Intravenous Bolus Insulin Delivery: Implications for Closed-Loop Control and Hospital Care

Abstract

I ntravenous insulin infusion to treat diabetic ketoacidosis has been one of the cornerstones of therapy since the 1920s. Few appreciate that this route of insulin delivery was not approved by the U.S. Food and Drug Association until 2005. Even more surprising are the minimal published data on the kinetics of intravenous insulin, by either intravenous infusion or intravenous bolus. Most appreciate that the actual pharmacokinetics of intravenous insulin are quite short, but it was quite surprising to learn the pharmacodynamics may be longer than appreciated.¹ However, those who previously worked with the Biostator^® (Ames Division, Miles Laboratories, Elkhart, IN) in the late 1970s and 1980s saw firsthand how slow intravenous insulin could be.²

The precise action times of intravenous insulin have taken on greater interest with the goal of developing an artificial pancreas. The real problem is the painfully slow pharmacodynamics of our current “rapid-acting” subcutaneous insulin.³ If a safe and effective system using intravenous insulin could be developed for an artificial pancreas, waiting for a faster subcutaneous insulin analog wouldn't be required.

The article by Skjaervold et al.⁴ provides us with a better understanding about the time–action profile of intravenous insulin. In some respects, it is surprising it has taken so many decades for this type of work to be completed. On the other hand, intravenous insulin is still rarely used today, although given the more common use of intravenous insulin infusion in the hospital this route of therapy is more often used now than in past decades.⁵ It should also be appreciated that the infusions and boluses done by Skjaervold and co-workers were performed in a pig model, and thus human studies should be considered before serious attention is paid to the use of intravenous insulin in a closed-loop system.

The findings by the study by Skjaervold et al.⁴ can be summarized as follows: (1) Despite the short half-life of intravenous insulin, when it is infused at rates up to 0.08 IU/kg/h blood glucose declines slowly. (2) Higher rates lead to more rapid decreases in blood glucose, but eventually hypoglycemia will ensue. (3) Hypoglycemia can be prevented with these higher rates of infusion leading to quicker reduction of blood glucose if the infusion lasts for only 2 min. (4) Intravenous bolus insulin at doses of 0.01–0.04 IU/kg results in a decrease in blood glucose on average at 4 min. (5) With intravenous bolus insulin there appears to be a clear dose dependency for degree and length of glucose level lowering. (6) Because of the short glucose response of the intravenous bolus, hypoglycemia as seen with the intravenous infusion after 30 min does not occur.

What are the major implications of these results? First, it is unlikely that subcutaneous insulin will ever have effects as quick as intravenous insulin. For a closed-loop system, intravenous bolus insulin appears to have desirable kinetics, although many future studies in humans will need to confirm these initial findings. Although we are obviously a long way from the use of outpatient intravenous bolus insulin in a true “artificial pancreas,” insulin provided by this route could possibly be used in a standardized manner for inpatients. Intraoperative blood glucose control by anesthesiologists is frequently performed with intravenous bolus insulin, yet I am not aware of insulin kinetic data, safety, or efficacy studies using insulin in this manner. In fact, my main concern with the use of intravenous bolus insulin in any part of the hospital is that these patients often have poorly controlled diabetes with resultant potassium, magnesium, and phosphate deficiency. The use of intravenous bolus insulin in this population may actually be quite dangerous. At the very least, intravenous bolus insulin should be specifically studied in different patient populations (specifically, levels of glycemic control) in the hospital, including the operating room.

Next, assuming the data from Skjaervold and colleagues can be repeated and confirmed in humans, for a closed-loop system having such a quick and effective insulin delivery will require an equally rapid glucose sensor without a lag time. This is unlikely to occur with interstitial glucose measurement, and thus an intravascular glucose sensor will be required. Like the intravenous bolus insulin delivery, besides a closed-loop implication for type 1 diabetes, there are also inpatient implications if this type of sensing device can be perfected. Indeed, it is likely an inpatient intravascular sensor used in combination with intravenous bolus insulin is a more realistic initial goal, depending on the duration the sensor lasts and the ease of delivering the insulin.

Finally, besides repeating these intravenous bolus insulin studies in humans, it will be interesting to perform these studies specifically in individuals with type 1 diabetes. Besides assessing fluxes of potassium, magnesium, and phosphate, close attention to electrocardiogram changes will be required. In particular, patients with long-standing diabetes with coronary artery disease should be studied to ensure safety. Other metabolites such as free fatty acids and β-hydroxybutyrate should also be assessed. We would expect complete suppression of lipolysis with intravenous bolus insulin, but it likely depends on the frequency with which the insulin is provided.

Although it is too soon to know if intravenous bolus insulin will someday be used for a closed-loop system, the study by Skjaervold et al.⁴ gives us initial and important information about the glycemic impact of this type of insulin delivery. Hopefully these studies can be repeated in humans with a true real-time intravascular sensor so that the elusive goal of an artificial pancreas can be achieved. Although not mentioned by the authors, these data could have important impact for the use of insulin in the hospital. As we approach the centennial birthday of the discovery of insulin, it seems appropriate to finally understand the best way to use intravenous insulin.

References

Heinemann

, Woodworth

. Pharmacokinetics and metabolism of insulin lispro. Drugs Today, 1998; 34,Suppl C:23–36.

Clemens

, Chang

, Myers

. The development of Biostator, a glucose controlled insulin infusion system (GCIIS) Horm Metab Res, 1977; Suppl 7:23–33.

Mudaliar

, Lindberg

, Joyce

, Beerdsen

, Strange

, Lin

, Henry

. Insulin aspart (B28 Asp-Insulin): a fast-acting analog of human insulin: absorption kinetics and action profile compared with regular human insulin in healthy nondiabetic subjects. Diabetes Care, 1999; 22:1501–1506.

Skjaervold

, Lyng

, Spigset

, Aadahl

. Pharmacology of intravenous insulin administration: implications for future closed-loop glycemic control by the intravenous/intravenous route. Diabetes Technol Ther, 2012; 14:23–29.

Moghissi

, Korytkowski

, DeNardo

, Einhorn

, Hellman

, Hirsch

, Inzucchi

, Ismail-Beigi

, Kirkman

, Umpierez

. American Association of Clinical Endocrinologists and American Diabetes Association consensus statement on inpatient glycemic control. Diabetes Care, 2009; 32:1119–1131.