In menstruating patients with type 1 diabetes, insulin sensitivity is not constant across the month, and automated systems do not yet know that. A decentralised study of 77 adults using automated insulin delivery, contributing 380 menstrual cycles of real-world data, quantified it: insulin sensitivity was about 2.6% higher in the early follicular phase (credible interval 0.3 to 5.0) and about 2.6% lower in the mid-luteal phase (-5.2 to -0.1). Total daily dose and carbohydrate intake both peaked in the luteal phase, alongside higher mean glucose and less time in range.
The useful part is not the average — it is that 84.7% of individuals followed the pattern while the size of the swing varied enormously between them. That makes it a question to ask, not a rule to apply.
So when a menstruating patient's download shows an unexplained bad fortnight, before you change carbohydrate ratios or basal profiles, ask when in the cycle it happened. If the pattern repeats month on month, the fix is a phase-aware adjustment for that week — a temporarily stronger correction factor or a higher automated target override during the luteal phase — not a permanent change that will then run too hot in the follicular phase.
- Ask menstruating patients to note cycle day 1 in the pump or app, so downloads can be read by phase
- Before changing settings on a bad two weeks, check the same fortnight in the previous one or two cycles
- If the luteal dip repeats, use a temporary phase-specific adjustment rather than a permanent ratio change
- Expect the size of the effect to differ hugely between patients — the population average of 2.6% understates what some individuals experience
- Include this in the first education session for adolescent girls starting a pump, not as an afterthought years later
The statistics, in plain English
The population-level swing is small and the credible intervals only just exclude zero, so nobody should manage a patient on the 2.6% figure. The finding that carries clinical weight is the individual-level one: 84.7% of participants moved in the same direction as the group, which makes cycle phase a reliable thing to look for in a given person even though its size cannot be predicted in advance. This is observational, real-world donated data, so it describes what happens rather than proving a mechanism.
Read the rest in the app
You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.
Scan to keep reading on your phone. No account needed to start.
Tomorrow morning, before your first patient
One edition a day for diabetes & endocrinology, written by the desk, every claim tied to its paper. Six minutes.
Get the app — free