My Blood Tests Were Getting Better. I Was Still Waking Up at 3am.
Your blood tests are heading the right direction. The GP is pleased. You’ve done the work, and the numbers are starting to show it. And you’re still wide awake at three in the morning, staring at the ceiling, with no explanation you can point to.
That’s not a sleep problem. Or at least, it wasn’t mine. I spent months assuming it was, which is why it took so long to understand what was actually happening. I’ve got NHS blood data going back to 2012, and I tracked 57 nights of nine-out-of-ten sleep quality in 2025 — none of it with medication or melatonin. But the sleep didn’t arrive when my biomarkers improved. It arrived six months later. And if you’re in that gap right now — numbers heading the right way, nights still broken — the next few minutes might explain why.
Because blood sugar instability runs a pattern at night that nobody warned me about. And it affects a lot of people going through any kind of metabolic change.
The Thing Nobody Told Me About Getting Better
The idea that improving your blood markers means everything improves together, roughly at the same time — that’s a reasonable thing to believe. It’s what the logic suggests. Fix the underlying problem, and the downstream effects resolve. It feels like the way repair ought to work.
And to be fair, that’s mostly what happens. Energy improves. Weight moves. Inflammation markers come down. The picture gets better across the board. So when one thing stubbornly refuses to improve — when sleep stays broken even as everything else progresses — the natural conclusion is that it must be a separate problem. Unrelated. Just one of those things.
That was my conclusion too. For months.
By mid-2024, the numbers were genuinely moving. HbA1c had dropped from 47 — that’s pre-diabetic territory — down to 40. ALT, which had been elevated, was coming back into range. My GP was pleased. I was pleased. On paper, the inputs were working.
And then I’d wake up. Three in the morning. Three-fifteen. Sometimes quarter to four. Completely alert. Not groggy. Not needing the bathroom. Just — wide awake, for no reason I could name.
I thought it was stress. Then room temperature. Then I decided it was probably just age and I’d have to live with it. I never once thought it might be connected to the same metabolic repair I was tracking everything else against. That connection took a while longer to surface.
What Blood Sugar Does at Night That Your Daytime Tests Won’t Show
Here’s the mechanism, because it took me too long to understand it and I’d rather you didn’t spend months in the same fog.
Your blood sugar drops during the night. That’s normal. You’re fasted, you’re asleep, that’s what should happen. But if your system is still recalibrating — if insulin sensitivity is improving but not fully repaired — the overnight drop can go slightly further than intended. Your body reads that as a low-fuel signal. It responds by releasing cortisol. Cortisol raises blood sugar. You wake up.
Not groggy. Not confused. Clean awake. Because cortisol is essentially an alert hormone. It’s designed to bring you round.
Here’s where the blood sugar picture gets interesting. The metric I’d been tracking most closely — HbA1c — is a three-month average. It tells you where blood sugar has been over the past twelve weeks. It doesn’t tell you what’s happening at two in the morning on a Tuesday when your liver decides to top up your glucose supply and slightly overshoots. The daytime picture can look fine while the nocturnal pattern is still running its old programme.
Jessie Inchauspé — biochemist, wrote Glucose Revolution — helped me understand it properly. If glucose rises and then drops during the night, the body may respond with a cortisol release. And that cortisol, in some people, produces a clean wide-awake moment at three or four in the morning. Not a sleep disorder. A fuel management response wearing insomnia’s coat.
Think of it like a central heating thermostat that hasn’t been recalibrated after you’ve properly insulated the house. The house is warmer now — the work is done — but the thermostat doesn’t know that yet. So it keeps firing the boiler at three in the morning to hit a temperature the house doesn’t need anymore. The system is improving. The controller is still running the old programme.
That’s exactly what was happening to me.
data going back to 2012 and a Six-Month Gap
January 2024. HbA1c at 43, still pre-diabetic. I’d started the Boxing Day dietary shift in late December 2023, so we’re five or six weeks in. Sleep at this point: broken. Four, five hours. Waking consistently in the small hours.
March 2024. HbA1c coming down. Evening meals moving earlier. The picture improving on paper.
May 2024. Blood sugar markers continuing to improve. Sleep: still disrupted. Still the 3am wake. I’d go to bed fine, drop off without trouble, and then somewhere between two-thirty and four I’d be completely awake. Sometimes forty minutes. Sometimes two hours.
The blood tests were improving faster than the sleep. That gap — that’s the pattern.
HbA1c is a lagging indicator. It reflects where you’ve been, not what’s happening tonight. The nocturnal glucose pattern can still be dysregulated while the quarterly average looks better. They’re measuring different things across different timeframes. I’d been treating them as the same signal, which is why the 3am waking made no sense to me. The signal I was watching was three months behind the signal I should have been watching.
There was a second moment that crystallised something else entirely. Five consecutive nights of 8.5 hours sleep — and I woke every morning feeling as though I’d barely rested. The hours were there. But the timing was wrong. I’d gone to bed around one and woken naturally at nine-thirty. And here’s what I’d missed: cortisol and glucose start rising an hour or two before your body expects to wake. If your body has been calibrated for a seven o’clock rise but you’re still asleep at nine-thirty, that preparation process — the glucose mobilisation, the cortisol lift — has been running for two and a half hours into your sleep. You wake feeling mildly stressed. Because physiologically, you were.
Duration isn’t the whole picture. Timing is a separate variable. I’d been measuring one and ignoring the other.
Blood sugar isn’t about having as little as possible — it’s about stable delivery at the right time. Think about the difference between the rolling Yorkshire Dales and a ski slope in Val d’Isère. Both involve elevation changes. But the Dales roll up and down gently, gradually, predictably. The ski slope goes straight up and then drops vertically. After every sharp spike comes a sharp drop — and at night, that drop is what triggers the cortisol response. The goal isn’t to eliminate the hills. It’s to avoid the ski slopes, especially in the hours before sleep.
What this suggested to me — and I want to be clear this is N=1, one person, data going back to 2012 — was that sleep isn’t a parallel track. It’s downstream. You fix the metabolic inputs, and sleep repairs. But it repairs last. The blood tests improve first. The energy improves. The weight moves. And then, eventually, the sleep catches up.
Six months, in my case. That’s the lag I tracked.
What Actually Changed — and When
Not a supplement. Not a sleep protocol. Not a particular bedtime routine. The thing that shifted was the eating window, and when I moved it.
Moving the evening meal earlier was the single change that altered the nocturnal pattern. Not dramatically, not overnight. But consistently, across the data, that’s what moved it. Less glucose activity in the final hours before sleep meant a smoother overnight curve. The thermostat had less reason to fire the boiler.
My GP’s read on this, when I raised it: the overnight cortisol pattern is well-established. Not surprising to him. What surprised him was that I’d pieced it together from my own data rather than presenting with a complaint and waiting for a diagnosis. That’s the advantage of tracking this across years. You see the pattern before anyone names it.
He also made a point I think is worth sitting with: the overlap between metabolic repair and sleep repair is real, but the lag is poorly communicated. People get told their numbers are improving and reasonably expect to feel it across the board. When sleep doesn’t follow immediately, they assume it’s a separate problem requiring a separate intervention. Sometimes it isn’t. Sometimes it’s just the sequence.
Three Things Worth Knowing If This Is Where You Are
Track sleep timing alongside sleep duration from the start. Not just hours — when. The five nights of 8.5 hours that felt like nothing would have taught me the timing lesson six months earlier if I’d been logging the window rather than just the total. Hours in isolation don’t tell you the whole story.
If your daytime markers are improving but you’re still waking consistently between two and four, it’s worth checking what you’re eating in the three hours before bed, and the timing of your last meal. The cortisol and glucose pattern is worth raising with your GP. They may recognise it. They can help you assess whether it fits your situation.
If you feel wrong but don’t have a label for it yet, this is for you. Waking at a specific time — three, three-thirty, four — consistently, with no obvious explanation — it might not be stress and it might not be your mattress. This kind of blood sugar and cortisol pattern can show up before anything is clinically labelled. Years before, sometimes. It shows up in the sleep first. The daytime markers can look fine while this is happening at night. Worth being curious about. Not a diagnosis. Just a thing to track. Note what you ate and when for a week. See if there’s a pattern. That’s all I did.
The obstacle here wasn’t willpower. It wasn’t a character flaw. It wasn’t even a sleep disorder. It was a metabolic system still learning the new normal, running a nocturnal programme that the daytime data hadn’t caught up with yet. Once I understood that, the six-month lag stopped feeling like failure. It felt like sequence.
That’s what the data showed me. You can look at your own and see if it applies.
NHS blood tests going back to 2012 — the actual numbers, the GP conversations, what changed and what didn’t. If you want to follow someone who documents this properly rather than just talks about it, that’s what this channel is. Subscribe if that’s useful to you.
If you want to start tracking your own patterns — sleep timing, eating window, how they interact — the 14-Day Rhythm Reset Guide covers the fundamentals: eating windows, what to eat, how to structure your days, and how to start building from there.
👉 Get the 14-Day Rhythm Reset Guide (free)
💪 Want ongoing support? Join Rhythm Club for £30/month (includes FREE app access, community, and weekly accountability calls)
🔽 Read the Video Transcript
My blood tests were coming back better every month. And I was still waking up at three in the morning for no reason I could find.
Not occasionally. Consistently. Month after month of improving markers, and still this — completely wide awake in the early hours, staring at the ceiling, no idea why. If I’d fixed the things that were supposed to be broken, why was my sleep still behaving like I hadn’t?
I’m Neil — I’ve got NHS blood data going back to 2012 and I tracked 57 nights of nine-out-of-ten sleep in 2025, all of it without medication or melatonin. But it didn’t arrive when my biomarkers improved. It arrived six months later. And if you’re in that gap right now — numbers heading the right way, sleep still chaotic — the next ten minutes might explain why, because blood sugar instability runs a pattern at night that nobody warned me about, and it affects most people going through any kind of metabolic change. First — the mechanism: what blood sugar instability actually does between midnight and four in the morning, and why it produces a cortisol spike that looks exactly like insomnia but isn’t. Then — my data: what the six-month lag looked like in NHS records going back to 2012, and the specific pattern that finally told me sleep was last in the repair sequence. Then — what actually changed: not a supplement, not a sleep protocol, not a bedtime routine. The thing that shifted, and when.
THE PUZZLE
So here’s the situation I was in.
By mid-2024, the numbers were genuinely moving. HbA1c had dropped from 47 — that’s pre-diabetic — down to 40. Cholesterol still high, but trending. ALT, which had been elevated, coming back into range. My GP was pleased. I was pleased. On paper, the inputs were working.
And then I’d wake up. Three, three-fifteen, sometimes quarter to four. Completely alert. Not groggy. Not needing the bathroom. Just — wide awake, for no reason I could name.
I thought it was stress. Then I thought it was the room temperature. Then I decided it was probably just age and I’d have to live with it. I’d been assuming it was a sleep problem. Took me months to work out it wasn’t a sleep problem at all.
The thing I’d missed — and this is the pattern nobody flagged for me — was that my blood sugar, even as it improved over months, was still doing something at night. Something that my daytime readings weren’t showing. And that nocturnal pattern had a very specific consequence at a very specific time.
Three in the morning, give or take.
THE DATA
January 2024. HbA1c 43. Still pre-diabetic range. This was before the dietary changes had fully compounded — I’d started the Boxing Day shift in late December 2023, so we’re five or six weeks in. Sleep at this point: broken. Four, five hours. Waking consistently in the small hours. I wasn’t tracking it formally yet, just noting it.
March 2024. HbA1c coming down. The dietary window had shifted. I’d stopped eating late. Evening meals moved earlier — mid-summer is when that fully locked in, but the change started earlier in the year. And here’s the thing — the HbA1c was improving faster than the sleep.
That gap is the pattern.
May 2024. Blood sugar markers continuing to improve. Sleep: still disrupted. Still the 3am wake. I’d go to bed fine, drop off without trouble, and then somewhere between two-thirty and four I’d be completely awake. Sometimes forty minutes. Sometimes two hours.
What I didn’t understand at the time — and this took me a while to piece together — is that HbA1c is a three-month average. It tells you where blood sugar has been. It doesn’t tell you what’s happening at two in the morning on a Tuesday when your liver decides it needs to top up your glucose supply and overshoots.
The nocturnal pattern works like this. Your blood sugar drops during the night — that’s normal, you’re fasted, that’s what should happen. But if your system is still recalibrating, if insulin sensitivity is still improving but not fully repaired, the overnight drop can go a bit further than intended. Your body reads that as a low-fuel signal. It responds by releasing cortisol. Cortisol raises blood sugar. You wake up.
In my case, it looked less like a sleep disorder and more like a fuel management problem wearing insomnia’s coat.
Think of it like a central heating thermostat that hasn’t been recalibrated after you’ve properly insulated the house. The house is warmer now — the work is done — but the thermostat doesn’t know that yet. So it keeps firing the boiler at three in the morning to hit a temperature the house doesn’t need anymore. The system is improving. The controller is still running the old programme.
That’s exactly what was happening to me.
Now — the data point that made this click. I’d been tracking sleep quality scores informally alongside the blood markers. And I noticed — not immediately, I had to lay things out and look — that the 3am waking was still present even on the best dietary days. Even when I’d eaten well, eaten early, done everything right. The sleep disruption was lagging behind the dietary change by weeks. Then months.
The second moment came later in 2024. Five consecutive nights of 8.5 hours sleep. I felt rough. Woke feeling like I’d barely slept. And the reason, once I’d thought it through, was timing. I’d gone to bed late — around one — and woken naturally at nine-thirty. The hours were there. But the window was wrong. Cortisol and glucose start rising an hour or two before your body expects to wake. If your body expects to wake at seven but you’re still asleep at nine-thirty, that preparation process — the glucose mobilisation, the cortisol lift — has been running for two and a half hours into your sleep. You wake up feeling like you’ve been mildly stressed all night. Because you have.
Duration isn’t the whole story. Timing matters.
NHS blood tests going back to 2012 — the actual numbers, the GP conversations, what changed and what didn’t. If you want to follow someone who documents this properly rather than just talks about it, that’s what this channel is. Subscribe if that’s useful to you.
THE INSIGHT
So the penny dropped, as these things tend to, while I was reading.
I came across Jessie Inchauspé — biochemist, wrote Glucose Revolution — and she helped me understand a possible pattern: if glucose rises and then drops during the night, the body may respond with a cortisol release — and that cortisol, in some people, is enough to produce a clean wide-awake moment at three or four in the morning. Not a sleep disorder. A fuel management response.
Later that day it properly landed. The mechanism I’d been experiencing had a name. And more importantly, it had a cause that was addressable — not with a sleep supplement, not with a bedtime routine, but with what I was eating and when I was eating it relative to sleep.
The cortisol spike at three in the morning wasn’t anxiety, wasn’t age, wasn’t insomnia. It was my body correcting a blood sugar dip. And the blood sugar dip was happening because my system — as it repaired — was still learning the new normal.
What this suggested to me — and I want to be clear this is N=1, one person, data going back to 2012, make of that what you will — was that sleep isn’t a parallel track. It’s downstream. You fix the metabolic inputs, and sleep repairs. But it repairs last. The blood tests improve first. The energy improves. The weight moves. And then, eventually, the sleep catches up.
Six months, in my case. That’s the gap I tracked.
My GP’s read on this, when I raised it: the overnight cortisol pattern is well-established. Not surprising to him. What surprised him was that I’d pieced it together from my own data rather than presenting with a complaint and waiting for a diagnosis. That’s the advantage of tracking this stuff. You see the pattern before anyone names it.
THE LESSON
Three things. Quick.
What I’d do differently. I’d track sleep timing alongside sleep duration from day one. Not just hours — when. The five nights of 8.5 hours that felt like nothing would have taught me the timing lesson six months earlier if I’d been logging the window rather than just the total. Hours in isolation don’t tell you enough.
What to watch for — if you’re already tracking. If your daytime markers are improving but you’re still waking consistently between two and four — don’t assume it’s a separate problem. Check what you’re eating in the three hours before bed. Check the timing of your last meal. In my case, moving the eating window earlier was the single change that shifted the pattern. Not dramatically, not overnight. But that’s what moved it. If you’re seeing this in your own data, it’s worth raising with your GP — your GP may recognise the cortisol and glucose pattern, and they can help you decide whether it fits your situation.
The early warning signal — if you don’t have a diagnosis yet. This one’s for the people who feel wrong but haven’t got a label for it yet. If you’re waking at a specific time — three, three-thirty, four — and it’s consistent, and you can’t explain it, it might not be stress and it might not be your mattress. This kind of blood sugar and cortisol pattern may show up before anything is labelled clinically. Years before, sometimes. It shows up in the sleep first. The daytime markers can look fine while this is happening at night. If that’s you — just worth being curious about. Not a diagnosis. Just a thing to track. Note what you ate and when, for a week. See if there’s a pattern. That’s all I did.
CTA + SAFETY
If you want to start tracking your own patterns — sleep timing, eating window, how they interact — I’ve got a free tracker template in the description. It’s what I used to spot this. Nothing complicated. Just a log.
Link below.
And the standard note — everything here is educational, not medical advice. Speak to your GP before changing your fasting pattern, diet, or medication. All the biomarker changes I’ve documented are on my NHS record.
Next Monday: a different pattern, same detective process.
⚠️ Important: This is educational content based on evidence and personal experience. It is not medical advice. Speak to your GP before making changes to your fasting pattern, diet, or medication.
Questions? Email hello@waymorethanweightloss.com
© 2025 Way More Than Weight Loss. All Rights Reserved.
