DailyDoctor Archive Specialties Get app
Back to the 1 September 2026 edition

Research · 05 of 07

Once-weekly efsitora's hypoglycaemia looks like daily basal insulin's

Hypoglycaemic episodes on once-weekly efsitora lasted about as long as on daily basal insulin, addressing the main concern about weekly dosing.

The worry about a once-weekly basal insulin is not whether it lowers glucose but what happens when it goes too low — with a week of insulin already administered, an episode cannot be undone by skipping tomorrow's dose. This exploratory analysis characterises those episodes across QWINT-2, QWINT-3 and QWINT-4, using masked continuous glucose monitoring for duration and both investigator reports and a pre-specified algorithm for recurrence.

Episodes were not longer. Level 1 and 2 events (below 70 mg/dL) lasted 40 to 42.5 minutes on efsitora against 40 minutes on the once-daily comparators; level 2 events (below 54 mg/dL) lasted 35 to 39.9 minutes against 35. Persistent-recurrent hypoglycaemia was uncommon in both. The contributing factors participants reported, and how they treated episodes, did not differ noticeably between groups.

Why it matters: this is the specific fear about weekly insulin, examined directly rather than inferred from event counts. Durations within a few minutes of daily comparators is a reassuring answer to it.

What to do about it: nothing changes today for prescribing, and this is an exploratory analysis rather than a powered comparison — it describes episodes, it does not test a hypothesis about them. But it is the analysis to have read when a patient asks what happens if a weekly insulin takes them low.

  • Addresses the specific concern about weekly basal insulin: episode duration, not frequency
  • Exploratory and descriptive — no hypothesis was tested, so read the numbers as characterisation
  • Masked CGM measured duration, which is more reliable than participant recall
  • Type 2 diabetes only; nothing here speaks to type 1
  • Usual hypoglycaemia advice is unchanged

The statistics, in plain English

These are medians with no confidence intervals and no significance tests, because the analysis set out to describe rather than to compare. That is a real limitation and the authors say so: a difference of a few minutes between 40 and 42.5 could be noise, a genuine small difference, or an artefact of how episodes were bounded. What the numbers do establish is the absence of anything dramatic — a weekly insulin producing episodes lasting hours rather than minutes would have been obvious here, and it did not.

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.

QR code to install Daily Doctor
Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.

obesityglp1type1insulinpediatrictechnologytype2hypertension

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
Daily Doctor All 27 specialties, every morning. Free.
Get the app