Clinical Context
Current automated insulin delivery (AID) systems—often called “artificial pancreas” technology—have dramatically improved glycemic control in type 1 diabetes, but they still require user input: patients must count carbohydrates and announce meals to trigger meal-time insulin boluses. This reliance on carbohydrate counting remains a significant burden. Carb counting is imprecise (studies show 50% estimation errors are common), and missed or late meal announcements result in postprandial hyperglycemia.
The holy grail of diabetes technology is a “fully closed-loop” system that requires no user input—a truly automated pancreas that detects meals from glucose excursions and responds appropriately. However, insulin’s slow pharmacokinetics (onset 15-30 minutes, peak 60-90 minutes for rapid analogs) make reactive dosing challenging: by the time glucose rises enough to trigger insulin delivery, it’s often too late to prevent significant hyperglycemia.
Pramlintide, a synthetic analog of amylin (a hormone co-secreted with insulin), offers a potential solution. Pramlintide slows gastric emptying, suppresses postprandial glucagon, and increases satiety—effectively delaying and blunting glucose absorption. When combined with insulin, pramlintide may provide the pharmacokinetic “cushion” needed for fully closed-loop control without meal announcement.
Study Summary (PICO Framework)
Summary:
In adults with type 1 diabetes, a fully automated closed-loop insulin and pramlintide system without carbohydrate counting maintained comparable time in target glucose range (~79%) compared to standard hybrid closed-loop with carbohydrate counting, though it was associated with numerically higher (but not statistically significant) hypoglycemia rates.
| PICO | Description |
|---|---|
| Population | 12 adults with type 1 diabetes (mean age 39.5 years, HbA1c 7.4%). |
| Intervention | Fully closed-loop insulin and pramlintide delivery without carbohydrate counting (8 µg/U and 10 µg/U pramlintide-to-insulin ratios). |
| Comparison | Standard faster aspart insulin with carbohydrate counting (control). |
| Outcome | Time in range (3.9-10.0 mmol/L): 78.8% with closed-loop vs 78.6% with control. Higher hypoglycemia frequency with closed-loop, but not statistically significant. |
Clinical Pearls
1. Eliminating carb counting with comparable time-in-range is a major proof-of-concept. Achieving ~79% time in range without meal announcements matches what current hybrid closed-loop systems achieve with carbohydrate counting. This suggests that pramlintide’s glucagon suppression and delayed gastric emptying can compensate for the lack of anticipatory insulin bolusing.
2. Pramlintide’s mechanisms are ideal for closed-loop synergy. By slowing glucose appearance (delayed gastric emptying) while insulin works to increase glucose disappearance, pramlintide effectively “buys time” for the algorithm to respond. The glucagon suppression also reduces hepatic glucose output, further smoothing the glucose profile.
3. The hypoglycemia signal deserves attention. Although not statistically significant in this small pilot, the numerically higher hypoglycemia rates with the fully closed-loop system are concerning. Pramlintide itself can cause hypoglycemia when combined with insulin, and the algorithm may over-deliver insulin reactively. Larger studies must carefully evaluate this safety signal.
4. Dual-hormone delivery is technically challenging. This system requires two infusion sets (or a dual-chamber pump), doubling potential site failures and user burden. Pramlintide also has stability and compatibility issues that complicate pump delivery. These engineering challenges must be solved for commercial viability.
Practical Application
Current implications for practice: This is a pilot study (n=12) and not yet clinically applicable. No fully closed-loop insulin-pramlintide system is commercially available. However, the findings inform the direction of diabetes technology development and may influence clinical trial participation opportunities.
Pramlintide use in current practice: Pramlintide (Symlin) is FDA-approved as adjunctive therapy for type 1 and type 2 diabetes but is underutilized due to injection burden (separate from insulin), GI side effects, and hypoglycemia risk. Patients using current AID systems cannot easily add pramlintide, though some have experimented with off-label use. If considering pramlintide, reduce meal-time insulin by 50% initially and titrate based on response.
For patients interested in advanced technology: Explain that fully closed-loop systems remain investigational. Current options (Medtronic 780G, Omnipod 5, Tandem Control-IQ) are “hybrid” closed-loop requiring meal announcements. Patients who find carbohydrate counting burdensome should know that simpler meal announcement strategies (e.g., fixed doses for “small/medium/large” meals) can reduce burden while maintaining reasonable control.
Clinical trial opportunities: Patients interested in cutting-edge technology should be encouraged to explore clinical trial participation. Fully closed-loop studies are ongoing at multiple academic centers, and participation helps advance the field while providing access to investigational systems.
How This Study Fits Into the Broader Evidence
The development of automated insulin delivery has progressed from threshold-suspend systems (stopping insulin at low glucose) to hybrid closed-loop (automated basal adjustment plus meal boluses) to the current frontier of fully closed-loop control. This study represents the latter, testing whether meal announcement can truly be eliminated.
Previous pramlintide studies in closed-loop contexts have shown promise but also challenges. The BionicPancreas team has explored insulin-glucagon dual-hormone systems (using glucagon to prevent hypoglycemia), while this approach uses insulin-pramlintide (using pramlintide to blunt hyperglycemia). Both represent the dual-hormone strategy but with different mechanisms.
The FDA has signaled openness to fully closed-loop systems, and multiple companies are pursuing this goal. Whether pramlintide, faster insulin analogs (like Lyumjev or FiAsp), glucose-responsive insulins, or other innovations will ultimately enable fully closed-loop control remains to be determined.
Limitations to Consider
This was a small pilot study (n=12) with short duration, designed to assess feasibility rather than definitive efficacy. The outpatient setting is a strength (real-world conditions), but the study period and number of meals analyzed are limited. The trend toward more hypoglycemia requires careful evaluation in larger trials. Long-term pramlintide tolerability (GI side effects often improve with time but can be limiting) was not assessed.
Bottom Line
A fully automated insulin-pramlintide closed-loop system without carbohydrate counting achieved time-in-range comparable to standard hybrid closed-loop with meal announcements in this pilot study. While not yet ready for clinical use, this proof-of-concept demonstrates that eliminating carb counting is achievable and brings the field closer to a true “artificial pancreas.” The increased hypoglycemia signal must be addressed in future development.
Source: Madison Odabassian, et al. “A Pilot Outpatient Assessment of a Fully Closed-Loop Insulin and Pramlintide System.” Read article here.
