Automated insulin delivery in type 2 diabetes: HbA1c down 0.7%, time in range up 16 points

Design

Single-arm, multicentre pivotal trial, 13 weeks

Population

307 adults aged 19-85 with insulin-treated type 2 diabetes

Primary outcome

Change in HbA1c from baseline to 13 weeks

Effect

-0.7% (95% CI -0.9 to -0.6), 8.1% to 7.4%; time in range +16% (14 to 18)

The twiist-T2D pivotal trial enrolled 307 adults aged 19 to 85 with insulin-treated type 2 diabetes at multiple US centres. After baseline CGM, each used the twiist AID system for 13 weeks. There was no control group.

HbA1c fell from 8.1% to 7.4% (change -0.7%, 95% CI -0.9 to -0.6), with the largest fall in those starting at 9.0% or above. Time in range 70-180 mg/dL rose from 57% to 73%. HbA1c also fell in people who stayed on a GLP-1 receptor agonist or SGLT2 inhibitor, so AID added to rather than replaced them. Hypoglycaemia stayed low, there was no severe hypoglycaemia, and one person had diabetic ketoacidosis. Satisfaction and sleep quality improved.

Closed-loop systems have so far been a type 1 diabetes tool. This trial supports considering AID for people with type 2 diabetes on multiple daily injections who remain above target despite good engagement. Cost and device access will limit this in India, and the result is against baseline rather than a randomised comparator.

• Consider AID for insulin-treated type 2 diabetes when HbA1c stays high despite adherence to injections.
• The largest benefit was seen at baseline HbA1c of 9.0% or more.
• GLP-1 receptor agonists and SGLT2 inhibitors were continued alongside AID in the trial.
• Teach ketone checking and pump-failure plans, especially for anyone also taking an SGLT2 inhibitor.
• Patients need a working CGM and the ability to manage pump supplies before starting.

Why it matters

It widens the group who may be offered closed-loop insulin beyond type 1 diabetes.

Don't overread it

Single-arm and 13 weeks long: part of the improvement may reflect closer trial follow-up rather than the device.

The statistics, in plain English

The HbA1c confidence interval (-0.9 to -0.6) is narrow and well below zero, so the fall from baseline is real. Without a control group, though, it is a before-and-after change, and some of it could come from the attention every trial participant receives.

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.

Get Daily Doctor — free

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