COMBINE 4 was a 40-week, open-label, treat-to-target phase 3b trial across 97 sites in nine countries, funded by the manufacturer. It randomised 485 insulin-naive adults with type 2 diabetes and an HbA1c of 8.0% or above, all already on oral glucose-lowering drugs, to either IcoSema (a once-weekly fixed combination of basal insulin icodec and semaglutide) or once-daily insulin glargine U100. Median age was 58 years. Both arms were titrated to a fasting target of 3.9-5.0 mmol/L (70-90 mg/dL).
Starting from a baseline HbA1c of about 9.5%, the weekly combination fell 3.32 percentage points against 2.44 with glargine, a difference of 0.88 percentage points (95% CI -1.12 to -0.63). Weight moved in opposite directions: down 0.79 kg with IcoSema, up 3.81 kg with glargine, a gap of 4.61 kg (95% CI -5.46 to -3.75). Clinically significant or severe hypoglycaemia was less frequent too, at 0.29 against 0.59 episodes per person-year (rate ratio 0.56, 95% CI 0.32-0.97). Gastrointestinal side effects were the commonest problem with the combination.
This is the scenario that fills clinics: someone on maximal oral therapy with an HbA1c in the nines, facing their first injection. The usual bargain has been better glucose in exchange for weight gain and hypoglycaemia risk. Here that bargain does not have to be struck, and the injection is weekly rather than daily. Two honest caveats: the trial was open-label, which matters most for how side effects were reported, and availability and price will decide whether this changes anything in Indian practice, where the same combination taken as separate basal insulin and a GLP-1 (glucagon-like peptide-1) receptor agonist is already an option for those who can afford it.
- Identify patients on maximal oral therapy with HbA1c above 8% who have refused insulin because of weight gain or injection frequency; this is the group the trial addresses
- Counsel on gastrointestinal symptoms at initiation, as with any GLP-1 receptor agonist, and titrate against fasting glucose rather than symptoms
- Do not extrapolate to patients already on basal insulin; this trial enrolled insulin-naive patients only
- Check whether a separate basal insulin plus GLP-1 receptor agonist is available and affordable now, rather than waiting for the fixed combination
- Reassess sulfonylureas at the point of starting any GLP-1-containing regimen, since much of the residual hypoglycaemia risk sits there
The statistics, in plain English
The 0.88 percentage point difference in HbA1c has a confidence interval of -1.12 to -0.63, entirely on the side of benefit, so the direction is secure even at the least favourable end. The weight difference of 4.61 kg is mostly the glargine arm gaining rather than the combination arm losing, which is worth saying plainly to patients. The hypoglycaemia rate ratio of 0.56 has an interval of 0.32 to 0.97 that only just clears 1.0, and the p value of 0.04 reflects that; read it as a real but imprecisely sized reduction rather than a firm halving. This was a treat-to-target design, meaning both arms were pushed towards the same fasting glucose, so the HbA1c gap reflects what the drugs do beyond fasting control, largely after meals.
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