A once-weekly insulin matched daily glargine with fewer lows and less titration

A once-weekly basal insulin given at fixed doses controlled blood sugar about as well as a daily injection in adults with type 2 diabetes starting insulin for the first time, and did so with fewer dangerous low-sugar episodes and far less dose-adjusting, according to the phase 3 QWINT-1 trial [s1]. Over 52 weeks, glycated haemoglobin fell 1.19 percentage points on weekly efsitora and 1.16 on daily insulin glargine — a difference of 0.03 points, well inside the margin set for calling the two equivalent [s1].

The appeal here is not a bigger number but a simpler regimen. Basal insulin has always demanded frequent self-adjustment: patients test fasting glucose and nudge the dose up or down, often weekly, a burden that keeps many from starting insulin or from using it well. Efsitora is engineered for a once-weekly injection on a fixed schedule of preset doses, and QWINT-1 was built to test whether that simplicity costs anything in control [s1].

What the trial did

The 52-week, open-label, treat-to-target trial, funded by Eli Lilly, randomised 795 adults with type 2 diabetes who had never used insulin to once-weekly efsitora or once-daily glargine U100 [s1]. Efsitora was started at a single 100-unit weekly dose, with adjustments made only every four weeks and only among fixed steps of 150, 250, and 400 units, aiming for fasting glucose of 80 to 130 mg per decilitre; glargine was titrated weekly or more often by a standard algorithm toward the same target [s1]. The primary endpoint was the change in HbA1c at 52 weeks, tested for noninferiority against a margin of 0.4 percentage points [s1].

What it found

HbA1c fell from 8.20% at baseline to 7.05% at week 52 on efsitora (least-squares mean change 1.19 percentage points) and from 8.28% to 7.08% on glargine (1.16 points); the between-group difference of 0.03 percentage points (95% CI 0.18 to 0.12) confirmed noninferiority, and efsitora was not superior (P=0.68) [s1]. So on the primary measure the two are a statistical tie.

The differences showed up around that tie. The combined rate of clinically significant hypoglycaemia (glucose below 54 mg per decilitre) or severe hypoglycaemia was 0.50 events per participant-year on efsitora against 0.88 on glargine, an estimated rate ratio of 0.57 (95% CI 0.39 to 0.84) [s1]. At week 52 the mean total insulin dose was 289.1 units per week on efsitora and 332.8 on glargine, a difference of 43.7 units per week (95% CI 25.0 to 62.4) [s1]. And the median number of dose adjustments over the year was 2 with efsitora against 8 with glargine — the clearest expression of the regimen's central claim [s1].

How to read it

This is a well-sized phase 3 trial reporting exactly what it set out to test, and the result is honestly a modest one: equal sugar control, achieved with a fixed weekly dose that most patients adjusted twice a year instead of eight times, at a lower rate of the low-glucose episodes that make insulin frightening to start [s1]. It is worth being precise about the claim. The primary endpoint was tested for noninferiority against a margin of 0.4 percentage points, not for superiority, and superiority was formally not shown (P=0.68) [s1]. Noninferiority means efsitora did not do meaningfully worse on blood-sugar control — a lower bar than beating glargine — so the case for the weekly drug rests on the secondary advantages in hypoglycaemia, total dose, and titration burden, not on tighter control [s1].

Two caveats sit alongside that. The trial was open-label — patients and clinicians knew which drug they were on — which can colour soft outcomes like how aggressively glargine was titrated, though HbA1c and measured hypoglycaemia are hard to bias. And it was funded by the manufacturer, so the independent replication that matters will come from the rest of the QWINT programme and from real-world use. A fixed-dose regimen also cuts both ways: the same rigidity that removes the titration chore removes the flexibility to fine-tune, which is why the endpoint that matters next is how the design performs in people who need larger or more variable doses [s1].

Efsitora is not the first weekly insulin: insulin icodec cleared the same insulin-naive hurdle in the ONWARDS 1 trial [s2] and has already reached the market, covered on this site in the approval of once-weekly Awiqli. What QWINT-1 adds is a fixed-dose design that removes most of the titration entirely. For readers weighing what the underlying number means, what HbA1c actually measures sets the context, and the opposite end of the diabetes-technology spectrum — automated minute-to-minute dosing — is covered in closed-loop insulin delivery.

What to watch

The open questions are whether the whole QWINT set of trials — in insulin-experienced patients and in type 1 diabetes, where fixed weekly dosing is riskier — holds the same balance, how the once-weekly insulins are priced against daily generics given the long-standing cost problem in insulin, and whether the reduced titration burden actually gets more people to start and stay on insulin in practice, which is the benefit the trial can only imply.

This article describes research and is not medical advice. Decisions about diabetes treatment and insulin are for patients and their treating clinicians.

Sources

• Weekly Fixed-Dose Insulin Efsitora in Type 2 Diabetes without Previous Insulin Therapy (QWINT-1) — New England Journal of Medicine, published online 22 June 2025
• Weekly Insulin Icodec versus Daily Insulin Glargine U100 in Type 2 Diabetes without Previous Insulin Treatment (ONWARDS 1) — New England Journal of Medicine, 24 June 2023