A blood glucose unit is the measuring scale a lab or meter uses to express how much sugar is circulating in your bloodstream. Two units dominate worldwide: milligrams per deciliter (mg/dL) and millimoles per liter (mmol/L). They report the same thing in different languages of measurement, and converting between them requires only one number: 18. Which one you encounter depends mostly on what country you live in, but misreading one for the other can lead to genuine confusion or even dangerous dosing errors.
What Each Unit Actually Measures
The difference comes down to whether you are describing glucose by weight or by molecular count. A milligram per deciliter (mg/dL) tells you the mass of glucose in a given volume of blood: how many milligrams of sugar sit in one-tenth of a liter. A millimole per liter (mmol/L) tells you the number of glucose molecules in a full liter, expressed as a fraction of a mole, the chemist’s standard counting unit for molecules. Both are perfectly valid ways to describe concentration, just as you could describe the density of traffic by “cars per mile” or “cars per kilometer.”
The reason the conversion factor is 18 has to do with the molecular weight of glucose. One mole of glucose weighs about 180 grams, and the volume difference between a deciliter and a liter is a factor of 10. Combined, those two facts produce a neat divisor. To go from mg/dL to mmol/L, you divide by 18. To go from mmol/L to mg/dL, you multiply by 18. So a fasting reading of 90 mg/dL is the same as 5.0 mmol/L, and a reading of 126 mg/dL equals 7.0 mmol/L.
Which Countries Use Which Unit
The United States reports blood glucose in mg/dL. So do a handful of other countries, including Japan, India, and several nations across the Middle East and parts of Latin America. Most of the rest of the world, including the United Kingdom, Canada, Australia, and nearly all of Europe, uses mmol/L. The split is not random. It tracks the broader and still-incomplete adoption of the Système International d’Unités (SI), the metric framework that standardized scientific measurements globally starting in the 1960s and 1970s.
American medical journals have been gradually shifting toward selected SI units, but reaching full consensus in the U.S. medical community on how far and how fast to convert clinical lab results has proven politically and logistically difficult. The debate has stretched over decades with no final resolution, which is why American labs, glucometers, and clinical guidelines still default to mg/dL while most international bodies use mmol/L.1PubMed Central. The international system of units in historical perspective
For a person with diabetes, the practical consequence is straightforward: if you buy a glucose meter overseas, or read a research paper from another country, or download a health app built for a different market, the numbers may look alarmingly different until you realize which unit is being used. A reading of 5.5 in mmol/L is perfectly normal. A reading of 5.5 in mg/dL would suggest you are nearly dead from hypoglycemia. Context matters.
Common Glucose Thresholds in Both Units
Because so much of diabetes care revolves around specific cutoff numbers, it helps to see the major thresholds side by side. These reflect widely used clinical guidelines:
- Normal fasting glucose: below 100 mg/dL (5.6 mmol/L)
- Prediabetes range: 100 to 125 mg/dL (5.6 to 6.9 mmol/L)
- Diabetes diagnosis: 126 mg/dL (7.0 mmol/L) or higher on two separate tests
- Post-meal target (general): below 140 mg/dL (7.8 mmol/L) at two hours
- Hypoglycemia alert: below 70 mg/dL (3.9 mmol/L)
These numbers appear constantly in patient education materials, lab reports, and continuous glucose monitor (CGM) apps. Knowing the pair in both units saves you from needing to convert every time you read an international study or travel with your meter.
Why Your Result Is Reported as Plasma Glucose
You might notice that lab results and modern meters say “plasma glucose” rather than “blood glucose.” That is not an accident. The international standard, set by the International Federation of Clinical Chemistry (IFCC), recommends reporting glucose concentration in plasma with the unit mmol/L, regardless of what type of sample was actually drawn or what technology measured it.2PubMed. Approved IFCC recommendation on reporting results for blood glucose The rationale is consistency: if every device reports in the same “language,” results from different labs and meters become comparable.
When a fingerstick meter actually measures glucose in whole blood (which includes red blood cells), the device applies a built-in correction to convert the result to an equivalent plasma value. The IFCC-recommended correction factor is 1.11: multiply the whole-blood glucose concentration by 1.11 to get the plasma-equivalent number.3PubMed. IFCC recommendation on reporting results for blood glucose That roughly 11% bump exists because red blood cells contain less glucose than the surrounding plasma, so whole blood reads lower than plasma alone. Most consumer meters handle this conversion internally, so the number on your screen already reflects plasma-equivalent glucose.
Where things get more complicated is the difference between venous and capillary blood. A sample drawn from a vein at the lab and a fingerstick from a capillary do not always agree. Research shows that venous plasma glucose tends to run higher than capillary blood glucose for fasting and random samples, but lower for samples taken two hours after drinking a glucose solution, the kind used in oral glucose tolerance tests.4PubMed. Comparability of venous and capillary glucose measurements in blood The discrepancy is small enough that it rarely changes a diagnosis on its own, but it does mean your fingerstick meter and your lab’s venous draw may not produce identical numbers even when both are “correct.”
How Lab Methods Can Shift the Number
Even after settling on units and sample type, the laboratory technique used to measure glucose can introduce variation. Two enzymatic methods dominate clinical labs. The hexokinase (HK) method is considered more specific. The glucose oxidase (GOD) method is widely used and cheaper but can be thrown off by other substances in the blood. Compounds like uric acid, ascorbic acid (vitamin C), bilirubin, and hemoglobin can cause falsely low glucose readings when the glucose oxidase method is used.5Scientific Reports. The impact of differences in plasma glucose between glucose oxidase and hexokinase methods on estimated gestational diabetes mellitus prevalence
For most people, this difference is clinically negligible. But it becomes meaningful in specific situations, such as screening for gestational diabetes, where the diagnostic thresholds are tight and a small shift in the measured value can change whether someone is classified as having the condition. If you have ever wondered why a glucose result from one lab does not perfectly match a result from another, the measurement method is often part of the answer, alongside timing, fasting status, and sample handling.
HbA1c Has Its Own Unit Confusion
Blood glucose is not the only diabetes-related number that suffers from competing units. HbA1c, the test that reflects your average blood sugar over roughly two to three months, has gone through its own standardization drama. For decades, HbA1c was reported solely as a percentage: a result of 7.0% meant that about 7% of your hemoglobin had glucose attached to it. That system, anchored by the National Glycohemoglobin Standardization Program (NGSP), is still widely used in the United States and many other countries.
In 2007, an international agreement introduced a second reporting format: millimoles of glycated hemoglobin per mole of total hemoglobin, abbreviated as mmol/mol. This IFCC-derived unit was designed to be more chemically precise and internationally comparable.6PubMed Central. HbA1c standardisation: history, science and politics Under the new system, an HbA1c of 7% translates to roughly 53 mmol/mol.7The British Journal of Diabetes & Vascular Disease. HbA1c – changing units
Countries like the U.K. and Australia have largely switched to mmol/mol for HbA1c, though many still include the percentage equivalent in parentheses during the transition period. If you see an HbA1c value of 48 and wonder whether your diabetes has gone wildly off the rails, check the units: 48 mmol/mol is equivalent to about 6.5%, which is actually right at the diagnostic threshold for diabetes and perfectly within many treatment targets. The confusion is understandable, and it catches even experienced clinicians who trained under the percentage system.
One additional wrinkle: HbA1c can be translated into an estimated average glucose (eAG) expressed in either mg/dL or mmol/L. This gives patients a way to connect their two- to three-month average to the numbers they see on their daily meter. The relationship holds well in non-pregnant adults, but in pregnancy the correlation shifts. Research on pregnant women with diabetes found that the same HbA1c level corresponds to a lower average glucose than it would in a non-pregnant person, likely due to changes in red blood cell turnover and other physiological shifts during pregnancy.8PubMed Central. Translating HbA1c measurements into estimated average glucose values in pregnant women with diabetes An HbA1c of 6.0% (42 mmol/mol) in pregnancy corresponded to an average glucose of about 6.4 to 6.7 mmol/L depending on the stage of pregnancy, rather than the roughly 7.0 mmol/L that standard eAG tables would predict.
When Unit Confusion Gets Dangerous
Most of the time, mixing up mg/dL and mmol/L is an inconvenience that a quick mental calculation can fix. Occasionally, though, the stakes are higher. Insulin dosing is the clearest example. If someone traveling abroad checks their glucose on a locally purchased meter that reads in mmol/L and interprets the number as if it were mg/dL, they might think their sugar is dangerously low and skip an insulin dose, or think it is dangerously high and take far too much. A reading of 8.0 mmol/L (a mildly elevated 144 mg/dL) could be mistaken for 8 mg/dL, which would be incompatible with consciousness. The reverse is equally hazardous: reading 144 from a mg/dL meter and thinking the unit is mmol/L would suggest a glucose level so extreme it borders on a medical emergency.
Some modern meters and CGMs let you toggle between units in the settings menu, and a few will even detect your location and default to the local standard. But not all do. Before using any new device, check which unit it displays. This single step prevents a category of errors that is entirely avoidable.
Unit awareness also matters in recognizing atypical emergencies. Euglycemic diabetic ketoacidosis, a rare but serious complication, is defined as ketoacidosis occurring with blood glucose below 250 mg/dL.9PubMed Central. Metabolic ketoacidosis with normal blood glucose: A rare complication of sodium–glucose cotransporter 2 inhibitors In mmol/L, that threshold is about 13.9. A patient or clinician accustomed to mmol/L who sees a glucose of 12.0 and assumes ketoacidosis requires much higher numbers could miss the diagnosis. The condition is already tricky because the blood sugar is not sky-high, and unit confusion only compounds the risk of underreaction.
Glucose Meters and International Standards for Accuracy
Regardless of which unit a glucose meter displays, all consumer devices are held to accuracy standards. The ISO 15197 standard, most recently updated in 2013, requires that 95% of meter readings fall within 15 mg/dL of the lab reference value when glucose is below 100 mg/dL, and within 15% of the lab value when glucose is 100 mg/dL or above. In mmol/L terms, that lower threshold is about 0.83 mmol/L. These margins mean a meter can be considered accurate even when its reading diverges noticeably from a lab draw taken at the same moment.
Continuous glucose monitors add another layer of variability. CGMs measure glucose in interstitial fluid (the liquid between your cells), not in blood directly. Interstitial glucose lags behind blood glucose by several minutes, and the two do not always move in lockstep during rapid rises or falls. CGM manufacturers calibrate their algorithms to report values as close to plasma-equivalent blood glucose as possible, but the underlying measurement is fundamentally different. The unit on the screen (mg/dL or mmol/L) is the same as on a fingerstick meter; the path to that number is not.
Blood Glucose Units in Veterinary Medicine
If you have a diabetic pet, you have encountered glucose units in a different context. Veterinary clinics use the same mg/dL and mmol/L scales as human medicine, and the choice typically follows the same geographic lines: U.S. veterinary labs report in mg/dL, and clinics in the U.K. and Europe use mmol/L. However, normal glucose ranges differ between species. A healthy fasting dog typically has a blood glucose between about 74 and 143 mg/dL (roughly 4.1 to 7.9 mmol/L), which overlaps with but extends higher than the normal human range.
Measurement logistics add complexity. A study comparing glucose readings from a veterinary point-of-care glucometer with a biochemical analyzer found mean differences between whole blood, plasma, and serum glucose concentrations of about 1.0, 6.3, and 6.7 mmol/L respectively, with the point-of-care meter tending to read higher.10PubMed Central. Comparison of glucose concentrations in canine whole blood, plasma, and serum measured with a veterinary point-of-care glucometer Those are substantial differences that can influence treatment decisions in a diabetic animal. Veterinarians are generally aware of this issue and interpret point-of-care readings with the device’s known bias in mind, but it underscores the broader point that a glucose number never stands alone. The unit, the sample type, and the device all shape the result.
Why a Universal Standard Has Proved Elusive
Given all the confusion, you might wonder why the world does not just pick one unit and stick with it. The IFCC has tried. Its formal recommendation is clear: report glucose as plasma concentration in mmol/L, period.2PubMed. Approved IFCC recommendation on reporting results for blood glucose In practice, adoption has been uneven. The United States, the largest single healthcare market, has an enormous installed base of clinical systems, educational materials, patient records, and guidelines expressed in mg/dL. Converting all of that carries real costs and introduces transition-period risks where clinicians and patients must think in two systems at once.
There is also a psychological dimension. Patients who have managed their diabetes for years develop an intuitive feel for their numbers. Someone who knows that 120 is fine and 250 is trouble does not want to suddenly navigate a world where 6.7 is fine and 13.9 is trouble. The underlying biology is unchanged, but the cognitive remapping takes time, and during that time, mistakes happen. Countries that have switched units for HbA1c reporting have documented transition-period confusion among both patients and healthcare workers, lending weight to the argument that unit changes, even when scientifically justified, carry practical risks that have to be managed carefully.
For now, the dual system persists. If you manage diabetes or work with someone who does, the most useful skill is not advocating for one unit over the other. It is recognizing both instantly, knowing the conversion factor of 18, and always confirming which unit any new device, lab report, or research paper is using before acting on the number.