Baricitinib's beta-cell protection in type 1 diabetes faded within a year of stopping

BANDIT was a randomised, double-blind, placebo-controlled trial in recent-onset type 1 diabetes in which 48 weeks of oral baricitinib preserved beta-cell function and lowered insulin requirements. This report follows what happened after the drug was stopped. Of 91 randomised participants, 88 (58 on baricitinib, 30 on placebo) reached the week 96 assessment, which measured C-peptide and glucagon responses to a mixed meal, HbA1c, CGM (continuous glucose monitoring) measures, and CD8+ T-cell phenotype and function.

C-peptide was still higher in the baricitinib group at week 72 (mean plus or minus SEM 0.54 plus or minus 0.05 versus 0.38 plus or minus 0.06 pmol/mL; p = 0.015), but the gap had closed by week 96 (0.43 plus or minus 0.05 versus 0.35 plus or minus 0.06; p = 0.336). There were no differences in insulin dose, HbA1c or CGM measures at any point during follow-up. The immune changes disappeared alongside the clinical ones: the reduced frequency and cytokine signalling of effector memory CD8+ T cells seen at week 48 had resolved by week 96. A post hoc split by age found that adults (n = 28) kept their beta-cell function fully after stopping, while children (n = 32) lost C-peptide by 0.09 pmol/mL relative to baseline (p = 0.022). Baricitinib also did not correct the paradoxical rise in glucagon after a mixed meal.

The authors' own conclusion is the honest one: durable benefit is likely to require continuous treatment. That reframes baricitinib in type 1 diabetes from a short immunological reset into an indefinite therapy, with all the safety, monitoring and cost questions that follow. It remains research-stage - baricitinib is not licensed for type 1 diabetes - and today's practical instruction is to be able to say this clearly to a family who has read about the original result, rather than to look for a way to prescribe it.

• If a family asks about baricitinib for new-onset type 1 diabetes, explain that a fixed course does not appear to give lasting benefit.
• Do not prescribe baricitinib off-label for type 1 diabetes on the strength of the original 48-week result.
• Measure C-peptide in recent-onset type 1 diabetes where you can; it is the outcome these trials turn on and it anchors the conversation.
• Treat the adult-versus-child difference as hypothesis-generating, not as a way to select patients.
• Keep the focus on what does have durable evidence: insulin, structured education and CGM access.

The statistics, in plain English

The week 96 p value of 0.336 does not prove the groups were the same - with 88 people in total, this follow-up could not rule out a modest lasting benefit even if one existed. Absence of a significant difference is not evidence of no difference, and the honest statement is that any remaining effect was too small for this trial to see. The results are also given as mean plus or minus SEM, the standard error of the mean, which makes the groups look tighter than they are; SEM shrinks as the sample grows, whereas the spread between individual patients does not. The adult-versus-child split was decided after the results were in, in groups of 28 and 32, which is exactly the size and timing of subgroup finding that most often fails to replicate.

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.