This is the post-treatment phase of BANDIT, a randomised, double-blind, placebo-controlled trial of oral baricitinib given for 48 weeks in recent-onset type 1 diabetes. Of 91 people randomised, 88 (58 on baricitinib, 30 on placebo) were followed to week 96, which is 48 weeks after the drug was stopped. The team measured C-peptide and glucagon responses to a mixed meal, HbA1c, continuous glucose monitoring measures, and the phenotype and function of CD8+ T cells.
C-peptide was still higher in the baricitinib group at week 72 (0.54 plus or minus 0.05 versus 0.38 plus or minus 0.06 pmol/mL, p = 0.015), but by week 96 the difference had gone (0.43 versus 0.35, p = 0.336). Insulin dose, HbA1c and continuous glucose monitoring measures never separated during follow-up. The immunological signal reversed in step: the reduced frequency and cytokine signalling of effector memory CD8+ T cells seen at week 48 had resolved by week 96. Baricitinib also failed to correct the paradoxical rise in glucagon after a meal that characterises type 1 diabetes.
One split is worth noting and not over-reading. In a post hoc comparison, adults who stopped at week 48 (n = 28) fully preserved beta-cell function, while children (n = 32) lost 0.09 pmol/mL of C-peptide relative to baseline (p = 0.022). That is a small, unplanned analysis and should be treated as a hypothesis rather than a finding. The message that does hold is that this is suppressive treatment rather than a cure. Durable benefit will need continuous therapy, and that changes the risk-benefit conversation entirely when the patient is a child facing decades of a drug rather than a single course.
- Describe beta-cell preservation therapies to patients as ongoing treatment, not a fixed course.
- Expect the benefit to disappear within about a year of stopping.
- Do not quote the adult versus child difference as established - it is a post hoc split of 60 people.
- Note that HbA1c and continuous glucose monitoring metrics never separated even while C-peptide did.
- Anyone finishing an immunotherapy trial needs explicit counselling about what happens after the drug stops.
The statistics, in plain English
A p value of 0.336 at week 96 does not prove the groups are identical, but with a real difference of only 0.08 pmol/mL against a background of considerable variation between people, there is nothing here a clinician could act on. The week 72 result, p = 0.015, is the more informative one because it shows the benefit was still measurable a full six months after the drug stopped and then decayed - the pattern is a fade, not an abrupt loss. The child versus adult comparison carries the weakest evidence in the paper: it was not planned in advance, it splits an already small trial into groups of about 30, and unplanned subgroup results are wrong far more often than their p values suggest.
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