How to read a blood test: your results in plain language

How to read a blood test comes down to four things on each line: the name of the test, your result, the units it is measured in, and the reference range the laboratory printed beside it, with a flag such as H or L when the result falls outside that range. Then read related results together, and against your own earlier results where you have them. A result just outside the range is common in healthy people. It is something to talk through with a doctor, with the report in front of you, not a diagnosis.

How to read a blood test report, column by column

Here is an example row, a fasting glucose, written in both of the unit systems laboratories use. The result is made up. The ranges are the ones MedlinePlus, the US National Library of Medicine's health site, gives for a fasting glucose.

• Test. The marker measured, often shortened: TSH for thyroid-stimulating hormone, eGFR for estimated glomerular filtration rate.
• Result. The value found in your sample, measured by the laboratory or, for some markers, calculated from other results.
• Units. What the number is counted in. Always read the result against a range written in the same units.
• Reference range. The span of results the laboratory expects in healthy people, also called the normal range or reference interval. Some markers, such as ferritin and creatinine, have separate ranges for men and women.
• Flag. H for high or L for low, sometimes an asterisk or bold type, when the result sits outside the range. A flag says where the number falls, not what it means.

The top of the report matters too. The date and time the sample was taken, and whether you had fasted, change how some results read. Glucose, for example, varies through the day and needs a fasting or timed sample, according to the American Diabetes Association (ADA).

What is a reference range based on?

A reference range describes other people, not you. A laboratory typically sets one by testing at least 120 healthy people, setting aside the lowest 2.5% and the highest 2.5% of their results, and calling the middle 95% the range. By definition, then, 5% of healthy people fall outside it on any one test.

The effect adds up across a panel. A 2025 review in the Journal of the American Board of Family Medicine worked through the arithmetic: on a panel of 20 tests, about 64% of healthy people would have at least one result outside the range, assuming the tests are independent of each other. On a panel of more than 100 results, some flags are to be expected even in good health.

Ranges also differ between laboratories. Each laboratory is meant to set its own, but the same review notes that in practice many adopt them from textbooks and from the makers of their equipment and reagents, and that methods differ. For ferritin in women, MedlinePlus gives 13 to 150 ng/mL, while Mayo Clinic Laboratories gives 6 to 175 µg/L for women aged 18 to 50 (the two units give the same number). That is why MedlinePlus says to read your result against the range on your own report, and not to compare results from different labs.

Who the range was built on matters as well. In a study of 394 healthy Emirati adults of Bedouin Arab origin, with an average age of about 25, 53% had at least one red cell value outside the usual range: a low haemoglobin, MCV or MCH, or a high RDW. The researchers pointed to how common iron deficiency and inherited thalassaemia traits are in this group: they estimated alpha-thalassaemia trait in 32% of the men, and found a ferritin below 30 µg/L in 65% of the women. They argued for red cell reference standards drawn from the local population.

Two more points. Some ranges are not population spreads at all but decision limits that experts set from outcome data: the review notes that the normal range for cholesterol has been lowered over the years for this reason, and HbA1c is read against thresholds set by the ADA. And a result inside the range does not prove good health: MedlinePlus notes that some people with health problems have results considered normal.

Why do the same results look different in SI and conventional units?

Laboratories report in one of two systems: SI units, most of them per litre (mmol/L, µmol/L, nmol/L), and conventional units, most of them per decilitre or millilitre (mg/dL, ng/mL). The same result looks very different in each.

• A fasting glucose of 70 to 99 mg/dL is 3.9 to 5.5 mmol/L, in MedlinePlus's figures.
• An HbA1c of 6.5% is 48 mmol/mol, and 5.7% is 39 mmol/mol. The ADA writes its thresholds in both.
• Vitamin D of 30 to 80 ng/mL is 75 to 200 nmol/L, again in MedlinePlus's figures.

So before you compare a result with an older report, a range you found online or a friend's numbers, check the units. A vitamin D of 50 means something quite different in ng/mL than in nmol/L.

What do blood test results mean when they are slightly out of range?

Usually less than the flag suggests. MedlinePlus puts it plainly: it is common for healthy people to sometimes have results outside the reference range. The 2025 review makes the same point from the other side: the number of results outside the normal range far exceeds the number that are clinically meaningful.

Ordinary things move results. MedlinePlus lists food and drink, medicines and supplements, hard exercise before the test, a menstrual period at the time of the test, and not following the preparation instructions, fasting included. Some effects are specific:

• Glucose reads falsely low if the sample is not processed promptly, the ADA notes.
• An acute illness can affect thyroid results, which is why NICE advises against testing the thyroid during one, unless the illness itself may be thyroid-related.
• A high intake of biotin from supplements can make thyroid results falsely high or low, according to NICE, and Mayo Clinic Laboratories asks people not to take supplements containing biotin for 12 hours before a ferritin test. Tell the person taking your sample about any supplements you use.

A doctor usually reads a borderline result in context: how far outside the range it is, whether related markers agree, how you feel, your earlier results, and whether a repeat test makes sense. For diabetes, for example, the ADA says that without the classic symptoms, a result above a diagnostic threshold has to be confirmed by repeat testing.

The 2025 review suggests reports sort results into three plain tiers: minor variations, results to discuss at the next visit, and results that call for contacting a doctor soon. Which tier applies is a judgement for a doctor who knows your history.

Measured or calculated: what is the difference?

Not every number on a report comes straight off an analyser. Some are calculated from other results. MedlinePlus notes that the LDL cholesterol on a report may say "calculated", meaning it is an estimate based on total cholesterol, HDL and triglycerides, while some laboratories measure it directly. eGFR, an estimate of how fast the kidneys filter blood, is worked out from creatinine with an equation. The US National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) stresses that eGFR is an estimate, not a precise measure, and asks laboratories to name the equation and date any change so that trends read correctly.

Calculated results inherit whatever moves their inputs: if creatinine shifts, eGFR shifts with it. So it helps to know which is which, and Clare's panel says so for every result. Of its more than 100 results, 88 are measured by a licensed lab partner and 23 are calculated from those measurements, and the full panel labels each one "From the lab" or "Derived biomarker". Some of the calculated ones:

• Non-HDL cholesterol: total cholesterol less HDL.
• Transferrin saturation: serum iron as a percentage of TIBC, the total iron-binding capacity.
• HOMA-IR: fasting insulin and fasting glucose combined into one estimate of insulin resistance.
• eGFR: worked out from creatinine with a validated equation.

How do you read blood test results together?

Many markers make sense in pairs or groups, because each one answers part of the same question. Three examples from Clare's panel:

Ferritin with the full blood count

Ferritin reflects how much iron the body holds in reserve. The full blood count shows whether the red cells are affected: according to British Society of Gastroenterology guidelines, iron deficiency tends to make red cells smaller (a lower MCV) and paler (a lower MCH). Read together, the two tell a doctor more than either does alone. Small red cells with a low ferritin usually point toward depleted iron stores. Small red cells with a ferritin that is not low send a doctor looking at other causes, inherited haemoglobin traits such as thalassaemia among them, which the same guidelines list as a cause of small, pale red cells.

There is a catch. Ferritin rises with inflammation, so it can look normal when stores are low. That is why a marker of inflammation such as CRP is often read alongside it, and why the WHO uses a ferritin below 15 µg/L to indicate iron deficiency in apparently healthy adults, but below 70 µg/L in adults with an infection or inflammation.

TSH with free T4

TSH is the pituitary gland's signal to the thyroid. Free T4 is the main hormone the thyroid releases in response. For adults, NICE suggests starting with TSH, then measuring free T4 in the same sample if TSH is above the range, and free T4 and free T3 if it is below. TSH reacts strongly to small shifts: the authors of a study of healthy men note that it responds with logarithmically amplified variation to minor changes in thyroid hormones, so TSH can leave its range while free T4 stays inside its own. NICE uses the term subclinical hypothyroidism for exactly that pattern, a TSH above the range with free T4 within it, and what it means for a given person is for a doctor to judge. How we read a thyroid panel goes through the patterns.

HbA1c with fasting glucose

Fasting glucose is one moment. According to the ADA, it moves with food, stress, recent illness and activity, and varies a lot within the same person, while HbA1c reflects glucose over roughly the past two to three months and varies little. But HbA1c depends on red cells, so conditions that change their turnover change it too, anaemia, iron status, blood loss and transfusion among them, and some haemoglobin variants interfere with certain methods. That links it back to the blood count and ferritin. When the two disagree, the ADA's guidance is to repeat the test that is above its diagnostic cut point, considering what may be affecting either result. HbA1c, glucose and insulin: reading them together covers this in more depth.

Why is one blood test result a snapshot?

Even in good health, results move. Each person's values drift around a personal set point from week to week, which laboratory scientists call biological variation. The European Biological Variation Study (EuBIVAS) measured it by sampling 91 healthy adults every week for 10 weeks at six European laboratories. Some markers barely move; others swing widely.

In plain terms, a healthy person's TSH typically varies by about 18% around their own average from week to week, while their HbA1c varies by about 1%. The glucose and HbA1c figures come from a 2020 meta-analysis; the others from EuBIVAS.

That has two consequences. A personal range is narrower than a laboratory range: in a Danish study of 16 healthy men tested monthly for a year, each man's own range for thyroid tests was about half as wide as the group's, and a single result pinned down his usual TSH to within about 50% either way. So a result inside the laboratory range can still be a change for you. And a change between two tests has to be large enough to stand out from that background: in the same study, TSH had to move by about 0.75 mU/L on average before a change could be called real with 95% confidence. Laboratories use biological variation figures to work out these thresholds, which they call reference change values.

This is also why MedlinePlus advises using the same lab when you are looking for trends, and why the 2025 review notes that repeat results within one laboratory vary much less than results between laboratories.

When should you talk to a doctor about your blood test results?

Talk to a doctor about any report, with the report in front of you. MedlinePlus advises talking with your provider about your results even when they are normal. It matters most:

• When a result is flagged, especially if it sits far from the range or several related results point the same way.
• When a result has changed noticeably from your earlier ones, even if it is still inside the range.
• When you have symptoms, whatever the report says. A normal result does not rule a problem out, and MedlinePlus notes that more testing may be needed, especially when there are symptoms.
• When the laboratory or clinic contacts you. Laboratories report critical values to the clinician.
• Before you change any medicine or supplement because of a result.

Bring earlier reports if you have them, from the same laboratory where possible.

One report compares you with other people. A series of reports from the same laboratory compares you with yourself, and that is where small, real changes show up. When Clare opens, members' results will be kept together over time in one account. Clare is opening with a waitlist at clare.health.