How to read your lab results and keep them in order, year after year

You get the lab report, your eyes hunt for the asterisks or the bold numbers, and you find one. Total cholesterol: 215, range 0–200. Your heart beats a little faster, you open your phone and search "cholesterol 215 dangerous". The scene repeats millions of times a day, and almost every time the same two things are missing: context and history.

This guide does not teach you to diagnose yourself; that is your doctor's job and it is not done from a report. It teaches you to read a lab report without panic, to understand what the ranges mean, why two laboratories give different numbers for the same test and, above all, how to keep your results so that, five years from now, you see a trend and not an isolated number.

The anatomy of a lab report

Every report, from every laboratory, has the same columns, even if arranged differently:

• The test: what was measured. Sometimes the full name ("Glycated haemoglobin"), sometimes the abbreviation ("HbA1c"), sometimes both.
• The result: the number.
• The unit: mg/dL, mmol/L, g/L, ×10⁹/L, IU/L. The same number in different units means completely different things: a glucose of 5.5 is normal if it is in mmol/L and impossibly low if it were in mg/dL.
• The reference range: the values between which most healthy people fall, for that laboratory's method.
• A flag (asterisk, arrow, bold, "H"/"L") when the result is outside the range.
• The method and sometimes the analyser: information you ignore, but which explains why laboratories differ.

The report should also carry the date and time of collection and whether you were fasting. If they are missing, note them yourself; they matter for comparisons.

What the "reference range" actually means

This is the biggest misunderstanding. The reference range is not the line between healthy and ill. It is usually the interval into which 95% of a reference population considered healthy falls, for that laboratory's measurement method. Three consequences:

1. Statistically, 5% of perfectly healthy people have a value "outside the range" on any test. If you have 20 tests done, the chance that one comes back flagged without anything being wrong is over 60%. A lone asterisk on a value slightly outside the range is first of all an invitation to repeat and to add context, not a diagnosis.
2. Ranges differ between laboratories, because methods and analysers differ. A TSH of 4.2 may be flagged at a laboratory with a 0.4–4.0 range and normal at another with 0.3–4.5. Always compare the result with the range on the same report, not with one found online.
3. Ranges depend on age, sex and sometimes pregnancy or time of day. A good laboratory adjusts them; a good report shows the ones that apply to you.

There is a second category, more important than the reference range: clinical decision thresholds. For LDL cholesterol, for example, there is no single "normal"; the target depends on your cardiovascular risk, from below 116 mg/dL for someone without risk factors to below 55 mg/dL after a heart attack, according to European guidelines. For HbA1c, the 6.5% threshold defines diabetes, but 5.7% to 6.4% is a zone of attention. These thresholds are set by your doctor, based on your whole situation, not by the laboratory.

Why results vary from day to day

Even with no change in your health, the same test can give different numbers on two consecutive days. Sources of variation:

• Biological: cholesterol varies by a few percent from day to day; iron and cortisol vary with the hour; glucose, with the previous meal and with stress.
• Pre-analytical: how long you fasted, whether you drank water, whether you exercised beforehand, how long the sample waited before processing, how the tourniquet was applied. High potassium after a difficult draw is a classic.
• Analytical: the laboratory's analyser and reagents have their own tolerance, usually a few percent.

The practical conclusion: a 5–10% difference between two measurements is usually not a real change. A trend over three or four measurements, in the same direction, is. That is why history matters more than a single number.

The common tests, briefly

Not a complete list, but the ones you see most often, with what they measure and why they are ordered. Interpretation remains your doctor's.

Full blood count (FBC / CBC). Red cells (with haemoglobin and haematocrit), white cells (with the differential) and platelets. Low haemoglobin suggests anaemia and calls for context (iron, B12, folate); raised white cells can mean infection, inflammation, or simply recent exercise or smoking.

Glucose and HbA1c. Fasting glucose is a snapshot; HbA1c is a film of the last two to three months, because it measures how much haemoglobin has bound glucose. HbA1c is the test to follow over time for diabetes and prediabetes.

Lipid profile. Total cholesterol, LDL ("bad"), HDL ("good"), triglycerides. LDL is what treatments target; high HDL is protective; triglycerides rise with alcohol, sugar and a recent meal. Interpreted only together with cardiovascular risk.

Kidney function. Creatinine and, more usefully, the estimated glomerular filtration rate (eGFR), calculated from creatinine, age and sex. An eGFR above 90 is normal; below 60, sustained, indicates chronic kidney disease. Urea completes the picture.

Liver function. ALT, AST, GGT, alkaline phosphatase, bilirubin. Raised ALT is the most liver-specific; GGT rises with alcohol and some medications.

Thyroid. TSH is the screening test: high suggests an underactive thyroid, low an overactive one. FT4 and FT3 are ordered when TSH is abnormal. TSH varies with the time of day and with acute illness; it is repeated before any decision.

Inflammation. ESR and C-reactive protein (CRP). CRP rises quickly in infections and acute inflammation; high-sensitivity CRP (hs-CRP) is used for cardiovascular risk.

Vitamins and minerals. Vitamin D (25-OH), B12, ferritin (iron stores), magnesium. Ferritin is also an acute-phase protein, so it can be raised in inflammation even when iron is low.

Urine. Urinalysis (specific gravity, protein, glucose, leukocytes, nitrites, red cells) and, when needed, urine culture. Simple, cheap, often neglected.

How to compare results from different laboratories

This is where standards come in. Two problems always appear:

Same test, different names. "Glucose", "Serum glucose", "Blood sugar", "GLU". The LOINC standard gives each test a unique code (fasting glucose: 1558-6), whatever the name on the report. An app that stores the LOINC code can put results from three laboratories on the same line; one that stores only the text cannot.

Same test, different units. Cholesterol in mg/dL in the US and Romania, in mmol/L in the UK. The conversion factor is fixed (for cholesterol, 38.67; for glucose, 18.02), and the UCUM standard writes units in one way so software can convert them automatically. When you enter a result, enter the unit exactly as on the report; the app does the conversion.

Different ranges. You cannot unify them and you do not need to. Keep, with each result, the range of the laboratory that produced it. For the trend, your value over time matters; for "is it outside the range?", the range at the time matters.

How to keep your results so you see the trend

Five rules, from the most important.

1. Enter values as numbers, not just as a PDF. The PDF is the evidence; the number is what you can compare. At least for the tests you follow: HbA1c, LDL, TSH, creatinine, ferritin, haemoglobin, whatever your doctor said matters for you.
2. Keep the unit and the reference range with each value. Without them, the number is ambiguous a year later.
3. Note the context: fasting or not, time, acute illness, new medication. A CRP of 40 on day three of the flu is not a CRP of 40 for no reason.
4. Put results on the same axis as everything else: medications started, weight, symptoms. An LDL falling three months after starting a statin confirms the treatment works; without the medication line, you only see a number that went down.
5. Link the result to its episode. The tests from the 2023 admission sit next to the 2023 discharge letter. Five years on, the doctor will want to know why they were done, not just what they showed.

Good digital medical record apps do three of these automatically: they extract the values from the laboratory's PDF, code them with LOINC and put them on a timeline alongside medications and symptoms. What stays yours is the context and the decision to enter them the day you receive them.

When to talk to your doctor

Always, when you have a question; this is only a hierarchy of urgency, not medical advice.

• Without delay, when a result is flagged as "critical" or a "panic value" on the report, or when the laboratory calls you. Very high or low potassium, very high or low glucose, very low haemoglobin, very low platelets.
• In the next few days, when a value is clearly outside the range (not by 5%, but by 50%) or when several tests in the same category are abnormal together (ALT, AST and GGT all raised).
• At the next consultation, when a value is slightly outside the range, isolated, without symptoms. Usually the doctor will ask for a repeat in a few weeks.

And one rule for home: do not change a medication yourself on the basis of a result. Neither up nor down.

How Anpheros helps

Anpheros Daily reads the lab report from a PDF or a photo, extracts the values with the laboratory's unit and reference range, codes them with LOINC and UCUM and places them on the record's timeline, alongside medications, symptoms and measurements. Results from different laboratories with different names land on the same curve. Each value stays linked to its source document and to the episode it belongs to, and the whole record exports in FHIR format at any time. The app's AI assistant can explain what a test measures and what the range means, with its sources marked; medical interpretation remains your doctor's.

Frequently asked questions

Does a flagged value mean I am ill?

No, not by itself. The reference range covers 95% of healthy people, so 1 in 20 tests comes back flagged in someone with no problem at all. What matters is how far outside the range it is, whether it repeats, whether it comes with other abnormal values or symptoms, and what the doctor who knows you says.

Why do two laboratories give different results for the same test?

Different methods, analysers and reagents, plus day-to-day biological variation. Differences of a few percent are normal. For fine comparisons (following a value over time, for example), try to use the same laboratory; for the rest, look at the trend, not the decimals.

Do I need to fast for every test?

No. Classically, glucose, the lipid profile and iron were drawn fasting; recent guidelines accept a non-fasting lipid profile for screening. HbA1c, TSH, the blood count and most others do not require fasting. Ask the laboratory or your doctor when in doubt, and note on the report how you came.

How long should I keep old results?

For life, if you can. Values from ten years ago are the starting point of your trend. The longer the history, the earlier a real change shows and the less a normal variation frightens.

Can an AI assistant interpret my results?

It can explain what a test measures, what the range means and why it might vary; that is useful for understanding the report and preparing your questions. It cannot replace your doctor, because interpretation needs your whole situation: history, medications, symptoms, clinical examination. Use it to understand, not to decide.