The pancreas secretes both insulin and amylin, so should an artificial pancreas system provide insulin and pramlintide?
In individuals without diabetes, amylin is secreted alongside insulin, which people with type 1 diabetes lack. Pramlintide, an analog of amylin, is meant to be injected at mealtimes to help delay gastric emptying, decrease glucagon secretion, and increase satiety. An artificial pancreas system has been developed to deliver insulin based on glucose sensor readings automatically. This has been shown to have better results than sensor-augmented insulin pump therapy with decreased hyper- and hypo-glycemia. Previous studies looked at adding injectable pramlintide to insulin used in an artificial pancreas and reducing the change in blood glucose after a meal.
This open-label, randomized, crossover study looked at the use of pramlintide when delivered alongside insulin, not given as a separate injection as previously studied. Pramlintide was delivered in a fixed ratio alongside insulin to mimic the body’s natural co-secretion. This study evaluated artificial pancreas systems administering insulin alone or with pramlintide in adults with type 1 diabetes. Insulin given with pramlintide was further divided into two groups: rapid-acting or regular insulin. The authors tested regular insulin because they believed it would better match pramlintide absorption. This study sought to compare the rapid-acting insulin with regular insulin when combined with pramlintide in an artificial pancreas system with rapid-acting insulin by itself. Each intervention in this study lasted for 24 hours and was separated by at least 14 days. Participants had to be >18 and have used an insulin pump for ≥6 months. Those with gastroparesis were excluded. Before each intervention, insulin therapy was optimized for 10-14 days, using a treatment like the intervention, e.g., those that would receive rapid-acting insulin and pramlintide wore two pumps, one delivering insulin and the other delivering pramlintide. Participants wore a glucose sensor for 1-2 days before the intervention. The artificial pancreas yielded a fixed ratio of 6 μg of pramlintide per unit of insulin. The rapid-acting insulin group received one injection of insulin at each meal. The rapid-acting insulin with pramlintide group received three mini-boluses, one at mealtime and two separated by 10 minutes each. The regular insulin with pramlintide group also received three mini-boluses, but these were started 20 minutes before the meals. The primary outcome was the amount of time spent in the target range (70-180 mg/dL). The secondary results were the amount of time spent in hyper- or hypoglycemia, glucose variability, and gastrointestinal adverse effects.
The artificial pancreas providing rapid-acting insulin and pramlintide increased the mean percentage of time spent euglycemic from 74% to 84% (P = 0.0014), reduced mean glucose from 142 to 126 mg/dL (P = 0.0053), and reduced time spent in hyperglycemia from 22% to 12% (P = 0.00012). Compared to this, the artificial pancreas providing regular insulin and pramlintide showed no significant advantages. Most of the benefits seen in these groups were caused by glycemic control during the day (0800-2300 h), raising the percentage of time spent in the range from 63% to 78% (P = 0.0004) while reducing mean glucose from 157 to 142 mg/dL (P = 0.0011). During the nighttime (2300-0800 h), all three groups had a similar time in range (94-95%). The study found that glucose levels after meals peaked after 60 minutes in the group that received rapid-acting insulin alone, regardless of the level before meals. In the group receiving insulin and pramlintide, glucose levels after meals depended on those before the meals. If glucose was >180 mg/dL before eating, the level immediately decreased towards the goal. If premeal glucose was 90-180 mg/dL, postprandial levels remained relatively flat. If premeal glucose meal was <90 mg/dL, levels immediately rose towards euglycemia. Patients received more insulin when it was paired with pramlintide than when insulin was given alone. There were 11 hypoglycemic events with rapid-acting insulin monotherapy, 12 with rapid-acting insulin and pramlintide, and 18 with regular insulin and pramlintide. 6 of the 108 meals given to those receiving pramlintide and rapid-acting insulin were followed by gastrointestinal symptoms, with three being mild and three being moderate. There were 11 such incidences out of 104 meals when regular insulin was given with pramlintide, with two being mild, six being reasonable, and three being moderate-to-severe.
Artificial pancreas systems delivering insulin increase glycemic control. Adding pramlintide appears to improve that control further: it increased time in range, reduced mean glucose levels, and reduced glucose variability. Pramlintide caused postprandial glucose to trend towards euglycemia, no matter the premeal glucose. As this is an early study looking at the possibility of using pramlintide in addition to insulin in an artificial pancreas, there have not been other studies to support or refute these findings.
- The authors recommend using more immediate pramlintide when premeal glucose readings are high and less when levels are low to mimic the body’s natural response to food more closely.
- There were few gastrointestinal side effects seen when pramlintide was added.
- Adding pramlintide to the artificial pancreas system resulted in lower glycemic variability, decreased mean glucose levels, and increased time spent at goal.
Haidar, Ahmad, et al. “A Novel Dual-Hormone Insulin-and-Pramlintide Artificial Pancreas for Type 1 Diabetes: A Randomized Controlled Crossover Trial.“ Diabetes Care, American Diabetes Association,
George McConnell, PharmD. Candidate, LECOM School of Pharmacy