How to read your lab results

Every lab report, no matter which hospital or lab produced it, is built from the same five columns. Learn them once and the formatting stops being intimidating. The first column is the test name, sometimes a plain word like 'Potassium,' sometimes an abbreviation like 'WBC' or 'eGFR,' sometimes a long chemical name. The second is your result, the actual measured number. The third is the unit (mg/dL, mmol/L, g/dL, cells per microliter) which is just the scale the number is measured on and which you can mostly ignore as long as you are comparing like with like.

The fourth column is the reference range, the band of values the lab considers typical for a healthy reference group. It is the single most misunderstood part of the report, and the next step is entirely about reading it correctly. The fifth, when it appears, is a flag column, usually a letter in the margin. 'H' means the result is above the reference range, 'L' means below, and some labs add a stronger marker (an asterisk, 'HH'/'LL,' or the word 'critical') for a value far enough out to need urgent attention. A blank flag column means the value landed inside the range.

Two more details are worth finding before you read anything else. One is the collection date and time, usually near the top. It tells you when the sample was drawn, which matters when you are tracking a value over weeks. The other is the name and address of the performing lab, often in small print at the bottom. That detail seems boring until you are comparing two reports and discover the reference ranges do not match, at which point the lab name is the explanation.

If the report runs several pages, the tests are usually grouped into panels: the complete blood count on one block, the metabolic panel on another, and so on. You do not have to read it top to bottom. Find the panel you came to understand, or the line with the flag next to it, and start there. The rest of this guide is organized the same way the report is.

Here is the idea that defuses most lab-result anxiety: a reference range is not a healthy-versus-unhealthy line. It is the band where the middle 95 percent of a reference population falls. By definition, that leaves a healthy person in twenty above or below it on any given test, perfectly well. A result a hair outside the range is common, usually unremarkable, and almost never an emergency on its own. The flag in the margin is a prompt for a human to look, not a diagnosis.

The ranges also vary from lab to lab, and this trips up nearly everyone tracking a value over time. Different labs run different machines and methods, so each one publishes its own reference band. MedlinePlus puts it bluntly: you cannot compare test results from different labs. If your father's potassium was drawn at the hospital in March and at the clinic in June, the two 'normal' columns may not match, and the safest way to track a trend is to use the same lab each time. When you cannot, compare each result against the range printed on its own report, never against the other report's range.

This is also why a medical record organizer that keeps each result with its own reference range beats a shoebox of printouts. The range is part of the result; separated from it, a bare number is hard to interpret and easy to misread. Keep the two together and the trend becomes legible. Pull them apart and you are guessing.

One more distinction saves a lot of confusion: a reference range is not the same as a healthy target. Cholesterol is the clearest example. A lab may flag your LDL as 'in range' against its broad reference band while your doctor is aiming for a much lower target because of your heart history. The range answers 'is this typical for the population?' The target answers 'is this where we want it for you?' Those are different questions, and for a few tests (cholesterol, A1C, blood pressure) the target your care team sets matters more than the range the lab prints.

The complete blood count is the panel you will see most often, and it counts the three cell types in your blood. Reading it is mostly a matter of knowing what each line is for. Note before you start: the ranges below are the typical adult values MedlinePlus publishes, and your lab's printed range is the one that governs. Normal bands shift with the lab, and with your sex, your age, and pregnancy.

White blood cells (WBC) are the immune system's count, and MedlinePlus puts the usual adult band at 4,500 to 11,000 cells per microliter. A count above the range can follow an infection, inflammatory disease, tissue damage, pregnancy, or stress. A count below it can follow a bone marrow problem, certain medicines, an autoimmune disorder, a severe infection, radiation therapy, or a viral illness. For anyone in cancer treatment, the WBC and its 'neutrophil' subset are watched closely, because there are points in a chemotherapy cycle when neutrophils run particularly low and the risk of infection goes up. If the differential also flags a line called immature granulocytes, that has its own plain-English decoder, and so does a flagged absolute lymphocyte count.

Red blood cells carry oxygen, and the two lines to read are hemoglobin and hematocrit. MedlinePlus gives the usual adult hemoglobin band as 12.1 to 15.1 g/dL for women and 13.8 to 17.2 for men; hematocrit, the percentage of blood made of red cells, tracks alongside it. Low hemoglobin is anemia, which has many causes and can leave someone tired, weak, short of breath, or lightheaded. A low hemoglobin alongside new symptoms is worth a call rather than a wait. The helper lines nearby, like the MCV (average cell size) and the RDW (how much the sizes vary), exist to explain an anemia rather than stand alone, and the RDW has its own plain-English decoder. Two more helper lines sit just below them, MCHC and MCH, which are not measured at all but calculated from the hemoglobin, hematocrit and cell count, which is why a mildly low value with everything else in range is so often arithmetic rather than anemia.

Platelets are what let blood clot, and the usual adult count is 150,000 to 400,000 per microliter, with bleeding risk rising below 50,000. Low platelets raise bleeding and bruising risk and, like the white count, are watched during chemotherapy. High platelets can follow inflammation or other conditions. As with every line on the CBC, a single value a little outside the band is rarely the story; the pattern across the whole panel, and across time, is what the care team actually reads.

The comprehensive metabolic panel (CMP) measures fourteen substances; the basic metabolic panel (BMP) covers eight of those fourteen and leaves out total protein, albumin, bilirubin, and the liver enzymes. It is a snapshot of how the kidneys, liver, blood sugar, and electrolyte balance are doing. You do not need to read all fourteen lines. Knowing the four groups is enough to find the one that matters.

Blood sugar is the glucose line. Drawn fasting, the CDC's thresholds are 99 mg/dL or below as normal, 100 to 125 as prediabetes, and 126 or above as diabetes. A single high glucose on a non-fasting draw means much less than a fasting one, which is why the report often notes whether you had eaten, context that changes the meaning of the same number.

Kidney function shows up as BUN, creatinine, and eGFR. Creatinine is a waste product from the normal breakdown of muscle that the kidneys clear out of your blood; eGFR estimates how well they are filtering, and the NIDDK puts 60 or more in the normal range, with less than 60 possibly meaning kidney disease. These three move together, and a drift over time matters more than one reading. If the report also flags a BUN/creatinine ratio, that line is calculated from the other two rather than measured, and it has its own plain-English decoder.

Electrolytes (sodium, potassium, chloride, and bicarbonate) are the body's chemistry balance, and potassium is the one to respect. Its normal band is narrow, about 3.7 to 5.2 mmol/L, and potassium that runs too high or too low can bring on an irregular heart rhythm, which is why a value well outside that small window is one of the few lab results that can genuinely be a call-today situation rather than an ask-next-visit one. Sodium typically runs 135 to 145 mEq/L. The liver group (ALT, AST, alkaline phosphatase, bilirubin) plus calcium, total protein, and albumin round out the panel; mild ALT and AST elevations show up in roughly 10 to 20 percent of the general population and often settle on a recheck, but a marked rise is something the care team will want to look at.

Three more panels show up at almost every routine visit, and each rewards a slightly different way of reading. The lipid panel (cholesterol) is the one where the target matters more than the range. The CDC's optimal levels are about 150 mg/dL for total cholesterol, about 100 for LDL (the 'bad' cholesterol), at least 40 in men and 50 in women for HDL (the 'good' kind), and under 150 for triglycerides, with a total above 200 counting as high. But your doctor may aim well below those numbers if you have heart disease or diabetes, so read your LDL against the target your care team set, not just the lab's range.

A1C is the diabetes screen, and it is the best example in the whole report of why the trend beats the snapshot. A1C reflects your average blood sugar over the past two or three months, so it moves slowly and meaningfully. The CDC's thresholds are below 5.7 percent for normal, 5.7 to 6.4 percent for prediabetes, and 6.5 percent or above for diabetes. But the more useful reading is direction: an A1C that went from 7.2 in the winter to 7.8 in the spring is heading the wrong way even though both numbers sit in the same category, and that drift is the thing worth raising at the next appointment.

Thyroid screening usually starts with TSH, and MedlinePlus gives typical values as 0.4 to 4.8 microunits per milliliter, with labs varying. TSH works backward from what you might expect: a high TSH usually points to an underactive thyroid (the pituitary is shouting at a sluggish gland), and a low TSH usually points to an overactive one. If TSH is off, the care team often adds free T4 and other tests before drawing conclusions, so a single out-of-range TSH is a reason to look further, not a diagnosis.

A urinalysis often rides along with these, and it is read differently again, because most of its lines are color pads rather than measured numbers. The two that cause the most worry are leukocyte esterase and nitrite, and they are meant to be read as a pair; a flagged leukocyte esterase can come from an infection or from contamination during collection, and it has its own plain-English decoder. The microscopy section below the pads carries the same theme: a count of squamous epithelial cells describes how the sample reached the cup rather than anything about the kidneys or bladder. A line reading hyaline casts works the same way: it is the least specific finding in the sediment, and what it means is decided by the protein, blood, and cast-type lines printed beside it. The chemistry pads carry the same lesson about scale, and urobilinogen is the clearest example, since a printed 0.2 and a printed normal are the same statement and the pad cannot read below that floor at all.

Across all three, the pattern from the rest of this guide holds: read the value against the right yardstick (range or target), pay attention to direction over time, and remember that one panel is a frame from a film, not the whole story. A result that worries you is a reason to ask, not a reason to conclude.

This is the step the commodity guides skip, and it is the one families actually need. A flagged result is not a single category of bad news; it sorts into three. Knowing which bucket you are in is the difference between a sleepless night and a note on the fridge to mention something next week. Read this as orientation, not medical advice. When in doubt, the call is always reasonable, and your care team's read is the one that counts.

Call today (or use the after-hours line) when a value is flagged with a stronger marker than a plain H or L (an asterisk, a double letter, or the literal word 'critical') or when it is a value that affects the heart or oxygen and is far outside its band: a potassium well outside its narrow window, a markedly low hemoglobin in someone newly short of breath, a very high blood sugar with symptoms, a platelet count low enough to risk bleeding. Federal rules require the lab to immediately alert whoever ordered the test when a result points to an imminently life-threatening condition. That alert goes to the ordering provider, not to you, so a quiet phone is not the same as reassurance. If you are not sure whether a value qualifies, that uncertainty is itself a good reason to call the nurse line and ask.

Ask at the next visit covers the large middle: a value modestly outside the range with no critical marker, a liver enzyme up a little, an A1C that crept up, a cholesterol that moved the wrong way. These are real and worth raising, but they are conversations, not emergencies. Write the question down ('my A1C went from 7.2 to 7.8, what should we do') and bring it. A specific question gets a specific answer; a folder of numbers gets a shrug.

Probably nothing is the most common bucket and the hardest to sit with. A single value a hair outside the band on a one-off draw, with no symptoms and no pattern, is usually noise: the same person retested next week often lands back inside. This is exactly where the temptation to search the number and spiral is strongest, and where it helps most to have something that has read the whole report in context. KeptWell's AI reads every result against its own reference range and explains what a flag likely means in plain English, with the source paragraph cited so you can take it back to the care team, and, just as important, it does not diagnose. It orients. The judgment stays with your clinician.

The most useful skill in reading lab results is the one a single report cannot teach: watching a value over time. A creatinine of 1.3 means one thing if it has been 1.3 for years and something else entirely if it was 0.9 last spring. An A1C is its direction. A tumor marker is its slope. The care team reads trends as a matter of course because they have the prior results in front of them; the family usually does not, because the priors are scattered across three portals and a kitchen drawer. Closing that gap is most of the value of keeping your own copy.

The practical move is to keep every result in one place, each with its date and its own reference range, so the trend is visible without assembling it by hand every time. A digital medical binder that reads each report as it arrives turns a pile of PDFs into a timeline: the same test plotted across visits, the drift made obvious, the context preserved. The upload-to-understand workflow is the same whether the result came from a portal export, a faxed printout, or a photo of the page the lab handed you on the way out.

Bringing the right question is the other half. A care visit is short, and the difference between a useful one and a wasted one is often a single well-formed question prepared in advance. Reading the report ahead of time, sorting the flags into the three buckets and noting the one value that moved, lets you walk in with 'my potassium has been climbing the last two draws, should we be watching it' instead of handing over a folder and hoping. If you want to check one flagged value before you get there, our free lab values explainer shows where a number sits against the typical range and what high or low usually means. The literacy this guide builds is in service of that one sentence.

A last word on the failsafe. Any tool that summarizes a medical result, including a good one, should show its work. A summary you cannot trace back to the source line is a guess dressed up as an answer, and a guess is the last thing a family in a hard stretch needs more of. Insist on the citation, check the summary against the actual number on the report, and treat the source line as the truth and the summary as the convenience. That habit is what makes reading your own results safe to do.