Summary: In a small pilot outpatient study in adults with type 1 diabetes, a fully closed-loop insulin-and-pramlintide system that needed no carbohydrate counting achieved time in range similar to standard carbohydrate-counted dosing, with numerically but not significantly more hypoglycaemia.
PICO Summary
| Element | Detail |
|---|---|
| Population | 12 adults with type 1 diabetes (mean age 39.5, HbA1c 7.4%); pilot, randomised, controlled outpatient study with 14-hour supervised arms, Canada. |
| Intervention | Fully closed-loop insulin and pramlintide without carbohydrate counting (faster aspart or aspart, at 8 or 10 µg/U pramlintide-to-insulin ratios). |
| Comparison | Faster aspart with carbohydrate counting (control). |
| Outcome | Median time in range was 78.6% (control) versus 76.2% and 78.8% for faster-aspart-plus-pramlintide (8 and 10 µg/U), and 65.9% and 77.4% for aspart-plus-pramlintide. Time below 3.9 and 3.0 mmol/L was numerically higher with the closed-loop arms. None of the differences were statistically significant. |
Closed-loop insulin-pramlintide, no carb counting
Pilot RCT · type 1 diabetes · 14-hour arms
Without any carb counting, the closed-loop insulin-pramlintide system reached a median time in range (78.8%) comparable to carb-counted control (78.6%), but with a non-significant trend toward more hypoglycaemia in this 12-patient pilot.
Expert Commentary
This is an early proof-of-concept study pursuing one of the field’s most desired goals, a fully closed-loop system that removes the burden and error of carbohydrate counting, and it should be read as feasibility rather than evidence of efficacy. The rationale is elegant: pramlintide slows gastric emptying and suppresses glucagon, buying time for slow-acting insulin to work reactively after a meal, which is exactly the cushion a no-announcement algorithm needs. The encouraging signal is that time in range without any carb counting was broadly comparable to carb-counted control. Two cautions dominate interpretation, however. First, this is a pilot of just 12 adults over very short supervised periods, and none of the differences reached statistical significance, so it cannot establish equivalence. Second, the numerically higher time below range is a safety signal that, while not significant here, fits pramlintide’s known hypoglycaemia risk and must be resolved in larger trials, alongside the practical challenges of dual-hormone delivery. Can I use this with my patients? Not yet; no such system is available, and pramlintide remains an underused adjunct rather than part of a closed loop. I would frame this honestly as a promising research direction and point interested patients toward clinical trials rather than any change in current care.
References
Odabassian M, Tsoukas MA, Cohen E, Pasqua MR, Rutkowski J, Haidar A. A pilot outpatient assessment of a fully closed-loop insulin and pramlintide system. J Diabetes Sci Technol. 2025;19(6):1457–1463. doi:10.1177/19322968251371046