quality of life (Stott et al., NEJM, 2017). Third, TSH swings within a person by 15–20% on its own, so a single value near the line is mostly noise. The honest synthesis: TSH is worth measuring once as a baseline, and worth repeating with symptoms — but the "optimal range" sold for it is thinner evidence than the industry admits.

Questions, Answered Briefly

• ❓ One of my values is one point outside the range — should I worry? Remember the construction: 5% of healthy people sit outside the band by definition, and one point out is usually noise. Repeat it and watch the trend — the trends page has the repeat-testing logic.
• ❓ Are wellness-lab "optimal ranges" worth paying for? Rarely. Only a handful of markers have outcome-based optima — ApoB, glucose/HbA1c, blood pressure, waist. For everything else the printed "optimal" band is decoration, sometimes borrowed from a different population than yours.
• ❓ Why did my clinician shrug at a flagged value? Because a clinician reads the number with context — symptoms, your prior values, your overall risk — while the lab report reads it in isolation. "Abnormal" is not "diseased"; it is "worth a second look," and sometimes not even that.
• ❓ Should I compare my numbers to friends' numbers? No — set-points differ. Your baseline and your slope are the units that matter; someone else's 82 mg/dL glucose tells you nothing about whether your 95 is a problem.

Related Topics