What is the difference between a reference range and an optimal range?
Written and medically reviewed by Zahraa Sater, M.D., M.P.H.
Reviewed August 12, 2026 · Next review due August 12, 2028
Short answer
A reference range is the interval holding the middle 95 percent of results from a laboratory's healthy reference group, measured on that laboratory's own assay. An optimal range is a narrower target that somebody has chosen. A few optimal targets come from outcome trials, such as LDL cholesterol and HbA1c goals. Most optimal ranges sold in wellness testing have no trial behind them.
Reference ranges and optimal ranges are different kinds of claim. A reference range is a description of a population: this is what healthy people's results look like on this machine. An optimal range is a prescription: this is where your result should sit for you to feel or function better. The first is a measurement standard. The second needs evidence that moving your number to the target actually changes an outcome, and that evidence exists for some tests and not for others.
How a reference range is built
Laboratories recruit apparently healthy reference subjects, run the analyte, and set the limits at the 2.5th and 97.5th percentiles of the resulting distribution. A 2016 review in Biochemia Medica states that at least 120 reference subjects are needed to estimate those limits with acceptable confidence, and that intervals belong to the specific assay used. CLSI EP28-A3c governs the process. This is why a result can be flagged high at one laboratory and unflagged at another, and why comparing a number against a range printed by a different laboratory is not meaningful.
Where optimal ranges come from
Some are genuine treatment targets built from outcome trials. The 2018 AHA/ACC cholesterol guideline uses LDL thresholds because lowering LDL lowers cardiovascular events. HbA1c cut points define prediabetes and diabetes in the ADA Standards of Care because they predict progression and complications. Other optimal ranges are invented by extrapolating from a population curve: someone observes that healthy young people cluster in one part of the reference range and declares that part optimal. That reasoning does not establish that moving your value there helps you.
Three questions that expose a weak target
Ask where the target came from, which outcome it improves, and what happens differently if you reach it. A target with a named trial and a named outcome is usable. A target defended by the phrase that the range is too wide, with no outcome attached, is a sales position. This matters because chasing an unvalidated target usually means taking something: thyroid hormone for a TSH of 3, high dose vitamin D for a level of 35 ng/mL, iron for a ferritin of 45 mcg/L. Each of those carries real risk.
| Target you may be given | Where it comes from | Status |
|---|---|---|
| HbA1c below 5.7 percent | ADA Standards of Care diagnostic threshold | Evidence based |
| Lower LDL cholesterol in high risk patients | 2018 AHA/ACC guideline, outcome trials | Evidence based |
| TSH below 2.5 mIU/L in non-pregnant adults | Extrapolation from population distribution | Contested, not a society treatment target |
| Ferritin above 100 ng/mL for hair or energy | No outcome trial | Unsupported |
| 25-hydroxyvitamin D of 60 to 80 ng/mL | No outcome trial | Unsupported; NIH lists above 50 ng/mL as potentially harmful |
When a clinician proposes a target outside the laboratory range, ask for the study and the outcome. When your result is inside the range but you feel unwell, the answer is usually a different test rather than a tighter target on the same one.
The clinical detail
Reference intervals are population based; clinical decision limits are outcome based. The two are often printed in the same column of a report, which causes confusion. LDL cholesterol, HbA1c and fasting glucose carry decision limits set by guideline committees, not percentile limits from a local healthy cohort. TSH, ferritin, testosterone and 25-hydroxyvitamin D are mostly reported against percentile intervals, and their treatment thresholds vary by guideline and by clinical context.
Analytical variation and biological variation both matter when comparing a result to any target. Repeat testing near a threshold is standard practice for testosterone, which the Endocrine Society requires on two separate fasting mornings, and for diabetes diagnosis, which the ADA requires on two abnormal results unless the patient has unequivocal hyperglycaemia.
Written and medically reviewed by Zahraa Sater, M.D., M.P.H.
Reviewed August 12, 2026 · Next review due August 12, 2028
Sources
- Reference intervals: current status, recent developments and future considerations. Biochemia Medica. DOI 10.11613/BM.2016.001
- EP28-A3c: Defining, Establishing, and Verifying Reference Intervals in the Clinical Laboratory. Clinical and Laboratory Standards Institute
- 2018 AHA/ACC/Multisociety Guideline on the Management of Blood Cholesterol. American Heart Association and American College of Cardiology. DOI 10.1161/CIR.0000000000000625
- Diagnosis and Classification of Diabetes: Standards of Care in Diabetes. American Diabetes Association
- Vitamin D: Fact Sheet for Health Professionals. NIH Office of Dietary Supplements
This library is educational and is not individual medical advice. Read our Editorial and Medical Review Policy.