Bihormonal iLet Bionic Pancreas (insulin + glucagon) feasibility study

What was tested and why it matters

Most automated insulin systems can only do one thing about a falling glucose: cut back insulin. They can't actively raise glucose. The bihormonal iLet adds a second hormone — automated micro-doses of glucagon, the body's own "raise glucose" signal — so the system can push back against lows instead of just easing off insulin.¹ This study was the first time the dual-hormone iLet was tested in people, and the first multi-day human use of dasiglucagon, a chemically stable glucagon analog that (unlike older glucagon) stays usable in a pump for days.²

The iLet itself is a purpose-built device: it reads a continuous glucose monitor (CGM) and runs a self-learning dosing algorithm that is set up using only the person's body weight, with no carb counting or manual dose settings.¹

Who it's for

Adults with type 1 diabetes who already use a pump and CGM. The 10 participants ranged from 21 to 74 years old, with starting HbA1c levels from 5.7% to 10.6% — a deliberately wide range.¹

Design

This was an open-label, random-order crossover home-use trial sponsored by Massachusetts General Hospital, with Beta Bionics (the iLet) and Zealand Pharma (dasiglucagon) as collaborators.³ Each person used the insulin-only iLet for 7 days and the bihormonal iLet (insulin lispro or aspart plus dasiglucagon at 4 mg/mL) for 7 days, in random order, at the same glucose target (110 mg/dL), with no restrictions on food or exercise.¹ The registry lists it as a Phase 2/Phase 3 feasibility study.³ All ten participants in the published report completed both arms; the registry separately lists twelve enrolled.¹

Key results

Both configurations hit every prespecified engineering target (CGM capture ~89-91%, drug channels available >99% of the time).¹ Descriptive glucose results favored the bihormonal configuration: median time below 54 mg/dL fell from 0.6% to 0.2%, and time below 70 mg/dL from 4.0% to 2.0%.¹ Time-in-range (70-180 mg/dL) rose from 72% to 79%, and mean glucose dropped from 149 to 139 mg/dL.¹ A single prefilled dasiglucagon cartridge lasted the whole week (mean 0.35 mg/day), with no infusion-site reactions or occlusions.¹ Safety was reassuring — no severe lows, no ketoacidosis, no serious adverse events; one bihormonal-arm vomiting episode.¹

What it means / what's next

The short feasibility study reported lower glucose exposure during home use with the dual-hormone configuration, while a practical liquid glucagon survives a week in the pump.¹ Five insulin-leak events at the cartridge connector were traced to off-center septum piercing and prompted a hardware redesign for the next generation; the authors concluded the system was ready for much larger, longer pivotal trials.¹