A home blood glucose meter doesn’t have to match the lab exactly. The FDA’s guidance for meters sold over the counter asks that 95% of readings land within 15% of a lab comparison, and 99% within 20%. So if your blood sugar is truly 100 mg/dL, a meter that meets that bar will usually show something between 85 and 115. And not every meter on the shelf keeps to it. In a 2018 test of 18 popular meters, only 6 hit their accuracy target every time.
I like knowing the wiggle room on any number a device hands me. With a glucose meter it matters more than most, because people make real decisions off the display. A reading that’s 10% off on a good day still comes from a good meter. A reading that’s off because of orange juice on your fingertip is a different story, and that one you can fix.
Before we start: despite the site’s name, I’m not a pharmacist, a nurse or a doctor. I can walk you through how accurate these meters are supposed to be and what throws the reading off. What your numbers mean for you, your normal range and any change to your treatment are your doctor’s call. Please don’t change a medication because of something you read here.
Here’s what I’ll cover:
- what “accurate” means for a glucose meter, by the FDA’s numbers
- what that error band looks like on the display
- how 18 meters did when researchers bought them off the shelf
- why washing your hands beats the alcohol wipe
- squeezing, cold hands and the second drop
- strip storage and the fine print in the insert
- why your meter and the lab disagree
- when to call your doctor, and when to call 911
What “Accurate” Means for a Glucose Meter
No meter reads your exact blood sugar every time. What there is instead is a standard. In the U.S., the FDA’s guidance for meters people buy and use at home puts it in plain numbers. In the maker’s own studies, 95% of the meter’s readings should be within 15% of a lab comparison, and 99% within 20%. That has to hold across the meter’s whole measuring range, not just at the comfortable numbers in the middle.
The same guidance says a meter should measure accurately from at least 50 to 400 mg/dL. Outside its range, it isn’t supposed to guess. It should show an error code instead, like a “LOW” message under 50. If your meter can show LO or HI instead of a number, the manual says what that means for your model, and it’s worth reading before the day you see it.
One more detail from the guidance: meters sold in the U.S. should report their results as “plasma equivalents.” In plain English, the meter tests a drop of blood from your finger but reports the number in the same terms a lab uses for its result. That’s why you can compare the two at all.
What That Error Band Looks Like on the Display
Percentages are hard to feel, so here’s the 15% band worked out on real numbers. At a true 100 mg/dL, any reading from 85 to 115 is inside it. At a true 70, it’s about 60 to 80. At 150, about 128 to 172. At 200, it’s 170 to 230, and at 300 it’s 255 to 345.
The 20% line is the outer fence. At a true 100, nearly every reading should fall between 80 and 120. For any result further off than 20%, the FDA asks the maker to explain why it happened and why it isn’t a safety risk.
Two things follow from this. First, two readings a few minutes apart that differ by 8 or 10 points can both be “right.” A second check that comes out a little different doesn’t mean the meter is broken. Second, the band gets wider in points as the number climbs. A 30-point gap at 200 is inside it. A 30-point gap at 100 isn’t.
While you’re looking at the display, check the units. U.S. meters show mg/dL. Much of the world uses mmol/L, where 100 mg/dL is about 5.55 mmol/L. The FDA’s guidance lists an accidental change of units among the user errors a maker has to plan for. If your numbers suddenly look tiny, that’s the first thing I’d check.
How 18 Meters Did When Researchers Bought Them Off the Shelf
The FDA’s numbers describe how a meter did in its maker’s studies before it went on sale. How meters do afterward, bought from a store like you’d buy one, is a separate question. A team led by David Klonoff, a diabetes researcher in California, set out to answer it and published the results in Diabetes Care in 2018.
They tested 18 meter systems cleared by the FDA, which together made up about 90% of the meters people were using from 2013 to 2015. The meters and strips came from regular stores. In all, 1,035 people had a fingertip reading taken on six different meters, plus a sample sent to a reference lab. The testing was triple-blinded, and every meter went through three separate rounds.
Their bar for a good reading was within 15% of the lab for results of 100 mg/dL or higher, or within 15 mg/dL below 100. The authors said their standard was a little more lenient than the regulators’. Here’s how the meters did:
- 6 meters met the standard in all three rounds.
- 5 met it in two rounds.
- 3 met it in just one.
- 4 didn’t meet it in any.
Those were meters from about a decade ago, and models change, so I wouldn’t shop off that list today. The bigger point holds, though. FDA clearance tells you a meter passed at the start. It doesn’t promise every box sold since performs the same way. If you’re choosing a meter, ask your doctor or your pharmacist which ones they trust, and why.
Wash Your Hands First, Even If They Look Clean
This is the part you control, and it’s the one I’d fix first. Sugar on your skin ends up in the drop of blood, and the meter reads it as yours.
A Dutch team checked this in 123 people with diabetes, in a study published in Diabetes Care in 2011. They compared readings from unwashed fingers, from fingers that had handled fruit, and from washed fingers against a control reading. Without washing, the first drop of blood was off by 10% or more in 11% of people. After handling fruit, the first drop was off by that much in 88% of them.
A Japanese study the same year went a step further. Healthy volunteers peeled an orange, grapes or a kiwi, then had a fingertip reading. Without washing, the readings came out abnormally high, even when the fingertip had been wiped with an alcohol swab once, or five times. After washing with plain tap water, the readings matched those of people who hadn’t touched fruit at all.
So the alcohol wipe is no substitute for the sink. The FDA tells meter makers to put a line in their instructions about washing with soap and water and drying the skin before you test, because anything on the skin can change the result. Hand cream, a sticky kitchen counter, the cookie you just handed a grandkid: it all counts. Wash, dry, then test.
Squeezing, Cold Hands and the Second Drop
On a cold morning, the drop doesn’t always want to come, and the temptation is to squeeze the fingertip hard. The same Dutch study tested that too. Pressing around the finger with different amounts of force changed the reading by 10% or more in 5% to 13% of people. If you have cold hands, wash them in warm water instead. You were going to wash them anyway, and it’s kinder than squeezing.
You’ve probably heard the old tip about wiping away the first drop and using the second. I had too, honestly, more than once. The Dutch researchers were clear about it: wash with soap and water, dry your hands well, then use the first drop. Only if you can’t wash, and your hands aren’t visibly dirty and haven’t touched anything sugary, is it acceptable to wipe away the first drop and use the second. Even then, the second drop from unwashed hands was off by 10% or more for 4% of people, and for 11% after fruit.
Where you prick matters too. Some meters are cleared for the forearm or palm, which can hurt less. The FDA’s guidance says blood from those spots can lag behind the fingertip when your sugar is changing fast: after a meal, after insulin, or during and after exercise. Meters cleared for those sites have to tell you to use the fingertip instead if you think you’re low, if the reading doesn’t agree with how you feel, after a meal or exercise, and when you’re sick or under stress.
Strip Storage and the Fine Print in the Insert
The meter gets the blame, but the strip does the chemistry. The FDA’s list of things a maker must plan for reads like a checklist of how strips go wrong: expired strips, damaged or dirty strips, strips that weren’t made for that meter, and a container that doesn’t keep them dry.
Then there’s the room you’re in. Heat and humidity are both on the FDA’s list of things that can change a result, and so is altitude, which surprised me. So is your own blood. The guidance asks that meters work across a hematocrit (the share of your blood made up of red cells) of about 30% to 55%, and that makers spell out the limits if theirs doesn’t.
Where do you find those limits for your meter? The fine print. The test strip insert, the folded paper in the strip box, is where the maker lists what applies to your strips. A few habits follow straight from that list:
- Keep strips in their original container, with the cap closed.
- Check the expiry date on the box, and the opened-vial date in the insert if it gives one.
- Don’t leave the kit in a hot car or a freezing one.
- Use only strips made for your meter, even if a cheaper box fits the slot.
Many meters also work with a control solution, a liquid with a known amount of glucose in it. The FDA counts skipping the meter checks the labeling calls for as a user error. The manual says when your meter wants one.
Why Your Meter and the Lab Disagree
Here’s what I’d think through before blaming the meter. A lab draw and a fingertip check are two different samples of blood, usually taken at different times. If the lab drew your blood at 8 a.m. after a fast and you tested at home after breakfast, there’s no reason they’d match.
Even side by side, they won’t match exactly. The lab is the reference, and your meter is allowed its 15% band around it. A home 112 against a lab 100 is inside the band. A home 140 against a lab 100 isn’t.
If you want a fair comparison, test with your meter close to the time your blood is drawn, with clean hands, and bring the meter with you. Some offices will look at your reading next to their own. Ask yours. And if your meter and the lab keep disagreeing by more than the band, tell your doctor. That’s worth a fresh box of strips at the least, and maybe a different meter.
Every home device has a band like this. A blood pressure cuff in the wrong size moves the reading, and I went through the same exercise with forehead vs ear thermometers. The trick is knowing how wide the band is before the number worries you.
When to Call Your Doctor, and When to Call 911
A meter’s number is a tool for you and your doctor, not a diagnosis. Call your doctor if:
- your readings keep landing outside the normal range your doctor gave you
- the reading doesn’t match how you feel, more than once
- your meter and your lab results keep disagreeing by more than the band
- you’ve ruled out your hands, the strips and the storage, and the numbers still don’t make sense
MedlinePlus, the National Library of Medicine’s health site, says blood sugar below 70 mg/dL is considered low. If your doctor has given you a plan for lows, follow it. If the signs of a low don’t get better after a snack that has sugar in it, MedlinePlus says to get a ride to the emergency room (don’t drive yourself) or call 911. Get medical help right away for anyone with diabetes or low blood sugar who becomes less alert or can’t be woken up.
And if you feel low but the display says you’re fine, don’t let the meter talk you out of it. Wash, re-test at the fingertip, and follow the plan your doctor gave you. If you’re in doubt, call someone.
The short version: a good meter can be 15% off and still be doing its job. Wash and dry your hands, use the first drop, keep the strips in date and dry, and read the insert once. Do that, and the number on the display is about as close as a home device gets.
If you have a meter question I didn’t answer, leave it in the comments. And take any reading that worries you to your doctor, not to a search engine.
