April 2025. My GP looked at my blood work and seemed genuinely pleased. HbA1c — the long-term blood sugar marker — had dropped from forty-seven to forty. Pre-diabetes range down to optimal. Thirteen months of eating differently, fasting regularly, tracking everything. The numbers were moving in the right direction for the first time in years.
I was still waking up at three in the morning.
Still tired by two in the afternoon. Still not sleeping properly, despite doing everything the blood markers said I was doing right. That gap — between when the clinical numbers improved and when the sleep actually repaired — turned out to be the most important thing I learned from thirteen months of correlation data. And nobody told me the gap would exist.
The Belief That Makes the Gap So Confusing
There is a completely reasonable assumption buried in how most health conversations go. Fix the metabolic problem, fix everything downstream. Sort the blood sugar, sort the sleep. It makes intuitive sense, and it is the kind of logic that gets reinforced every time a GP ticks a box and says “levels are improving.”
What that framing misses is that the metabolic environment and the sleep architecture are not the same clock. They are related. They influence each other. But they do not move together.
I held the reasonable assumption for most of 2024. The bloods are improving, the sleep will follow. Month after month, the bloods kept improving and the sleep kept being unreliable. Waking in the small hours. Fragmented. Unrefreshing in a way that is hard to articulate — not dramatic insomnia, just consistently poor quality, like a phone that charges to sixty percent and never quite gets to full.
The assumption was not wrong. It was just incomplete.
What Is Actually Happening When Poor Sleep Breaks Blood Sugar
The mechanism runs in both directions. That is the bit that surprised me, and the bit that most sleep advice never quite gets to.
Start with the direction most people know. Peter Attia cites work from Eve van Cauter’s group at the University of Chicago, confirmed across nine randomised controlled trials. Sleep deprivation of four to five hours — the kind a lot of people run on as a matter of routine — creates insulin resistance equivalent to early type 2 diabetes. Not after months of disruption. Within days.
Benjamin Bikman, a biochemist who communicates this better than almost anyone, puts a number on it. One week of insufficient sleep makes the body roughly thirty percent more insulin resistant. A separate study found that just two days of fifty percent sleep reduction was enough to make otherwise healthy men insulin resistant. Two days.
When you are sleep deprived, cortisol rises. Cortisol mobilises glucose — that is its job, it is a stress hormone preparing your body for effort. Blood glucose goes up. At the same time, the cells that would normally respond to insulin start becoming less sensitive. The combination of elevated cortisol and disrupted hormonal rhythms reduces how well insulin can do its job. The result: higher blood sugar, insulin working harder, the whole system running under increased load.
Sophie Bostock, a sleep scientist whose work I have found consistently rigorous, describes this as a full physiological stress cascade. Cortisol elevated. Blood pressure up. Blood glucose mobilised. Run that cascade night after night and it compounds into immune dysregulation, inflammation, and cardiovascular risk.
Now here is the direction nobody talks about. Michael Mosley cites research from a Cornell sleep lab where they controlled what participants ate and measured the sleep it produced. Diet rich in fibre and protein produced measurably more deep sleep. Diet rich in sugar and simple carbohydrates produced lighter, more fragmented sleep throughout the night. The relationship runs the other way. What you eat shapes the architecture of your sleep.
Which means if your blood sugar has been dysregulated for years — as mine had, HbA1c climbing from 38 in 2015 to 47 by 2022 — the sleep architecture has been deteriorating the whole time. And fixing the blood sugar number is not an instant fix for the sleep. Because the circadian rhythms that govern when you fall asleep, when you drop into deep sleep, when cortisol starts rising in the morning — those patterns do not snap back the moment your HbA1c comes down.
That is the lag. That is the gap. And it is not your fault for not sleeping well once the numbers improve. The mechanism you disrupted has to repair on its own timeline, not yours.
The Bin Lorry That Cannot Do Its Rounds
There is a particular cost to poor deep sleep that I think most people underestimate, partly because it is invisible until it is not.
Deep sleep is when the council bin lorry does its rounds in your brain. All day, your brain accumulates waste — dead cells, metabolic debris, the beta-amyloid proteins associated with Alzheimer’s disease. Deep sleep is when the lorry drives through, clears the lot, hoses down the street. Miss deep sleep — five-hour nights, fragmented sleep, too much alcohol before bed — and the rubbish piles up. Week after week, year after year.
But here is the connection to the metabolic story. If the metabolic environment is disrupted — blood sugar unstable overnight, cortisol elevated — the lorry cannot do its route properly. The bins do not get emptied. Everything backs up.
This is not theoretical risk. Brenda Davis, a registered dietitian with a solid clinical background, makes the point that both high and low overnight blood glucose directly disrupt sleep architecture. High overnight glucose disrupts sleep quality. Low blood glucose in the small hours triggers cortisol release, which wakes you up. The vicious cycle runs: poor blood sugar control worsens sleep, and poor sleep worsens blood sugar control.
Duration does not fix that cycle. Fixing the metabolic environment is what fixes the cycle. And fixing the metabolic environment takes considerably longer than most sleepThe Belief That Makes the Gap So Confusing
There is a completely reasonable assumption buried in how most health conversations go. Fix the metabolic problem, fix everything downstream. Sort the blood sugar, sort the sleep. It makes intuitive sense, and it is the kind of logic that gets reinforced every time a GP ticks a box and says “levels are improving.”
What that framing misses is that the metabolic environment and the sleep architecture are not the same clock. They are related. They influence each other. But they do not move together.
I held the reasonable assumption for most of 2024. The bloods are improving, the sleep will follow. Month after month, the bloods kept improving and the sleep kept being unreliable. Waking in the small hours. Fragmented. Unrefreshing in a way that is hard to articulate — not dramatic insomnia, just consistently poor quality, like a phone that charges to sixty percent and never quite gets to full.
The assumption was not wrong. It was just incomplete.
What Is Actually Happening When Poor Sleep Breaks Blood Sugar
The mechanism runs in both directions. That is the bit that surprised me, and the bit that most sleep advice never quite gets to.
Start with the direction most people know. Peter Attia cites work from Eve van Cauter’s group at the University of Chicago, confirmed across nine randomised controlled trials. Sleep deprivation of four to five hours — the kind a lot of people run on as a matter of routine — creates insulin resistance equivalent to early type 2 diabetes. Not after months of disruption. Within days.
Benjamin Bikman, a biochemist who communicates this better than almost anyone, puts a number on it. One week of insufficient sleep makes the body roughly thirty percent more insulin resistant. A separate study found that just two days of fifty percent sleep reduction was enough to make otherwise healthy men insulin resistant. Two days.
When you are sleep deprived, cortisol rises. Cortisol mobilises glucose — that is its job, it is a stress hormone preparing your body for effort. Blood glucose goes up. At the same time, the cells that would normally respond to insulin start becoming less sensitive. The combination of elevated cortisol and disrupted hormonal rhythms reduces how well insulin can do its job. The result: higher blood sugar, insulin working harder, the whole system running under increased load.
Sophie Bostock, a sleep scientist whose work I have found consistently rigorous, describes this as a full physiological stress cascade. Cortisol elevated. Blood pressure up. Blood glucose mobilised. Run that cascade night after night and it compounds into immune dysregulation, inflammation, and cardiovascular risk.
Now here is the direction nobody talks about. Michael Mosley cites research from a Cornell sleep lab where they controlled what participants ate and measured the sleep it produced. Diet rich in fibre and protein produced measurably more deep sleep. Diet rich in sugar and simple carbohydrates produced lighter, more fragmented sleep throughout the night. The relationship runs the other way. What you eat shapes the architecture of your sleep.
Which means if your blood sugar has been dysregulated for years — as mine had, HbA1c climbing from 38 in 2015 to 47 by 2022 — the sleep architecture has been deteriorating the whole time. And fixing the blood sugar number is not an instant fix for the sleep. Because the circadian rhythms that govern when you fall asleep, when you drop into deep sleep, when cortisol starts rising in the morning — those patterns do not snap back the moment your HbA1c comes down.
That is the lag. That is the gap. And it is not your fault for not sleeping well once the numbers improve. The mechanism you disrupted has to repair on its own timeline, not yours.
The Bin Lorry That Cannot Do Its Rounds
There is a particular cost to poor deep sleep that I think most people underestimate, partly because it is invisible until it is not.
Deep sleep is when the council bin lorry does its rounds in your brain. All day, your brain accumulates waste — dead cells, metabolic debris, the beta-amyloid proteins associated with Alzheimer’s disease. Deep sleep is when the lorry drives through, clears the lot, hoses down the street. Miss deep sleep — five-hour nights, fragmented sleep, too much alcohol before bed — and the rubbish piles up. Week after week, year after year.
But here is the connection to the metabolic story. If the metabolic environment is disrupted — blood sugar unstable overnight, cortisol elevated — the lorry cannot do its route properly. The bins do not get emptied. Everything backs up.
This is not theoretical risk. Brenda Davis, a registered dietitian with a solid clinical background, makes the point that both high and low overnight blood glucose directly disrupt sleep architecture. High overnight glucose disrupts sleep quality. Low blood glucose in the small hours triggers cortisol release, which wakes you up. The vicious cycle runs: poor blood sugar control worsens sleep, and poor sleep worsens blood sugar control.
Duration does not fix that cycle. Fixing the metabolic environment is what fixes the cycle. And fixing the metabolic environment takes considerably longer than most sleep advice acknowledges.
Why the Two Processes Do Not Sync at the Same Speed
Your sleep is governed by two completely separate systems running simultaneously. Understanding both of them explains why my blood work improved months before my sleep did.
The first is like an hourglass. It fills with sand the longer you are awake — a chemical called adenosine accumulating, building what researchers call sleep pressure. When enough sand has accumulated, you fall asleep. The second is like a pendulum — your circadian rhythm, swinging independently, controlling when those thresholds rise and fall across the 24-hour day.
Both have to align for sleep to arrive easily.
This is why you can be completely exhausted and still not sleep. The hourglass is full — the sleep pressure is there. But the pendulum is at the wrong point in its swing. The circadian system is out of sync, or something upstream — stress, cortisol, blood sugar instability — is interfering with the signal.
What I experienced for months was essentially this: the hourglass was full. I was tired. But the pendulum had been knocked off its rhythm by years of metabolic disruption, and it took time — considerably more time than I expected — for it to find its natural cadence again.
The blood marker improved first because it reflects what is happening in a single blood draw. The circadian system repairs over weeks and months, not days. It is a different clock entirely.
The body also has an anticipatory function that makes this concrete. An hour before your alarm goes off, your muscles are already warming up. Genes are switching on. Metabolic rate rising. Blood pressure preparing to support upright activity. Your body knows you are going to wake up and it is getting ready before you do — like a factory bringing the machines up to temperature before the shift starts.
This warm-up requires the factory to know when the shift starts. Consistent wake time. Consistent signals. If your circadian system has been disrupted — which metabolic dysfunction tends to do, because insulin, cortisol, and blood glucose are all circadian-regulated — the factory does not know when to start warming up. You get out of bed into a body that is not ready. Cold start every morning. And cold starts compound, for longer than feels fair.
What Thirteen Months of Correlation Data Actually Showed
I want to be straight about where my data is strongest and where it carries more uncertainty.
The blood sugar data is solid. NHS records going back to 2012. HbA1c at 38 in 2015, climbing to 47 by 2022, back down to 40 by January 2025, and 37 by April 2025 — documented, GP-reviewed at every stage.
The sleep quality data is more subjective. I have been scoring my sleep nightly on a ten-point scale, with over three hundred entries across 2025. A self-reported sleep score is still a self-reported sleep score. I am not wearing an Oura ring. I am not in a sleep lab. So I will tell you what the data showed and what I think it reflects, and I will be honest that this part carries more uncertainty than the blood work.
HbA1c 47 — October 2022. Sleep scoring around 5–6 most nights. The odd 7. Nothing consistently good.
HbA1c 40 — January 2025. Back in normal range. I assumed the sleep would follow quickly. It did not.
The blood sugar number improved first. The sleep repaired roughly six months later — slower than I expected, slower than felt reasonable, honestly. There were weeks in that period where I was eating better than I had in years, fasting regularly, blood tests looking good for the first time in a decade, and still waking at three in the morning. The mechanism I had spent years disrupting was repairing on its own schedule.
The endpoint, for what it is worth: 57 nights rated nine or ten out of ten across the full year of 2025. No sleeping pills. No melatonin. No alcohol as a sleep aid. Just the underlying metabolic environment finally stable enough for the circadian system to do what it is supposed to do.
Fifty-seven nights is not a remarkable achievement. But the contrast with 2022 and 2023 is stark enough that something genuinely shifted. Not overnight. Not in a month. But it shifted.
I will leave that with you.
What I Would Do Differently, Knowing What I Know Now
Not a programme. Not a rigid schedule. Just the order of priority I would change.
The eating window is the upstream variable, not the sleep hygiene. Every piece of sleep advice I read told me to focus on the bedroom environment — temperature, blackout blinds, screen time. That stuff matters. But it is downstream of the metabolic environment. If blood sugar is unstable overnight, no blackout blinds fix the cortisol spike at three in the morning.
Finishing eating by seven in the evening — or earlier if you can manage it — gives the overnight blood glucose profile a chance to settle before the sleep architecture has to run. The research Michael Mosley cites is clear: blood glucose spikes late in the evening directly disrupt deep sleep. Earlier closing of the eating window gives the system longer to stabilise.
Morning light comes next, before anything else. Ten to twenty minutes outside within thirty minutes of waking — ideally without sunglasses — tells the suprachiasmatic nucleus, the master circadian clock, when the day starts. That signal cascades to every organ and sets timing for the next twenty-four hours. Some mornings in Yorkshire the weather does not cooperate. But the mornings I manage it, the evenings are consistently better.
Track the trend, not the night. One bad night means very little. One bad month tells you something. The lag I described — the delay between metabolic improvement and sleep improvement — means you will not see the signal in individual nights. You will see it in the quarterly average. Weekly scoring, four-week rolling average. That is the data worth watching.
If you are starting today, the single thing worth testing first is finishing eating at least three hours before bed, consistently, for four weeks. Not fasting. Not counting anything. Just moving the last meal earlier and noting your sleep quality each morning on a scale of one to ten. Four weeks. See if the trend moves. That is enough data to tell you something real.
Key Takeaways
- Blood markers and sleep architecture run on different clocks: HbA1c improvement is measured in a blood draw. Circadian repair takes months, not weeks.
- The relationship runs both ways: Poor sleep raises blood sugar. Unstable blood sugar disrupts sleep architecture. Breaking the cycle requires addressing the metabolic environment, not just sleep hygiene.
- The bin lorry cannot do its rounds in a disrupted system: Deep sleep waste clearance requires metabolic stability overnight. Fix the upstream, and the downstream repairs itself.
- The two-process model explains the lag: The hourglass fills. But the pendulum — the circadian system — resets more slowly after years of disruption.
- The eating window is upstream of the bedroom: Finishing eating earlier is a more direct intervention than blackout blinds or screen time rules.
- Track the quarterly trend: Individual nights are noise. The four-week rolling average is the signal.
Ready to Track Your Own Pattern?
If this resonated — particularly the gap between what the blood tests show and what the body is actually doing — the 14-Day Rhythm Reset Guide covers the fundamentals I built my system on. Eating windows, sleep timing, what to track and why. It is where I would point anyonThe Belief That Makes the Gap So Confusing
There is a completely reasonable assumption buried in how most health conversations go. Fix the metabolic problem, fix everything downstream. Sort the blood sugar, sort the sleep. It makes intuitive sense, and it is the kind of logic that gets reinforced every time a GP ticks a box and says “levels are improving.”
What that framing misses is that the metabolic environment and the sleep architecture are not the same clock. They are related. They influence each other. But they do not move together.
I held the reasonable assumption for most of 2024. The bloods are improving, the sleep will follow. Month after month, the bloods kept improving and the sleep kept being unreliable. Waking in the small hours. Fragmented. Unrefreshing in a way that is hard to articulate — not dramatic insomnia, just consistently poor quality, like a phone that charges to sixty percent and never quite gets to full.
The assumption was not wrong. It was just incomplete.
What Is Actually Happening When Poor Sleep Breaks Blood Sugar
The mechanism runs in both directions. That is the bit that surprised me, and the bit that most sleep advice never quite gets to.
Start with the direction most people know. Peter Attia cites work from Eve van Cauter’s group at the University of Chicago, confirmed across nine randomised controlled trials. Sleep deprivation of four to five hours — the kind a lot of people run on as a matter of routine — creates insulin resistance equivalent to early type 2 diabetes. Not after months of disruption. Within days.
Benjamin Bikman, a biochemist who communicates this better than almost anyone, puts a number on it. One week of insufficient sleep makes the body roughly thirty percent more insulin resistant. A separate study found that just two days of fifty percent sleep reduction was enough to make otherwise healthy men insulin resistant. Two days.
When you are sleep deprived, cortisol rises. Cortisol mobilises glucose — that is its job, it is a stress hormone preparing your body for effort. Blood glucose goes up. At the same time, the cells that would normally respond to insulin start becoming less sensitive. The combination of elevated cortisol and disrupted hormonal rhythms reduces how well insulin can do its job. The result: higher blood sugar, insulin working harder, the whole system running under increased load.
Sophie Bostock, a sleep scientist whose work I have found consistently rigorous, describes this as a full physiological stress cascade. Cortisol elevated. Blood pressure up. Blood glucose mobilised. Run that cascade night after night and it compounds into immune dysregulation, inflammation, and cardiovascular risk.
Now here is the direction nobody talks about. Michael Mosley cites research from a Cornell sleep lab where they controlled what participants ate and measured the sleep it produced. Diet rich in fibre and protein produced measurably more deep sleep. Diet rich in sugar and simple carbohydrates produced lighter, more fragmented sleep throughout the night. The relationship runs the other way. What you eat shapes the architecture of your sleep.
Which means if your blood sugar has been dysregulated for years — as mine had, HbA1c climbing from 38 in 2015 to 47 by 2022 — the sleep architecture has been deteriorating the whole time. And fixing the blood sugar number is not an instant fix for the sleep. Because the circadian rhythms that govern when you fall asleep, when you drop into deep sleep, when cortisol starts rising in the morning — those patterns do not snap back the moment your HbA1c comes down.
That is the lag. That is the gap. And it is not your fault for not sleeping well once the numbers improve. The mechanism you disrupted has to repair on its own timeline, not yours.
The Bin Lorry That Cannot Do Its Rounds
There is a particular cost to poor deep sleep that I think most people underestimate, partly because it is invisible until it is not.
Deep sleep is when the council bin lorry does its rounds in your brain. All day, your brain accumulates waste — dead cells, metabolic debris, the beta-amyloid proteins associated with Alzheimer’s disease. Deep sleep is when the lorry drives through, clears the lot, hoses down the street. Miss deep sleep — five-hour nights, fragmented sleep, too much alcohol before bed — and the rubbish piles up. Week after week, year after year.
But here is the connection to the metabolic story. If the metabolic environment is disrupted — blood sugar unstable overnight, cortisol elevated — the lorry cannot do its route properly. The bins do not get emptied. Everything backs up.
This is not theoretical risk. Brenda Davis, a registered dietitian with a solid clinical background, makes the point that both high and low overnight blood glucose directly disrupt sleep architecture. High overnight glucose disrupts sleep quality. Low blood glucose in the small hours triggers cortisol release, which wakes you up. The vicious cycle runs: poor blood sugar control worsens sleep, and poor sleep worsens blood sugar control.
Duration does not fix that cycle. Fixing the metabolic environment is what fixes the cycle. And fixing the metabolic environment takes considerably longer than most sleep advice acknowledges.
Why the Two Processes Do Not Sync at the Same Speed
Your sleep is governed by two completely separate systems running simultaneously. Understanding both of them explains why my blood work improved months before my sleep did.
The first is like an hourglass. It fills with sand the longer you are awake — a chemical called adenosine accumulating, building what researchers call sleep pressure. When enough sand has accumulated, you fall asleep. The second is like a pendulum — your circadian rhythm, swinging independently, controlling when those thresholds rise and fall across the 24-hour day.
Both have to align for sleep to arrive easily.
This is why you can be completely exhausted and still not sleep. The hourglass is full — the sleep pressure is there. But the pendulum is at the wrong point in its swing. The circadian system is out of sync, or something upstream — stress, cortisol, blood sugar instability — is interfering with the signal.
What I experienced for months was essentially this: the hourglass was full. I was tired. But the pendulum had been knocked off its rhythm by years of metabolic disruption, and it took time — considerably more time than I expected — for it to find its natural cadence again.
The blood marker improved first because it reflects what is happening in a single blood draw. The circadian system repairs over weeks and months, not days. It is a different clock entirely.
The body also has an anticipatory function that makes this concrete. An hour before your alarm goes off, your muscles are already warming up. Genes are switching on. Metabolic rate rising. Blood pressure preparing to support upright activity. Your body knows you are going to wake up and it is getting ready before you do — like a factory bringing the machines up to temperature before the shift starts.
This warm-up requires the factory to know when the shift starts. Consistent wake time. Consistent signals. If your circadian system has been disrupted — which metabolic dysfunction tends to do, because insulin, cortisol, and blood glucose are all circadian-regulated — the factory does not know when to start warming up. You get out of bed into a body that is not ready. Cold start every morning. And cold starts compound, for longer than feels fair.
What Thirteen Months of Correlation Data Actually Showed
I want to be straight about where my data is strongest and where it carries more uncertainty.
The blood sugar data is solid. NHS records going back to 2012. HbA1c at 38 in 2015, climbing to 47 by 2022, back down to 40 by January 2025, and 37 by April 2025 — documented, GP-reviewed at every stage.
The sleep quality data is more subjective. I have been scoring my sleep nightly on a ten-point scale, with over three hundred entries across 2025. A self-reported sleep score is still a self-reported sleep score. I am not wearing an Oura ring. I am not in a sleep lab. So I will tell you what the data showed and what I think it reflects, and I will be honest that this part carries more uncertainty than the blood work.
HbA1c 47 — October 2022. Sleep scoring around 5–6 most nights. The odd 7. Nothing consistently good.
HbA1c 40 — January 2025. Back in normal range. I assumed the sleep would follow quickly. It did not.
The blood sugar number improved first. The sleep repaired roughly six months later — slower than I expected, slower than felt reasonable, honestly. There were weeks in that period where I was eating better than I had in years, fasting regularly, blood tests looking good for the first time in a decade, and still waking at three in the morning. The mechanism I had spent years disrupting was repairing on its own schedule.
The endpoint, for what it is worth: 57 nights rated nine or ten out of ten across the full year of 2025. No sleeping pills. No melatonin. No alcohol as a sleep aid. Just the underlying metabolic environment finally stable enough for the circadian system to do what it is supposed to do.
Fifty-seven nights is not a remarkable achievement. But the contrast with 2022 and 2023 is stark enough that something genuinely shifted. Not overnight. Not in a month. But it shifted.
I will leave that with you.
What I Would Do Differently, Knowing What I Know Now
Not a programme. Not a rigid schedule. Just the order of priority I would change.
The eating window is the upstream variable, not the sleep hygiene. Every piece of sleep advice I read told me to focus on the bedroom environment — temperature, blackout blinds, screen time. That stuff matters. But it is downstream of the metabolic environment. If blood sugar is unstable overnight, no blackout blinds fix the cortisol spike at three in the morning.
Finishing eating by seven in the evening — or earlier if you can manage it — gives the overnight blood glucose profile a chance to settle before the sleep architecture has to run. The research Michael Mosley cites is clear: blood glucose spikes late in the evening directly disrupt deep sleep. Earlier closing of the eating window gives the system longer to stabilise.
Morning light comes next, before anything else. Ten to twenty minutes outside within thirty minutes of waking — ideally without sunglasses — tells the suprachiasmatic nucleus, the master circadian clock, when the day starts. That signal cascades to every organ and sets timing for the next twenty-four hours. Some mornings in Yorkshire the weather does not cooperate. But the mornings I manage it, the evenings are consistently better.
Track the trend, not the night. One bad night means very little. One bad month tells you something. The lag I described — the delay between metabolic improvement and sleep improvement — means you will not see the signal in individual nights. You will see it in the quarterly average. Weekly scoring, four-week rolling average. That is the data worth watching.
If you are starting today, the single thing worth testing first is finishing eating at least three hours before bed, consistently, for four weeks. Not fasting. Not counting anything. Just moving the last meal earlier and noting your sleep quality each morning on a scale of one to ten. Four weeks. See if the trend moves. That is enough data to tell you something real.
Key Takeaways
- Blood markers and sleep architecture run on different clocks: HbA1c improvement is measured in a blood draw. Circadian repair takes months, not weeks.
- The relationship runs both ways: Poor sleep raises blood sugar. Unstable blood sugar disrupts sleep architecture. Breaking the cycle requires addressing the metabolic environment, not just sleep hygiene.
- The bin lorry cannot do its rounds in a disrupted system: Deep sleep waste clearance requires metabolic stability overnight. Fix the upstream, and the downstream repairs itself.
- The two-process model explains the lag: The hourglass fills. But the pendulum — the circadian system — resets more slowly after years of disruption.
- The eating window is upstream of the bedroom: Finishing eating earlier is a more direct intervention than blackout blinds or screen time rules.
- Track the quarterly trend: Individual nights are noise. The four-week rolling average is the signal.
Ready to Track Your Own Pattern?
If this resonated — particularly the gap between what the blood tests show and what the body is actually doing — the 14-Day Rhythm Reset Guide covers the fundamentals I built my system on. Eating windows, sleep timing, what to track and why. It is where I would point anyone starting out.
The tracker I built for my own sleep quality data — nightly scoring, weekly averages, quarterly trends — that is in the description below.e starting out.
The tracker I built for my own sleep quality data — nightly scoring, weekly averages, quarterly trends — that is in the description below. advice acknowledges.
Why the Two Processes Do Not Sync at the Same Speed
Your sleep is governed by two completely separate systems running simultaneously. Understanding both of them explains why my blood work improved months before my sleep did.
The first is like an hourglass. It fills with sand the longer you are awake — a chemical called adenosine accumulating, building what researchers call sleep pressure. When enough sand has accumulated, you fall asleep. The second is like a pendulum — your circadian rhythm, swinging independently, controlling when those thresholds rise and fall across the 24-hour day.
Both have to align for sleep to arrive easily.
This is why you can be completely exhausted and still not sleep. The hourglass is full — the sleep pressure is there. But the pendulum is at the wrong point in its swing. The circadian system is out of sync, or something upstream — stress, cortisol, blood sugar instability — is interfering with the signal.
What I experienced for months was essentially this: the hourglass was full. I was tired. But the pendulum had been knocked off its rhythm by years of metabolic disruption, and it took time — considerably more time than I expected — for it to find its natural cadence again.
The blood marker improved first because it reflects what is happening in a single blood draw. The circadian system repairs over weeks and months, not days. It is a different clock entirely.
The body also has an anticipatory function that makes this concrete. An hour before your alarm goes off, your muscles are already warming up. Genes are switching on. Metabolic rate rising. Blood pressure preparing to support upright activity. Your body knows you are going to wake up and it is getting ready before you do — like a factory bringing the machines up to temperature before the shift starts.
This warm-up requires the factory to know when the shift starts. Consistent wake time. Consistent signals. If your circadian system has been disrupted — which metabolic dysfunction tends to do, because insulin, cortisol, and blood glucose are all circadian-regulated — the factory does not know when to start warming up. You get out of bed into a body that is not ready. Cold start every morning. And cold starts compound, for longer than feels fair.
What Thirteen Months of Correlation Data Actually Showed
I want to be straight about where my data is strongest and where it carries more uncertainty.
The blood sugar data is solid. NHS records going back to 2012. HbA1c at 38 in 2015, climbing to 47 by 2022, back down to 40 by January 2025, and 37 by April 2025 — documented, GP-reviewed at every stage.
The sleep quality data is more subjective. I have been scoring my sleep nightly on a ten-point scale, with over three hundred entries across 2025. A self-reported sleep score is still a self-reported sleep score. I am not wearing an Oura ring. I am not in a sleep lab. So I will tell you what the data showed and what I think it reflects, and I will be honest that this part carries more uncertainty than the blood work.
HbA1c 47 — October 2022. Sleep scoring around 5–6 most nights. The odd 7. Nothing consistently good.
HbA1c 40 — January 2025. Back in normal range. I assumed the sleep would follow quickly. It did not.
The blood sugar number improved first. The sleep repaired roughly six months later — slower than I expected, slower than felt reasonable, honestly. There were weeks in that period where I was eating better than I had in years, fasting regularly, blood tests looking good for the first time in a decade, and still waking at three in the morning. The mechanism I had spent years disrupting was repairing on its own schedule.
The endpoint, for what it is worth: 57 nights rated nine or ten out of ten across the full year of 2025. No sleeping pills. No melatonin. No alcohol as a sleep aid. Just the underlying metabolic environment finally stable enough for the circadian system to do what it is supposed to do.
Fifty-seven nights is not a remarkable achievement. But the contrast with 2022 and 2023 is stark enough that something genuinely shifted. Not overnight. Not in a month. But it shifted.
I will leave that with you.
What I Would Do Differently, Knowing What I Know Now
Not a programme. Not a rigid schedule. Just the order of priority I would change.
The eating window is the upstream variable, not the sleep hygiene. Every piece of sleep advice I read told me to focus on the bedroom environment — temperature, blackout blinds, screen time. That stuff matters. But it is downstream of the metabolic environment. If blood sugar is unstable overnight, no blackout blinds fix the cortisol spike at three in the morning.
Finishing eating by seven in the evening — or earlier if you can manage it — gives the overnight blood glucose profile a chance to settle before the sleep architecture has to run. The research Michael Mosley cites is clear: blood glucose spikes late in the evening directly disrupt deep sleep. Earlier closing of the eating window gives the system longer to stabilise.
Morning light comes next, before anything else. Ten to twenty minutes outside within thirty minutes of waking — ideally without sunglasses — tells the suprachiasmatic nucleus, the master circadian clock, when the day starts. That signal cascades to every organ and sets timing for the next twenty-four hours. Some mornings in Yorkshire the weather does not cooperate. But the mornings I manage it, the evenings are consistently better.
Track the trend, not the night. One bad night means very little. One bad month tells you something. The lag I described — the delay between metabolic improvement and sleep improvement — means you will not see the signal in individual nights. You will see it in the quarterly average. Weekly scoring, four-week rolling average. That is the data worth watching.
If you are starting today, the single thing worth testing first is finishing eating at least three hours before bed, consistently, for four weeks. Not fasting. Not counting anything. Just moving the last meal earlier and noting your sleep quality each morning on a scale of one to ten. Four weeks. See if the trend moves. That is enough data to tell you something real.
Key Takeaways
- Blood markers and sleep architecture run on different clocks: HbA1c improvement is measured in a blood draw. Circadian repair takes months, not weeks.
- The relationship runs both ways: Poor sleep raises blood sugar. Unstable blood sugar disrupts sleep architecture. Breaking the cycle requires addressing the metabolic environment, not just sleep hygiene.
- The bin lorry cannot do its rounds in a disrupted system: Deep sleep waste clearance requires metabolic stability overnight. Fix the upstream, and the downstream repairs itself.
- The two-process model explains the lag: The hourglass fills. But the pendulum — the circadian system — resets more slowly after years of disruption.
- The eating window is upstream of the bedroom: Finishing eating earlier is a more direct intervention than blackout blinds or screen time rules.
- Track the quarterly trend: Individual nights are noise. The four-week rolling average is the signal.
Ready to Track Your Own Pattern?
If this resonated — particularly the gap between what the blood tests show and what the body is actually doing — the 14-Day Rhythm Reset Guide covers the fundamentals I built my system on. Eating windows, sleep timing, what to track and why. It is where I would point anyone starting out.
The tracker I built for my own sleep quality data — nightly scoring, weekly averages, quarterly trends — that is in the description below.
👉 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)
⚠️ 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.
