Lab reference range: why the normal on your report is not your personal norm
What a lab reference range is, why it differs between labs, how much a value moves in the same person and when a change inside the range still matters.
Visionary entrepreneur driving innovation at the intersection of healthcare, AI and global diagnostics. Author of the Previva methodologies. He has worked for many years in preventive medicine, longevity and early detection of disease. He writes from his own experience and vision.
A health professional is reviewing this text and will sign here when done.
Key points
- A reference range holds the central 95% of results from a group of healthy people, not the ideal value for one individual.
- By construction, 1 in 20 healthy people falls outside the range for any given marker, and that alone means nothing.
- Two labs can report different ranges for the same blood, because they use different methods, samples, populations and units.
- A value can sit inside the range and still be a meaningful change compared with your own history: each marker's biological variation tells you how far it has to move to count.
You open your blood test report and see, next to each result, two numbers in brackets. Inside the bracket, you relax; outside it, your pulse jumps. That bracket is the reference range, the lab reference values, and almost nobody explains where it comes from or why it changes from one lab to the next. The short answer: the range describes where most of a group of healthy people fall, not where you should be.[1] Here is how it is built, why two labs print different ranges, how much a value moves in the same person when nothing has changed, and when a change, inside or outside the range, deserves a conversation with your doctor.
What a reference range is and where its numbers come from
To build one, the lab measures a group of people considered healthy, sorts the results from lowest to highest and keeps the central 95%: the lower limit is the 2.5th percentile and the upper limit the 97.5th.[1] The international reference guideline, C28-A3 from the Clinical and Laboratory Standards Institute (CLSI), calls for at least 120 people per group so those edges are stable.[1] The European recommendation puts it this way: an interval that corresponds to 95% of the population, centered on the median.[2]
In other words, the range describes where most healthy people fall. It does not describe where you should fall, what your optimal value is, or what is right for your stage of life. That is why the Mexican standard for clinical laboratories requires every report to print the reference values for the method used and for the sex and age group they apply to:[8] the range describes a group, and the report tells you which one.
How a range from somewhere else gets verified
Not every lab recruits 120 healthy people for each marker. The usual approach is to adopt a published or manufacturer range and verify it in the lab's own population: the CLSI guideline describes that transfer using 20 people from the receiving lab,[1] and the European recommendation accepts the range if no more than 2 of those 20 fall outside.[2] That procedure checks that the range works for the population the lab serves. It does not check that it works for you.
Different reference values between labs: why it happens
This is what confuses people most: same test, same person, same week, and the brackets do not match. It is not a mistake. The range depends on the analytical method and the sample type, and differences between methods can be large, especially in immunochemistry tests such as many hormones.[2] It also depends on the equipment and reagents[7] and on the population it was built or verified with: age, sex, ethnic background, weight and eating habits.[1] MedlinePlus is blunt about it: results from different labs cannot simply be compared, and to follow a trend it is best to use the same lab.[6]
| What changes between two labs | Why it moves the result or the range | How to compare |
|---|---|---|
| Method, equipment and reagents | Two methods can give different numbers from the same sample; the gap tends to be larger for hormones and other immunochemistry tests.[2][7] | Compare results from the same lab. If you switched, note the method on each report and ask whether they are equivalent. |
| Reference population | The range is built or verified with people from a specific population, with their age, sex, background and habits.[1] | Do not use the range on one report to judge the result on another. |
| Sex and age | They are the most common reason for different ranges for the same marker.[7] The Mexican standard requires printing the range for the matching sex and age group.[8] | Check that the printed range is the one for your sex and age. |
| Sample type | The range can vary with the sample type, for example serum or plasma.[2] | Check that both reports used the same sample. |
| Units | The same value can come in mg/dL or mmol/L, in g/dL or g/L; the Mexican standard requires the general system of units of measurement.[8] | Convert both results to the same unit before comparing. The percentage change does not depend on the unit. |
| Conditions of the draw | Time of day, fasting and physical activity move several markers; some labs give different ranges depending on time or fasting.[1][7] | Repeat the draw at the same time, with the same hours of fasting and no intense exercise the day before. |
If this is your first time facing a full report, how to read your lab results without panicking walks through its parts, and the most common panels have their own guides: complete blood count and blood chemistry panel. For hormones such as TSH, which are measured by immunochemistry and where methods differ most,[2] it is worth reading the thyroid panel guide too before comparing two labs.
How much does a result have to change for the change to be real?
This is the point that most changes how you read a report. Each person moves around their own set point. Within-person variation (CVI) is how much a value fluctuates, on average, around that point; between-person variation (CVG) is how far apart the set points of different people are.[4] For many markers the second is several times larger than the first: each person has their own point, and that point can sit anywhere inside the population range.[3] When within-person variation is small compared with between-person variation, the population range is of little use for spotting changes in one individual, and the useful comparison is with yourself.[1]
The European Federation of Clinical Chemistry and Laboratory Medicine (EFLM) collects these figures in a public database, built from pooled reviews of studies appraised with a common method; by 2020 it already held 488 publications and 1,782 records, and it is the first source experts recommend checking.[4] From within-person variation and the imprecision of the lab's method (CVA) you can calculate the reference change value (RCV), the tool for judging whether the difference between two serial results from the same person is significant: RCV = √2 × Z × √(CVA² + CVI²), where Z is the number of standard deviations that matches the chosen probability.[5]
The table brings together the variation of seven common markers from the EFLM database and the minimum change calculated with that formula for a 95% probability (Z = 1.96), assuming a method with no imprecision. It is a floor: with each lab's real imprecision, the change needed is somewhat larger.[3][5]
| Marker | Within-person variation (CVI, EFLM) | Between-person variation (CVG, EFLM) | Minimum change between two results (calculated with CVA = 0) |
|---|---|---|---|
| Glucose | 4.9% | 8.1% | 14% |
| Total cholesterol | 5.4% | 16.0% | 15% |
| LDL cholesterol | 7.6% | 23.8% | 21% |
| Creatinine | 4.4% | 16.1% | 12% |
| Hemoglobin | 2.7% | 6.2% | 7.5% |
| Glycated hemoglobin (A1C, IFCC units) | 1.6% | 7.3% | 4.4% |
| ALT (liver enzyme) | 12.4% | 29.8% | 34% |
My glucose went from 88 to 97: is that a real change?
Probably not on its own. From 88 to 97 mg/dL (4.9 to 5.4 mmol/L) is a 10% rise, below the 14% that glucose's biological variation allows between two draws from the same person.[3][5] What helps is checking whether both draws were fasting and at the same time, and looking at the next result: if it rises again, the direction starts to count. Our guides to high fasting glucose and A1C explain what other data go with that number.
Creatinine from 0.80 to 0.95, hemoglobin from 14.0 to 13.2
Two more examples, with results that may still be inside the range on your report. Creatinine going from 0.80 to 0.95 mg/dL (about 71 to 84 µmol/L) is a 19% rise, more than the 12% minimum change: even if neither number is flagged, it is a change worth discussing with your doctor along with the conditions of the draw, because hard exercise before the test also moves results.[3][6] Our guide to high creatinine covers what is usually ordered next. Hemoglobin going from 14.0 to 13.2 g/dL is a 5.7% drop, below 7.5%: it fits within the normal fluctuation of one person.[3][5]
Inside the range but changed, outside the range but stable
When you cross the two readings, the range and your history, you get four situations. The European recommendation says it in technical language: a variation in a value, even within the reference range, can have pathophysiological significance.[2] And the other way around, a value outside the range may or may not be a sign of a problem, just as one inside does not guarantee good health.[6]
| Situation | What it usually means | What to do |
|---|---|---|
| Inside the range and stable compared with your earlier results | The most common case: the value fluctuates around your set point.[4] | Save it with the date and conditions of the draw; it is the baseline for the next one. |
| Inside the range, but changed more than the minimum in the table | The lab does not flag it, but relative to you there was a change larger than biological fluctuation.[2][5] For key markers, a rise or a drop can matter even if it stays within limits.[7] | Check that the conditions were the same and raise it at your next appointment, with your earlier results. |
| Outside the range, but the same as in your results for years | It may be your own set point: 1 in 20 healthy people falls outside for each marker.[7] An individual interval is built from several samples from the same person in a steady state.[1] | Do not ignore it: bring the full series to your doctor, who decides whether watching is enough.[6] |
| Outside the range and new | It may be a passing variation from the conditions of the draw, or a signal; the report alone cannot tell.[6] | A scheduled appointment with the result and its context. If it is a critical value, the lab notifies your doctor.[9] |
The reference range answers "where is most of the population?". Your history answers "what changed in me?". Those are different questions and you need both.
An example with round numbers: a value whose range runs from 12 to 30, a person who has been around 14 for years and this year comes out at 27. Still inside the range, so nothing gets flagged, but compared with themselves they nearly doubled. That change may mean nothing or be the most interesting thing in the report, and the bracket has no way of saying so: it does not know the person. Reviews on reference intervals propose exactly this for such cases: using the person as their own reference when several of their samples are available.[1]
How to compare two reports without fooling yourself
Before concluding that something changed in you, rule out that the draw changed. Certain foods and drinks, some medicines and supplements, hard exercise, your period and not following fasting instructions all move results.[6] The time of the draw relative to the body's rhythms, fasting and physical activity are variables the guidelines ask labs to control when building a range,[1] and it pays to control them the same way when you compare yourself with yourself.
- Same lab and same method. To follow a trend, MedlinePlus recommends using the same lab.[6] If you switched, note the method on each report.[2]
- Same units. Convert both results to the same unit before comparing them, using the conversions in the note above.[8]
- Same time of day. If one draw was at 7 a.m. and the other at noon, the body's rhythm is already a difference.[1]
- Same hours of fasting. The instructions are in fasting before a blood test; if one draw was fasting and the other was not, compare them with caution.[6]
- Same day before. No intense exercise, no alcohol and no oversized meal the day before both draws.[6]
- Same point in your cycle, if you menstruate: your period is among the things that move results,[6] and our guide to the menstrual cycle and lab results covers what changes in each phase.
- The same medicines and supplements, or any changes written down with dates: some interfere with results.[6][7]
- The range for your sex and age, which the report should have printed.[8]
If the list reveals a difference in conditions, the sensible step is to repeat the test under the same conditions before drawing conclusions, within the time frame your doctor suggests. If the conditions were the same and the change exceeds the minimum in the table, the next step is not another test on your own, but bringing the series to an appointment. And with three or more results moving in the same direction, the question your doctor will ask is exactly that: whether your current results differ from the ones you had before.[7]
What the reference range does well
None of this means reference ranges are useless. They are a serious tool, built with method and verified in each lab,[1][2] and they do exactly what they were made for: flag what deserves a closer look in a person the lab knows nothing else about.
And above all: an out-of-range result is not to be ignored. The fact that 5% of healthy people fall outside[7] does not turn your result into a false alarm. A flagged value is a question for your doctor, not an answer and not a verdict: it may or may not be a sign of a problem, and it is read together with your history and symptoms.[6] If you want to shape that question, our guide to health hypotheses explains how to build one to take to an appointment.
Warning signs
Most out-of-range results are discussed at a scheduled appointment. There are two exceptions that do not wait.
When the lab calls
Some reports include critical values: results that may represent a life-threatening abnormality.[7] International recommendations ask the lab to communicate them to the responsible clinician, preferably by voice and generally within an hour of identifying them.[9] If your lab or your doctor tries to reach you about a result, answer and follow their instructions that same day. The lab will not interpret it for you: the Mexican standard asks lab staff not to give opinions on your results,[8] so that conversation is with your doctor.
When there are symptoms
No number replaces what you feel. Call 911 or go to the emergency room, without waiting for any test, if you have:[10]
- Chest pain or discomfort lasting two minutes or more.[10]
- Trouble breathing or shortness of breath.[10]
- Confusion, fainting or loss of consciousness.[10]
- Sudden dizziness, weakness or vision changes, or inability to speak.[10]
- Sudden, severe pain anywhere in the body, or severe abdominal pain.[10]
- Bleeding that will not stop, or vomiting or coughing up blood.[10]
What to bring to your doctor
An appointment about a result is decided in a few minutes. What you bring determines whether the conversation is about one loose number or about your history.
- Every report for the same marker, with dates, not just the latest, and the range each lab printed.[8]
- The lab, method and unit on each report.[2][8]
- The conditions of each draw: time, hours of fasting, exercise, alcohol or a large meal the day before, point in your cycle if it applies.[6]
- Your list of medicines and supplements, with start dates and the date of every change.[6]
- Your calculation of the change as a percentage and the figure from this article's table, as a question: "is this change larger than the expected variation?".[5]
- Symptoms or changes in how you feel since the previous result, with dates.
With that, your doctor can decide whether to repeat the test, check it against other markers or simply keep watching. AI organizes, you understand, you and your doctor decide.
Did your result really change? Seven questions before you compare two reports
Not a diagnosis: it sorts out where your two results come from, under what conditions they were drawn and how much they moved, and leaves you with a list of what to record. If the lab or your doctor contacted you about a critical value, or you have chest pain, shortness of breath, confusion or sudden weakness, do not answer anything: call your doctor or 911 now.
Are the two results you are comparing from the same lab?
This quiz points to what to measure and what to discuss with your doctor. It is not a diagnosis and does not replace an appointment. Your answers never leave your browser.
Frequently asked questions
Why did my results come out different at two labs?
Because each lab uses its own method, equipment and reagents, and builds or verifies its range with a specific population. From the same sample, two methods can give different numbers, especially for hormones. On top of that, your blood naturally varies from day to day and the conditions of the draw are rarely identical. To follow a trend, compare results from the same lab and in the same unit.
What does it mean to be outside the reference range?
It means your result falls outside the central 95% of a group of healthy people. By construction, 1 in 20 healthy people is outside for each marker, so it does not equal being ill. It is not ignored either: it is a question for your doctor, who reads it alongside your earlier results, your symptoms and the conditions of the draw.
How much does a result have to change to matter?
It depends on the marker. With the biological variation from the European EFLM database you can calculate the minimum change between two results from the same person: with a method that has no imprecision, about 14% for glucose, 12% for creatinine and 7.5% for hemoglobin. With the lab's real imprecision the threshold is somewhat higher. A change above that figure, under the same conditions, is worth discussing even if it stays inside the range.
My glucose went from 88 to 97 mg/dL: is that a real change?
On its own, probably not. It is a 10% rise, below the 14% that glucose's biological variation allows between two draws from the same person. Check that both draws were fasting and at the same time, and look at the next result: if it keeps rising, the direction starts to count and is worth discussing with your doctor.
What is a personal norm?
It is the range your own values move within over time, built from your own test history rather than from a population. It takes shape from several results taken under similar conditions and lets you notice changes that matter to you even when the result stays inside the lab range. It does not replace the reference range: the two are read together.
Sources
- Ozarda Y (2016). Reference intervals: current status, recent developments and future considerations. Biochemia Medica 26(1)
- Henny J, Vassault A, Boursier G et al., EFLM Working Group Accreditation and ISO/CEN standards (2016). Recommendation for the review of biological reference intervals in medical laboratories. Clinical Chemistry and Laboratory Medicine
- European Federation of Clinical Chemistry and Laboratory Medicine (EFLM). EFLM Biological Variation Database: meta-analysis results
- Fraser CG (2020). Biological variation: a still maturing aspect of laboratory medicine. Advances in Laboratory Medicine 1(3)
- Fraser CG (2012). Reference change values. Clinical Chemistry and Laboratory Medicine 50(5)
- MedlinePlus, National Library of Medicine. How to Understand Your Lab Results
- Testing.com (2021). Reference Ranges and What They Mean
- Secretaría de Salud (2012). NOM-007-SSA3-2011, Para la organización y funcionamiento de los laboratorios clínicos. Diario Oficial de la Federación, 27 mar 2012
- Lippi G, Mattiuzzi C (2016). Critical laboratory values communication: summary recommendations from available guidelines. Annals of Translational Medicine 4(20)
- MedlinePlus Medical Encyclopedia, National Library of Medicine. Recognizing medical emergencies
Keep reading
Low vitamin D: symptoms, what your result means and which tests to ask for
Low vitamin D symptoms that official sources recognize, what a level of 10, 20 or 30 ng/mL means, who is at higher risk and what to bring to your doctor.
High insulin: what it means, what symptoms it causes and which tests to ask for
What a high fasting insulin level means, why there is no single normal value, what moves the result and which tests are read alongside it.
Hemoglobin A1C (HbA1c) normal range: what 5.7, 6.5 or 7 means
Hemoglobin A1C normal range, what 5.7, 6.2 or 6.5 % means, the average glucose table, what can distort the result and when to repeat the test.