Blood Test Reference Ranges Explained (UK, 2026)

By Aether, edited by a human editor · Last updated 22 May 2026 · ~12 min read

Almost every email we get about a private blood test result is, fundamentally, a question about reference ranges. "My ferritin is 28, the range says 15–200, so I'm fine — but why am I still tired?" "My TSH is 4.1, range 0.4–4.5, my GP says it's normal, the internet says it isn't." "Why is my Medichecks result different to the one my GP did three weeks ago?"

The reference range is doing more work than people realise — and less work than people think. This guide explains, in plain English, what UK blood test reference ranges actually are, why NHS and private labs publish different ones for the same marker, why "in range" is not the same as "optimal", and how to read your own result without misinterpreting it.

What a reference range actually is

A reference range — sometimes called a reference interval — is a statistical band. The lab collects results from a defined reference population (usually adults considered healthy at the time of sampling, sometimes split by sex), measures the marker, and publishes the central 95% of the resulting distribution as the range. The lower limit is the 2.5th percentile; the upper limit is the 97.5th.

That definition has three consequences most patients are never told:

• By definition, 5% of healthy people fall outside it. If you run 20 markers on a healthy person, statistically one will flag. This is why multi-marker panels almost always produce at least one "out of range" result that means nothing.
• The "healthy" population is not necessarily you. If the reference data came from blood donors aged 18–65, the range may not represent a postmenopausal woman, a teenager, a Black patient, or someone in pregnancy. Subgroup ranges exist for some markers and not others.
• It is a description of population variation, not a definition of health. A result inside the range is not proof of disease absence. A result outside the range is not proof of disease. It is one data point among many.

The UK clinical-biochemistry community publishes reference-interval guidance through the Association for Laboratory Medicine (formerly the Association for Clinical Biochemistry and Laboratory Medicine) and follows international Clinical and Laboratory Standards Institute (CLSI) methodology. UK accredited labs operate under ISO 15189 administered by UKAS, which requires reference intervals to be established, verified, or transferred (and then reviewed) for each assay.

Why NHS and private labs disagree

If you have ever compared an NHS thyroid report with a Medichecks, Thriva or Forth one, you have probably noticed the printed ranges are not identical. That is not a mistake. There are four legitimate reasons two UK labs publish different ranges for the same marker:

1. Different assay. The chemistry kit (the antibody and detection system) used to measure a hormone or protein varies by manufacturer. A Roche TSH assay and an Abbott TSH assay don't return identical numbers from the same sample. Each assay has its own reference interval, established or verified against that assay.
2. Different reference population. A regional NHS lab and a private lab sampling national mail-order customers can produce subtly different distributions even when both are healthy adults.
3. Different reference-interval methodology. Some labs use the classic parametric 95% interval; others use non-parametric, robust, or transferred intervals from large multi-centre studies. The maths is different and can shift a limit by 5–10%.
4. Different review cycle. Reference intervals are reviewed periodically. One lab may have updated its TSH upper limit from 5.5 to 4.2 mIU/L; another may not have. Both are defensible at the time they were set.

The practical rule: always read your result against the reference range printed on the same report. Comparing the number alone against a range from a different lab, an online calculator, or last year's NHS bloods is the single most common mis-reading we see. It is also why our test pages quote "typical UK reference ranges" rather than absolute ones — yours will sit somewhere in that neighbourhood, but the exact limits on your report are the ones that count.

Why "in range" is not the same as "optimal"

A reference range tells you that your result is one of the central 95% of the reference population. It does not tell you that your result is the value associated with the best long-term outcome. For some markers — and only some — there is evidence for a tighter "optimal" sub-range that sits inside the population reference range.

The most common examples in UK private testing:

• Ferritin. Population reference range typically 15–200 µg/L (women) and 30–400 µg/L (men). The British Society for Haematology has argued that anything below roughly 30 µg/L should be treated as iron-deficient and that values up to 50 µg/L can plausibly produce symptoms such as fatigue and hair shedding. "In range" at 18 µg/L is not the same picture as 80 µg/L. See the ferritin test page for the full nuance.
• Vitamin D. The reference range printed on a private report often starts at 25 or 30 nmol/L. NHS guidance via NICE treats below 25 nmol/L as deficient and 25–50 nmol/L as insufficient; many endocrinologists target ≥75 nmol/L for bone-health outcomes. "In range" at 32 nmol/L is technically true and clinically borderline. See the vitamin D test page.
• HbA1c. 42–47 mmol/mol is the NHS prediabetes range; below 42 is normal, ≥48 is diabetes. Within "normal" (e.g. 39 vs 41), the cardiovascular-risk gradient is real but small and rarely actionable on its own. See the HbA1c test page.
• TSH on levothyroxine. The population reference range often extends to 4.0–4.5 mIU/L, but patients on thyroid replacement are typically targeted toward 0.5–2.5 mIU/L. "In range" at 3.9 is technically true but may not be where a treated patient should sit.
• LDL cholesterol. Reference ranges describe the population; cardiovascular targets come from risk-based NICE guidance and depend on age, family history, smoking, blood pressure and existing cardiovascular disease. "In range" at LDL 4.5 mmol/L is statistically common but not the target for a 55-year-old with a family history of early cardiac disease.

The opposite is also true. For many markers there is no evidence-supported "optimal" sub-range and chasing one is futile or harmful. Pushing TSH below 0.5 mIU/L in a healthy person, or pushing ferritin into the 200s "to feel better", is not a strategy any UK clinical guideline endorses. The honest summary: "in range vs optimal" is a real and useful distinction for a short list of markers and a misleading idea for everything else. Be sceptical of any private report or wellness brand that paints an "optimal zone" across every marker — most of those zones are marketing, not evidence.

The lab-to-lab variation problem

Even with identical reference ranges, two real measurements of the same blood are never identical. The total variation in a result comes from three sources:

• Analytical variation. The assay itself has a coefficient of variation (CV) — typically 2–5% for routine biochemistry, 5–10% for hormone immunoassays, occasionally higher for low-concentration markers. Two runs of the same sample on the same machine will differ by roughly the CV.
• Biological variation within an individual. Your own TSH, ferritin or cortisol moves day to day, hour to hour, with sleep, illness, hydration, food and stress. For TSH this is on the order of 20–30% across a day; for cortisol it is many times that.
• Pre-analytical variation. Sample handling — time in the post for a fingerprick kit, haemolysis, fasting status, time of day, posture during a venous draw — moves results before the assay even starts. The post is a real and underappreciated source of noise on UK home-testing kits.

Add those up and a single result has a "true value" plus a meaningful noise envelope. A borderline flag on one report and a borderline-normal on another, from the same person two weeks apart, is usually noise — not a real change. This is why responsible clinicians repeat borderline abnormal results before acting, and why a single out-of-range value in an otherwise unremarkable panel is rarely worth losing sleep over.

The corollary: if you want to compare results over time, use the same provider and the same sample type (fingerprick or venous) each time. Switching between Medichecks, Thriva and an NHS lab gives you three numbers that are not directly comparable. For the rationale on which sample type to pick, see fingerprick vs venous blood tests.

Age, sex, ethnicity, pregnancy: when subgroup ranges matter

A single adult reference range is a simplification. For some markers, well-recognised subgroup ranges should apply:

• Sex. Ferritin, haemoglobin, creatinine, urate and several hormones have established sex-specific ranges. Most UK private reports apply these by default.
• Pregnancy. TSH, free T4, creatinine, alkaline phosphatase, D-dimer and several others have trimester-specific ranges. A standard adult range will misclassify a normal pregnancy result.
• Age. Creatinine and eGFR shift with age; PSA upper limits are sometimes age-adjusted; alkaline phosphatase is far higher in adolescents than adults.
• Ethnicity. HbA1c can run subtly different across ethnic groups even at the same average glucose; some haematology markers vary too. The clinical adjustment is small but real in some guidelines.

Most UK private reports default to a single adult range with sex adjustment, and that is almost always appropriate. The exceptions worth raising with a clinician are pregnancy (always — never read a pregnancy result against a non-pregnancy range), young adolescents (rare on private testing) and chronic kidney disease where eGFR formulas have changed recently. If your provider has not used a relevant subgroup range and your result is borderline, ask before acting.

Reading your own report without misinterpreting it

A short, opinionated checklist:

1. Use only the range on the same report. Not the range your GP used. Not the range from a Reddit thread. Not the range on a competitor's marketing page.
2. Note the units. UK reports use SI units (nmol/L, mIU/L, µg/L). US sources often quote different units (ng/dL, pg/mL). A number is meaningless without the unit.
3. Count the flags in context. One mild flag on a 30-marker panel is statistically expected. Three flags pointing at a coherent picture (e.g. low ferritin + low haemoglobin + high red-cell distribution width) is a signal.
4. Distinguish "out of range" from "clinically significant". An ALT of 56 when the upper limit is 50 is technically flagged and almost never significant on its own. A TSH of 12 when the upper limit is 4.5 is a different conversation.
5. Repeat before acting on a single borderline result. Same provider, same sample type, 4–8 weeks later (longer for ferritin, shorter for inflammatory markers).
6. Take the PDF to a GP if anything is meaningfully flagged. NHS GPs will look at private results, especially for clearly out-of-range values. Don't self-diagnose, don't self-medicate, and don't change existing medication based on a single home test.

For a marker-by-marker walkthrough of what flags actually mean, see how to read blood test results. For what the test itself measures, the test guides index covers every common UK private marker.

Regulation: who checks the labs and the ranges

UK clinical laboratories — NHS and private — operate under UKAS accreditation to ISO 15189. The standard explicitly requires labs to establish, verify or transfer reference intervals and to review them periodically. UKAS-accredited labs are the floor; you should expect a UK private provider to use one, and the report or the provider's website should say so. The major UK private labs (TDL, County Pathology, Eurofins, the Doctors Laboratory partner labs used by Medichecks, Thriva and Forth) are UKAS-accredited.

Patient-facing guidance on what blood tests mean is published by the NHS via the NHS Blood Tests hub and by the Royal College of Pathologists / Association for Laboratory Medicine via Lab Tests Online UK. Both are good starting points; neither replaces a clinician reviewing your specific report.

FAQ

Related on Blood Test Guide UK

• Private blood tests UK — the complete 2026 guide
• Ask Aether — paste a result, get plain-English meaning

• How to read your blood test results
• Private blood test cost UK (2026)
• Private thyroid panel UK cost
• All UK private blood test guides (index)
• Ferritin — why "in range" can still mean depleted
• Vitamin D — NHS thresholds vs "optimal" zones
• Thyroid — TSH, T4, T3 and antibodies
• HbA1c — what counts as prediabetic