What do the numbers on a blood test actually mean?

Each row is one marker (something measured in your blood), your value, and the unit it is measured in. Next to it is the reference range that lab expects for a healthy person, and a flag (H for high, L for low) if your value sits outside that range. The flag compares your number to that range only; it is a prompt to look closer, not a diagnosis.

What does H or L mean next to a result?

H means your value is higher than that lab's reference range; L means lower. It is a comparison to the printed range, nothing more. About 1 in 20 healthy people get a flag on any single test by chance, so one mild flag is often nothing, though the flagged marker is worth a closer look and sometimes a repeat.

Is one abnormal blood test result something to worry about?

Usually a single mild flag is not, and clinicians typically repeat an abnormal result before acting on it, because extreme values tend to drift back toward your own average on a second test. The exception is a very abnormal value with symptoms, such as chest pain, breathlessness, or a very high glucose with heavy thirst or confusion, which needs same-day care.

Why is my normal range different from a friend's or a website's?

Reference ranges are specific to the lab, method and machine that ran your test, and they shift with age, sex and pregnancy. Compare your result to the range printed next to it on your own report, not to a number from another lab or a general website.

How to read any lab result

The distinctions

01
A lab result is a comparison, not a verdict
A single value means something only against three references: the range printed on that report, your own earlier results, and the reason the test was ordered. Read in isolation, it is only a data point.
02
One flag on a broad panel is expected
Reference ranges are drawn to include the middle 95% of healthy people, so about 1 in 20 healthy results land outside by design. Order 20 tests and the chance at least one flags by chance alone is about 2 in 3.
03
The right range depends on who you are and which lab ran it
Reference intervals shift with age, sex and pregnancy, and they are specific to each lab's method and machine. A number is only high against the correct range, which is the one printed next to it on that report.

A lab report you did not order looks like a wall of numbers with a few letters flagged in red. Read it in four parts: the marker (what was measured), your value with its unit, the reference range that lab expects, and the flag. A flag means one result landed outside that lab's range, which is a signal to look closer before it is anything else.

The instinct is to read a flagged line as a verdict. A single value cannot be a verdict, because it only carries meaning against three things: the range printed on that report, your own earlier results, and the reason the test was run. Get those three references in view and most flags shrink to their real size.

First, read the page: the five columns

Almost every lab report, from any lab, lays out the same columns: four describe your result, and a fifth flags it. Once you can name them, the page stops being a blur. Here is one row read across, using a fasting glucose as the example.

Column	What it tells you	Example
Test / analyte	What was measured	Glucose, fasting
Result / value	Your number	108
Units	The scale the number sits on	mg/dL
Reference interval	The range that lab expects for a healthy person	70 to 99 mg/dL
Flag	How your value compares to that range	H (high)

1Find the marker and your valueRead across the row: the test name, then your number.
2Check the unitmg/dL, mmol/L, mIU/L. The unit is part of the number; the same amount reads as different figures in different units, which is why a foreign report can look alarming until you convert it. See [lab units, explained](/learn/lab-units-explained).
3Compare to the range on that reportNot a range from another lab or a website. The reference interval printed beside your result is the one that applies to it.
4Read the flag, then ask the three questionsAn H or L says your value is outside that range. Whether it matters comes from the three questions below.

The three questions that read any result

Four columns tell you what was measured. Three questions tell you whether it matters. Ask them in order.

The question	Why it matters	The trap to avoid
Is the flag real, or is it noise?	Ranges are built to leave some healthy people outside, so a lone flag is often just chance.¹	Reading a single mild flag as a diagnosis.
Is this the right range for me?	Ranges are specific to each lab and shift with age, sex and pregnancy.¹	Comparing your number to a range you found online instead of the one on your report.
Is it a trend or a one-off?	One value is a snapshot; the direction across several results carries the signal.⁴	Acting on one odd number before it is repeated.⁵

Question 1: is the flag real, or is it noise?

A reference range is drawn on purpose to leave some healthy people outside it. Exactly how many fall outside, and why the cutoff sits where it does, is unpacked in reference ranges, explained.

Reference ranges are drawn to include the middle 95% of a healthy population, so about 1 in 20 healthy people fall outside the range on any single test.

reference ranges, explained

One test carries that 1 in 20 chance of a meaningless flag. Run twenty independent tests and the chance at least one flags by chance alone climbs to about 2 in 3, because each independent test carries its own small chance and those chances compound.¹ On a broad panel, one flag is the expected result.

More tests, more flags by chance alone

Illustrative. With reference ranges covering the middle 95% of healthy people, a single test flags a healthy person about 5% of the time; the chance at least one of n independent tests flags is 1 minus 0.95 to the n.[[1]]

Question 2: is this the right range for me?

A number is only high against the correct range, and the correct range is the one printed next to your result. Reference intervals are specific to each lab's method and instrument, so a cutoff copied from another lab or a website may not apply to your report.¹ They also shift with who you are.

Age, sex and pregnancy all move the range. The estimating formula can change too: in 2021 the standard equation for eGFR, the kidney-function estimate calculated from creatinine, dropped its race coefficient (an adjustment the old formula applied based on a patient's race), so the same creatinine value can read as a different eGFR on a newer report than an older one.³ Nothing about your body changed; the formula did.

Marker	A range in one context	A different context	Why it moves
Hemoglobin	About 13.5 to 17.5 g/dL (adult men)⁶	About 12.0 to 15.5 g/dL (adult women)	Sex-specific physiology
TSH	About 0.4 to 4.0 mIU/L (general adult)⁶	Lower, trimester-specific ranges in pregnancy	Pregnancy shifts thyroid hormone handling
eGFR from creatinine	Older reports used a race coefficient	2021 equation removed it, so the same creatinine can read differently³	The estimating formula changed

Question 3: a trend or a one-off?

One value is a snapshot, and every marker carries built-in day-to-day wobble.⁴ A result can move between two draws without your health changing at all, from normal biological variation, the lab's own imprecision, and pre-analytical factors like whether you fasted or how long the tourniquet was on. A change counts as real only once it clears that noise, which is why a single odd number is usually repeated before anyone acts on it.⁵ The direction across several results, tracked in trends beat single readings, carries the signal.

This is standard clinical practice. The American Diabetes Association, for example, requires a repeat before diagnosing diabetes from a lab value.²

In the absence of unequivocal hyperglycemia, diagnosis requires two abnormal test results from the same sample or in two separate test samples.American Diabetes Association, Standards of Care in Diabetes 2024 [[2]]

Where an example fasting glucose of 108 mg/dL fallsmg/dL

108 mg/dLjust above the standard cutoff · same as about 6.0 mmol/L

Low54–69Typical fasting range70–99Above range100–125Diabetes range on repeat testing126–200

In this illustrative example, a fasting glucose of 108 sits just above the typical range, in the band clinicians usually confirm with a repeat test rather than act on immediately. Treat it as a reason to ask about retesting; on its own it does not diagnose diabetes.

Illustrative example. Fasting glucose tiers follow the American Diabetes Association 2024 cutoffs: 100 to 125 mg/dL and 126 mg/dL or higher define higher categories on repeat testing.[[2]]

One more piece of context sits off the page: why the test was run in the first place. A ferritin (an iron-storage marker) checked because you are exhausted reads differently from the same ferritin in a routine panel. If you do not know why a test was on the order, that is a fair thing to ask.

Link · MedlinePlus (US National Library of Medicine)Understanding your lab resultsA plain-language index of common lab tests and what each one measures, maintained by the US National Library of Medicine.medlineplus.gov

LifeFrom membershipWant your own results read in context instead of just flagged? LifeFrom explains each marker against your ranges, your history, and the reason it was measured, so a single flag comes with the interpretation it needs.Get started

Citations

Educational context only, not medical advice or a diagnosis. Always discuss your results with a clinician.