The same blood test, written a dozen ways: lab reports compared
HealthViewer reads lab reports from many labs, and every new layout has to be taught to the parser. That leaves us with something few people have: the same tests, as printed by labs that have never coordinated with each other. We counted what they print. Across 50 reports in 36 layouts from 8 countries, the unit for a white blood cell count is written 13 different ways, and the “normal” range for the same test changes from one report to the next.
The numbers on this page are generated from our parser’s test files and were last updated on November 9, 2026. The methodology and the limits of this data are below, and the underlying rows are a CSV you can check.
One unit, many spellings
Here is every way these reports write the unit for a white blood cell count:
x10^9/L /cmm 10*3/uL Thousand/uL x10E3/uL 10^3/μL 1000/uL k/cumm K/ul K/uL thou/mm3 X 10³ / μL x10(9)/L
Almost all of these are the same unit. A thousand cells per microliter is the same
amount as a billion cells per liter, so K/uL, x10E3/uL, Thousand/uL, and
x10^9/L all go with the same number. The exception is the per-microliter form
(/cmm, common on Indian reports), where the same count is printed a thousand times
larger: 7,500 instead of 7.5.
The pattern holds for the other tests that appear in at least 10 of the reports:
| Test | Reports | Unit spellings | As printed |
|---|---|---|---|
| Red blood cells | 21 | 14 | x10^12/L million/cmm Million/uL x10E6/uL 10*6/uL 10^6/uL m/cumm M/ul M/uL mil/cumm mili/cu.mm mill/mm3 X 10^6/μL x10(12)/L |
| White blood cells | 23 | 13 | x10^9/L /cmm 10*3/uL Thousand/uL x10E3/uL 10^3/μL 1000/uL k/cumm K/ul K/uL thou/mm3 X 10³ / μL x10(9)/L |
| Platelets | 23 | 13 | x10^9/L /cmm 10*3/uL Thousand/uL x10E3/uL 10^3/uL 10^3/μL 10E3/μL k/ul K/uL thou/mm3 X 10³ / μL x10(9)/L |
| Neutrophils (absolute) | 19 | 12 | x10^9/L /cmm 10*3/uL cells/uL 10^3/μL 1000/uL k/cumm K/ul K/uL thou/mm3 X 10³ / μL x10E3/uL |
| TSH | 14 | 7 | uIU/mL mIU/L μU/mL microIU/mL mU/l u[IU]/mL uIU/ml |
| Hemoglobin | 23 | 6 | g/dL g/dl g/L gm/dL gm% mmol/l |
| Creatinine | 15 | 5 | mg/dL mg/dl mg/L umol/L μmol/l |
For a person reading one report, none of this matters. It matters when you put two
reports side by side, type results into a spreadsheet, or paste them into a tool that
has to decide whether 10^3/μL and K/uL belong on the same chart. Some of the
differences are only spelling. Others change the number: hemoglobin in g/dL, g/L, and
mmol/L gives three different values for the same blood. The
lab unit converter handles the pairs that need arithmetic.
“Normal” depends on the report
For the comparison of ranges we used tests where most labs print a single adult range, so that a difference between two reports mostly reflects the lab and not the patient’s sex. Ranges printed in another unit were converted first, and the same range written two ways (3.5–5.1 and 3.50–5.10) was counted once.
| Test | Reports | Different ranges | Lower limit runs from | Upper limit runs from |
|---|---|---|---|---|
| TSH | 14 | 10 | 0.27 to 0.5 uIU/mL | 3 to 5.33 uIU/mL |
| Potassium | 13 | 6 | 3.3 to 3.7 mmol/L | 5 to 5.4 mmol/L |
| Sodium | 13 | 6 | 134 to 136 mmol/L | 143 to 146 mmol/L |
| Chloride | 11 | 6 | 96 to 101 mmol/L | 106 to 115 mmol/L |
| Calcium | 12 | 8 | 8.4 to 8.8 mg/dL | 10 to 10.5 mg/dL |
| Glucose | 13 | 9 | 60 to 74 mg/dL | 99 to 118 mg/dL |
| White blood cells | 22 | 12 | 3.2 to 4.1 x10^9/L | 8.15 to 11.4 x10^9/L |
| MCV (red cell size) | 23 | 13 | 78 to 86.1 fL | 96 to 102 fL |
Take TSH. In 14 reports there are 10 different ranges, and the top of the range runs from 3 to 5.33 uIU/mL. A result between those two numbers is flagged high on some of these reports and printed without comment on others.
None of these labs is wrong. A reference range describes the lab’s own method and the people it was checked against, and U.S. labs are required to set or verify their own. Why are lab reference ranges different between labs? explains how that works and which tests it affects most. What the table adds is how routine the differences are: every test we compared had at least 6 different ranges across these reports.
One test, many names
The parser keeps a list of the names it has to recognize for each test. It currently maps 534 printed names to 181 tests. Most tests have only two or three. The ones ordered most often have many more:
| Test | Names | Examples |
|---|---|---|
| hsCRP | 14 | c reactive protein high sensitivity, c reactive protein hs, c-reactive protein high sensitivity, c-reactive protein hs, crp high sensitivity |
| WBC | 14 | leucociti, leucocytes, leukocytes, leukozyten, total leucocyte count |
| eGFR | 13 | debit de filtration glomerulaire, debit de filtration glomerulaire selon ckd-epi, egfr, egfr crea nach ckd-epi, egfr non-afr american |
| LDL Cholesterol | 12 | cholesterol, ldl, calculated, colesterol ldl, colesterolo ldl, direct ldl, ldl |
| RBC | 12 | eritrociti, erythrocytes, erythrozyten, hematies, rbc |
| Glucose | 11 | blood glucose, fasting blood glucose, fasting blood sugar, fasting glucose, glucosa |
| Vitamin D | 11 | 25(oh) vitamin d, 25(oh)d, 25-hydroxy d total, 25-hydroxyvitamin d, 25-oh-vitamin d |
| Hematocrit | 10 | ematocrito, haematocrit, hamatokrit, hct, hematocrit |
Some of these are abbreviations, some are other languages (French, German, Italian, and Spanish reports are in the set), and some name a variant that a specialist would keep apart, such as a directly measured LDL next to a calculated one. They are listed together because a person following their own results over the years wants them on one line.
Why this matters if you keep your own records
Lab results now reach patients directly, often as a PDF from a portal, and many people see more than one lab over the years. Putting those results in one place means settling, for every test, which names are the same test, which units are the same unit, and which range a value should be judged against. A spreadsheet leaves all three to you. An AI chatbot asked to merge two reports has to make the same three decisions, and you would need to check each one.
This is the job we built HealthViewer for. It is a lab results tracker that runs on your own computer, reads the report, and keeps each result with the unit and range its lab printed, converting only when it draws a chart. How to track your lab results over time walks through it.
How we counted
- The reports. 50 lab reports in 36 distinct layouts, from labs in Canada, Colombia, France, Germany, Ghana, India, Italy, and United States. They are the files HealthViewer’s parsers are tested against. Most were sent in by users through the app’s “Help us support this lab” feature, which removes names, dates of birth, and ID numbers and scrambles the result values on the user’s computer before anything is sent. This page uses only what belongs to the lab: test names, units, and reference ranges. No result values are used.
- Units. For each test found in at least 10 reports, we listed every distinct spelling of its unit, exactly as printed. Where a report prints a test more than once, the first is used.
- Ranges. Only ranges with both a lower and an upper limit are compared. Each was converted to the unit most reports use for that test. A range was left out if its unit could not be converted, or if either limit was more than 1.6 times away from the middle value for that test, which marks a range that was scrambled along with the values or misread. 2 ranges were left out this way; they are marked in the CSV.
- Names. Counted from the parser’s list of names for each test, ignoring capitalization. Home measurements such as blood pressure are not included.
What this does not show
- It is not a sample of all labs. These are the reports our users happened to send, weighted toward the United States. The counts show that the differences exist and are easy to find. They do not estimate how common each one is.
- Ranges can depend on the patient. We chose tests that most labs give one adult range for, but some labs adjust even these by age or sex, and the reports come from different people. Part of the spread in the range table may be that.
- Glucose ranges mix fasting and non-fasting reports, and some labs print a guideline cutoff where others print a range from their own population.
- The name list is what we have had to support, not a census. It grows whenever a new report arrives.
Citing this page
You are welcome to quote these figures with a link to this page. Because the numbers are regenerated as new report layouts arrive, please include the date:
HealthViewer, “Lab report units and ranges compared,” data as of November 9, 2026. https://healthviewer.net/blog/lab-report-units-and-ranges-compared/
Questions about the data, or a report format we should add: support@healthviewer.net.