Smart MDI non-inferior to MDI plus CGM in type 1 diabetes with HbA1c of 8% or more

The study

This report is based on the abstract of a randomised controlled trial published in Diabetes, Obesity & Metabolism. It compared a smart multiple daily injections (MDI) system with MDI plus a continuous glucose monitor (CGM) in people with type 1 diabetes and HbA1c of 8% or more. The abstract says Smart MDI alerts prompt insulin bolus actions, unlike traditional CGM alerts.

After a 3-week run-in, 143 participants were randomised to Smart MDI or MDI plus CGM. The primary endpoint was the between-group difference in time-in-range (TIR) at 6 months, adjusted for baseline glycaemia, tested for non-inferiority with a margin of 7.5%. HbA1c was a secondary endpoint with a non-inferiority margin of 0.4%. Superiority was also tested for both. The authors then did post hoc analyses of "alert-responsiveness" and checked their findings in a real-world cohort of 5507 users.

What they found

At 6 months, the between-group difference in TIR was 0.19 percentage points (95% confidence interval: -4.9 to 5.3). This met the non-inferiority criterion. HbA1c also met non-inferiority (difference -0.13 percentage points, 95% confidence interval -0.46 to 0.19). Neither outcome showed superiority.

In the post hoc analysis, higher responders were those who responded to more than the median response rate of 28.3% of alerts (n = 31). They had greater TIR gains than lower responders (n = 28), by +8.12 percentage points (p = 0.04). Their HbA1c reduction was also greater (-0.6 percentage points; p = 0.02). The authors report that the TIR benefit in higher responders was confirmed in the real-world cohort.

For safety, one episode of diabetic ketoacidosis occurred during the run-in. One severe hypoglycaemic event occurred in the control arm.

How much weight to give it

The main result comes from a randomised comparison with stated non-inferiority margins. That is a reasonably strong design for the question of whether Smart MDI can be used instead of MDI plus CGM. The authors call Smart MDI safe and non-inferior. Two points limit this. Non-inferiority is not the same as better, and the trial found no superiority. Also, only two safety events are reported, so the abstract cannot support firm conclusions about rare harms.

The alert-response result needs more caution. The authors state that these analyses were post hoc. The groups were formed by how people behaved after randomisation, not by random allocation. People who respond to alerts may differ in motivation, engagement or other habits that also improve glycaemia. This shows association, not cause. The group sizes (31 and 28) are small. The real-world cohort supports the TIR finding, but the abstract does not say how that cohort was analysed. The statement that responding to "as few as 28%" of alerts resulted in improvement goes further than these data can show.

The abstract does not report funding, who made the device, blinding, participant age range, baseline HbA1c, or follow-up beyond 6 months. It also does not give the number of participants in each arm or the number of dropouts.

For your clinic

This trial suggests that an MDI system with alerts that prompt bolus actions met non-inferiority to MDI plus CGM for TIR and HbA1c in people with HbA1c of 8% or more. It does not show that the system is better. The post hoc finding links higher alert responsiveness with greater TIR gains and HbA1c falls, but this is an association that needs testing in a prospective trial before it shapes practice.

Before changing any local approach, read the full paper. Check the participant details, the amount of CGM use in the control arm, and how the system was supported. Device choice for any one patient should follow current guidelines and clinical judgement.