45 minutes at level 2: why I deliberately ride slowly in the mornings

The starting position: the value is already up in the morning

My day rarely begins at a low value. Between four and eight in the morning blood sugar rises, because cortisol and growth hormone prepare the body for waking and the liver releases more glucose during that window.³ My value is then typically somewhere between 115 and 145 mg/dl, without my having eaten anything.

Then comes the protein shake, and then I get on the bike. This sequence developed over years, and the reason for it is not ambition. It is about not letting the morning run away – but in a way where the correction does not become worse than the problem.

The protocol

The sequence is deliberately rigid. Get up, read and note the value, protein shake of 25 grams of powder in 300 millilitres of unsweetened almond drink, then 45 minutes of low impact riding on the Peloton bike with resistance at level two. Afterwards I record what the sensor did.

I almost always ride the same classes, mostly with Maya Wedekind or Charlotte Weiderbach. That is not advertising but method: if I want to compare the results of two mornings, as little as possible may differ. A different class means a different pace, different intervals, a different load curve – and therefore a different result whose cause I can no longer pin down.

Why 45 minutes and not 20

Shorter sessions lower my value too little to really catch the morning rise. After 20 minutes I see a dip, but the trend often turns upward again afterwards because the liver keeps supplying. Somewhere between about 40 and 50 minutes lies the range in which my value stays reliably in target without my driving it into the basement.

There is a second, unspectacular reason: 45 minutes fit into my day. A protocol that only works on good days does not produce comparable data. The best session length is the one you still complete after a night-time call-out.

Two mornings I documented in full

On 22 July I recorded a complete morning. Starting value 119 mg/dl at 08:19, protein drink at 08:22, start of the ride at about 08:27. The value initially kept rising – 131 at 08:30, 143 at 08:37, peaking at 149 mg/dl at 08:46. Only then did it turn: 121 at 09:11, 118 at 09:24 after the cool-down.

What is notable: the rise initially continued despite the training being under way. The movement did not prevent it, it caught up with it. At the end the value sat slightly below the starting value – a successful morning from my point of view, but not for the reason I had assumed beforehand.

Three days later, on 25 July, I rode at midday. Starting value about 143 mg/dl, after the session 87 mg/dl at 12:32 with a falling arrow. A drop of roughly 56 points in a single session. Back then I considered that a good result. Today I read it differently: training that pushes the value that far is no longer working against a meal, it is working against my reserves.

The pattern this is really about

The actual finding of my recent years lies not in a single curve but in the repetition: the deeper the value falls during a session, the more forcefully the rebound comes afterwards in my case.

Physiologically this is well explicable. When blood sugar falls towards 70 mg/dl, the body reads it as an alarm. Glucagon rises, with a steeper fall adrenaline joins in, and the liver releases glucose.¹ This protective mechanism does not ask whether the fall was intended. If intense exertion is added, which raises the value via adrenaline anyway⁴, two effects add up.

A calm ride, by contrast, consumes glucose without triggering the alarm system. The muscle takes it up via GLUT-4 transporters, independently of the insulin signal² – but slowly enough that counter-regulation does not kick in. That is the whole trick, and it is unspectacular.

What I experienced with more intense sessions

I tried it with classic rides and higher resistance levels, sometimes level four instead of two. The immediate effect was impressive: the value fell faster and further. The effect across the whole morning was not. The rebound caught up with the gain and in several cases went beyond it – at the end the value stood higher than before the training.

I have to be honest about this: I did not measure those intense sessions systematically under the same conditions as my current protocol. What I describe is a repeated impression from many mornings, not a measurement series with a clean comparison. That is why the graphic above is explicitly labelled as a schematic.

Why the shake comes before the ride and not after

This is the question I am asked most often. The common recommendation is to train fasted and eat afterwards. In my case the opposite has proved itself, and the reason lies in what I described above.

If I ride completely fasted, the value falls further – and that is precisely when counter-regulation kicks in. At around four grams of carbohydrates the shake is not a meal that raises the value appreciably. It works more like a signal to the body that something is on its way. Whether that effect really exists or whether I am rationalising, I cannot prove. What I do see: the mornings with a shake run more calmly than the ones without.

A second reason is more mundane. The shake is the one fixed point on which I hang the morning. Without it the sequence shifts, and then two days are no longer comparable. A protocol lives on being boring.

Morning or midday – two different starting positions

It helps to be clear that the same session at different times of day does not mean the same thing. In the morning the liver’s glucose release runs at a higher level anyway, and insulin action is at its weakest point of the day. In the afternoon both are more relaxed.

That explains why my midday sessions lower the value more markedly than the morning ones. On 25 July it fell at midday from about 143 to 87 mg/dl – a drop I never see in that form in the morning. Anyone measuring their morning protocol against afternoon values is therefore comparing apples with pears. It took me a while to understand that.

What sleep has to do with it

As a volunteer firefighter I regularly have nights interrupted by a call-out. On the mornings that follow, my curve looks different: the starting value sits higher and the session works less well. That fits the physiology – lack of sleep raises cortisol, and cortisol drives precisely the glucose release I ride against in the morning.

I did not conclude from this that I should skip training, but that I should expect something different. After a night-time call-out I ride the same 45 minutes but do not count on the same result. That sounds like a detail, yet it prevents the reaction that used to follow: ride harder because the value is not coming down. That is exactly how the rebound arose that I want to avoid.

How I measure and what I note

The sensor is a FreeStyle Libre 3. What gets noted is the starting value before the shake, the product with its amount, start and end of the session as physical activity with its duration, and the value after the cool-down. All directly in the app, not later from memory.

What I learned along the way: the spot readings are more reliable than the daily curve. In the twelve-hour view the display visibly smooths the peaks away – a peak showing 183 mg/dl as a spot reading appears in the daily overview as about 160. Anyone looking only at the curve underestimates both extremes.

Four mistakes I made

• Rode too hard because the value would not fall. The obvious reaction – and the one that produces the rebound.
• Measured only right after training. The value looks good then. What happens an hour later stays invisible.
• Switched classes. Different intervals, different result – and no way left to say what caused it.
• Credited the training with the progress. My weight changed over the same period. Which of the two is working, I do not know.

What the protocol does not achieve

It makes the morning more predictable, but it does not solve the underlying problem. My fasting value still sits higher than it should. The morning glucose release from the liver can be dampened by a calm ride, not switched off. I am not finished at this point, and I do not want to pretend otherwise.

What has changed in parallel is the overall picture: resting heart rate and weight have fallen over time, and the values as a whole sit calmer than they used to. How much of that is down to the training and how much to the weight I cannot separate. Both run at once, and I have no control group – I only have myself.

What I will test next

The obvious counter-test is still missing: the same morning, the same starting position, but a markedly more intense session – documented in full rather than recalled. Until that exists, the core of this article remains a reasoned assumption with matching physiology, no more.

Alongside that I want to record weight and resting heart rate daily, so that the question of cause can eventually be answered better. And I want to test whether the 45 minutes really are the best compromise or whether 30 minutes achieve the same. That can be measured – so far I have not done it.