7 Mistakes People Make Before a Blood Test (That Change Your Results)
Blood test results aren't as fixed as most people think. You can get tested twice in the same week and get genuinely different numbers — because of what you did beforehand, not because anything changed inside you.
Prefer to watch? This guide is also a 16-minute video. Watch it on YouTube.
So you get your blood test back. And there it is. One number with a little H next to it. Slightly too high.
Before you start worrying about that, there's something you should probably know.
That number might not be telling you about your health at all. It might just be telling you about your Tuesday morning.
Here's the thing. Blood tests aren't as fixed as most people think. You could get tested twice in the same week and get genuinely different numbers. Nothing changed inside you. What changed is what you did the day before.
So I want to go through seven things that move your results. None of them are complicated. Most people do at least two of them without anyone ever mentioning it. And number seven? That's the one that gets people sent back for repeat tests they never needed.
Quick word about me, and I'll keep it short. I'm a researcher. Not a doctor. I read the studies and the lab guidance, and I tell you what's actually in them. So none of this is medical advice, and nothing here should change any medication or supplement you're taking. That's between you and your doctor. Always.
What this is for is simpler than that. It's making sure that when they take your blood, what comes back actually looks like you.
How it goes wrong
Let me show you how this actually plays out. Because it's never just one thing. It stacks.
Picture a man in his late fifties. Sensible. Does what he's told. He books the first appointment of the day, because that's the only one you can ever get.
On the Sunday he goes out on his bike and does forty miles. Because he's been told exercise is good for him. Which it is.
Monday morning, eight o'clock. He's fasted properly since eight the night before. And he skips his morning glass of water, because he figures fasting means nothing at all. Seems reasonable.
Couple of weeks later, a letter turns up. His creatinine is up. His eGFR has come back at fifty-eight. And the letter uses the phrase "reduced kidney function."
He spends the next three weeks assuming his kidneys are packing in.
Then they retest him. Rested this time. Properly hydrated. Ordinary week. Comes back at seventy-nine.
Nothing about his kidneys changed between those two mornings. What changed was forty miles on a bike and a glass of water.
Now, that's a composite. It's not one person's notes. But every single ingredient in it is completely ordinary. That's why I'm starting with it. He didn't do anything daft. He did three sensible things that all happened to push the same number the same way.
What a "normal range" really is
Right. Before we get to the seven, there's one idea that makes all of them easier to understand. And honestly, I wish someone had explained this to me years ago.
A reference range isn't a health target. It's a description of a crowd.
Here's how they make one. A lab measures something across a big group of people they've decided are healthy. Then they draw a line around roughly the middle ninety-five percent of them. That's your range.
It's not saying what's good for you. It's saying where most people happened to sit.
Now follow that through, because something useful falls out of it.
If the range covers ninety-five percent of healthy people, then about one in twenty healthy people land outside it. On any given marker. Nothing wrong with them at all. That's just how the maths works.
So count the markers on a standard panel. Twenty is pretty normal. Which means on a routine blood test, for a completely healthy person, one flagged value is about what you'd expect.
Not unusual. Expected.
And that's why your doctor can look at a letter that's had you awake half the night and seem totally unbothered. They're not brushing you off. They're looking at one mildly out-of-range number on a twenty-marker panel and reading it as what it usually is. Noise. Until something says otherwise.
One more thing on this, and it matters if you ever compare your numbers to something you've read online.
Every lab's range is different. Different machines, different methods, different populations, different countries. The range printed on your report is the only one that applies to your result. Not the one in a video. Not the one in an article. Not mine. Yours.
So a mild flag is a reason to look closer. It isn't a diagnosis.
Right. With that in mind, here are the seven things most likely to have caused it.
1. Fasting, done wrong
Starting with the obvious one. Except even here, people go wrong in a direction nobody warns them about.
If you were told to fast, the two tests that really care are glucose and triglycerides. Both react fast to food. Eat a couple of hours before a fasting glucose test and you can turn a perfectly normal reading into one that looks like prediabetes.
But here's what people get wrong. More fasting isn't better.
The window is usually eight to twelve hours. And some people, trying to be really thorough, turn up having not eaten for sixteen or eighteen. A long fast does its own things to your blood sugar. So you haven't been extra careful. You've just run a different test.
And if you were meant to fast and you didn't? Just say so. A non-fasting sample that nobody knows is non-fasting is far worse than moving your appointment. Nobody's going to be annoyed with you. They'll be much more annoyed chasing a result that was never real in the first place.
So: stick to the window you were given. And if nobody gave you one, ask.
2. Turning up dehydrated
This one's almost universal. Because fasting appointments are early, and people assume fasting means no water either.
When you're dehydrated, the liquid part of your blood shrinks. So everything measured in it gets more concentrated. That can push up your haemoglobin, your haematocrit, your sodium. And your creatinine, which is the number they calculate your kidney function from.
So you turn up dry after an early start, and your eGFR comes back a bit low. Which is a genuinely frightening thing to read about your kidneys. And it might be describing nothing more than a glass of water you didn't drink.
Water is nearly always fine while fasting. Nearly. A few specific tests do restrict it. So check with the clinic rather than guessing.
Easiest thing in the world: when you book, just ask "can I drink water?" If they say yes, drink normally that morning.
3. A hard workout the day before
This one catches out exactly the people who least deserve it. The ones over fifty who've taken up lifting or cycling and are doing everything right.
Creatinine is a waste product from normal muscle turnover. Work your muscles hard, you make more of it. It's not a kidney problem. It's a muscle signal that happens to get measured through your kidneys.
So a heavy session a day or two before your blood draw can push creatinine up. And muscle enzymes too. If AST is on your panel, which is one of the liver markers, that can go up from exercise as well. Because AST isn't only in your liver. So you get a letter saying your liver enzymes are raised, and the actual cause was a deadlift session.
There's a related trap here. Just having a lot of muscle raises your baseline creatinine. Someone well built can sit above the range with completely healthy kidneys. The formula that works out your eGFR has no idea how much muscle you're carrying.
So keep the couple of days before a test boring. And if your creatinine or your liver enzymes come back up, that's a fair thing to ask about.
4. Supplements, and especially biotin
If you take one thing from this guide, I'd quite like it to be this one. Because it's invisible, and the results can look genuinely alarming.
High dose biotin. Vitamin B seven. It's the active ingredient in most hair, skin and nail supplements. And it's well documented that it interferes with a whole family of lab tests.
Not by changing anything in your body. By messing with the measurement itself.
A lot of lab tests use a biotin binding step in their chemistry. Flood the sample with biotin and that step misbehaves. Depending on how the test is built, your result can come out falsely high or falsely low.
The tests most affected include several thyroid tests and some heart ones. And this really matters. Biotin interference has produced thyroid panels that looked like serious thyroid disease, in people whose thyroids were completely fine. It's been reported to interfere with troponin, which is the test they use when someone turns up with chest pain.
The dose in an ordinary multivitamin is usually too small to matter. The dose in a dedicated hair and nails product is often much, much higher.
So tell whoever takes your blood about every supplement you take. Biotin especially. A lot of labs advise pausing high dose biotin beforehand, but how long depends on the dose and the test, so ask them rather than guessing.
And please don't stop a prescribed medication for a blood test unless the person who prescribed it tells you to.
5. The wrong time of day
Some things in your blood run on a daily rhythm. And for those, the time on the tube is part of the result.
Cortisol is the clearest one. Highest early in the morning, drops off through the day. A cortisol level that's perfectly normal at eight in the morning would look distinctly odd at eight at night. Testosterone does something similar, which is why they measure it in the morning. Iron moves about through the day too.
But the real issue isn't any single test. It's comparing them.
If last year's test was at eight in the morning and this year's was at four in the afternoon, and you're trying to work out whether something's going up or down, you might just be reading the clock.
So book the same sort of time each go. And if something time-sensitive comes back odd, check when it was taken before you read anything into it.
6. Alcohol in the last couple of days
A drink or two in the couple of days before your test can move several things at once.
Triglycerides react quickly to alcohol. A heavy night can push them up quite a bit. Liver markers react too. GGT especially, and that's one of the enzymes they use to work out whether a liver signal is alcohol related. And because alcohol messes with how you handle blood sugar, glucose can shift as well.
So one social evening can produce a panel with raised triglycerides and raised liver enzymes. Which together looks like a real metabolic picture. And might be describing a wedding.
If you can, keep the two days before clear. If you can't, just mention it, so somebody reads it in context.
7. Testing while you're ill, or just after
And here's the one that sends people back for repeat tests they never needed.
When you're fighting an infection, your inflammatory markers go up. CRP can climb sharply and then drop away just as fast. Your white cell count shifts. And ferritin, which is meant to show your iron stores, also goes up with inflammation.
Which means a cold can make your iron stores look better than they are, and hide a real deficiency.
That one has real consequences. Someone properly low in iron, tested while they've got a chest infection, can get a normal looking ferritin and be told everything's fine.
A routine panel taken in the fortnight after you've been ill is describing your recovery. Not your baseline.
So if it's a routine test and you've been poorly recently, ask whether it's worth waiting until you're properly better. And if it can't wait, which sometimes it can't, make sure the illness is on the record.
The habit that beats all seven
If you only take one thing from all of this, take this. Standardise.
Same lab. Same time of morning. Same fasting state. Same ordinary sort of week beforehand.
You're not chasing one perfect result. You're building a trend line. And a trend only means anything if the conditions stay the same.
One number on its own tells your doctor surprisingly little. Five years of numbers taken the same way tells them an enormous amount.
Which is also why the most useful question you can ask isn't "is this normal?" It's "what's my trend on this over the last few years?" Your practice has that history sitting there. Hardly anyone ever asks for it.
The thirty second checklist
Before they take your blood, make sure they know:
Every medication and supplement you take, with doses. Biotin especially. Whether you actually fasted, and how long for. Any illness in the last two weeks. Any hard exercise in the last two days. And any alcohol in the last two days.
That's it. Thirty seconds of context. And it's the difference between a harmless blip being understood, and a harmless blip being chased.
The short version
Now, none of this means you should ignore a flagged result. Some of them matter a great deal.
What it means is that one odd number, taken once, in conditions nobody wrote down, is a reason to look closer. It isn't a diagnosis.
So if something on your last set of results has been sitting at the back of your mind, take it to your doctor. Ask whether the conditions on the day could explain it. Ask whether it's worth repeating. And ask what your trend looks like.
Those three questions will get you a lot further than searching your symptoms at eleven o'clock at night.
I've put those, and a couple of others that work for any flagged result, on a one page card you can print off and take with you. It's free. It's just below.
Further reading
I'm a researcher, not a doctor. Health Unfiltered is education, not medical advice. Nothing here should be used to start, stop or change any medication or supplement. That is a decision for you and a qualified clinician. If a result worries you, speak to your doctor.
The other guides
- How to read a blood test report
- Full blood count
- Cholesterol and the lipid panel
- Blood sugar, HbA1c and prediabetes
- Blood pressure
- Kidney function, eGFR and creatinine
- Thyroid, TSH and free T4
- Iron and ferritin
- Liver enzymes
- CRP and inflammation
- Urea and electrolytes (U&E)
Want every marker in one place? The Over-50 Blood Test Decoder covers each line of a routine panel in plain English, with the question to ask for each one.