Design: Phase 2b, randomised, double-blind, placebo-controlled, 19 centres
Population: 141 people aged 6–35 years with type 1 diabetes, insulin for under 3 months
Primary outcome: Change in stimulated C-peptide AUC at 12 months
Effect: Geometric mean ratio 1.06 (95% CI 0.56–2.01) versus placebo; p=0.85

DIABIL-2 randomised 141 people aged 6 to 35 years within three months of starting insulin for type 1 diabetes to low-dose interleukin-2, given fortnightly or weekly after a five-day induction, or placebo, for almost a year.

Regulatory T cells expanded sharply in the IL-2 groups, confirming the drug did what it was meant to do biologically. But stimulated C-peptide at 12 months did not differ from placebo (geometric mean ratio 1.06, 95% CI 0.56 to 2.01), in either regimen or in children or adults. Serious adverse events were not more frequent with IL-2.

This is a clean negative result for IL-2 as a single agent at diagnosis. The investigators suggest earlier, presymptomatic stages or combinations may still be worth testing. For now, teplizumab in stage 2 disease remains the only immune therapy with evidence of delaying type 1 diabetes, and outside trials nothing alters insulin management at diagnosis.

Offer trial participation to newly diagnosed people who ask about immune therapy; nothing outside trials preserves β-cell function yet.
Explain that raising regulatory T cells alone did not slow β-cell loss after diagnosis.
Keep insulin education and early CGM as the core of care in the first year after diagnosis.
Consider islet autoantibody screening of first-degree relatives where a staging and prevention pathway exists.

Why it matters

Hitting the immune target is not the same as protecting the β cell, and this trial shows the gap plainly.

The statistics, in plain English

A geometric mean ratio of 1.06 means C-peptide was about the same in both groups; the interval from 0.56 to 2.01 is wide because the trial was modest in size, but it is centred on no effect.