Pancreatic secretion, hepatic extraction, and plasma clearance of insulin from steady-state insulin and C-peptide measurements in critically ill patients

Abstract

Hyperglycaemia is prevalent in critical illness and clinical practice is to reduce blood glucose by intravenous insulin infusion. We estimated pancreatic secretion, hepatic first-pass extraction ratio and plasma insulin clearance per minute from measured serum C-peptide and insulin in 9 critically ill adults in a pseudo steady-state situation, i.e. when nutrition and insulin infusion remained at a constant rate up to the test (min 20, median 240, max 510 min). To estimate pancreatic secretion, a population C-peptide kinetics model was used to convert C-peptide concentration to C-peptide secretion rate. Pancreatic secretion varied 18-fold from 5.2 to 93.5 mU·min -1 , and the lowest secretion rates were in patients older than 70 years or patients with type 2 diabetes. Pancreatic secretion correlated positively, but not significantly with blood glucose. Blood glucose was not correlated with plasma insulin. A two-dimensional regression analysis of hepatic first-pass extraction and plasma insulin clearance showed that the smallest relative error between estimated plasma insulin and measured plasma insulin was obtained for an extraction ratio of 72% and plasma clearance of 0.34 min -1 . Using these values, a negative correlation was found between post-hepatic insulin production and the rate of insulin infusion. These results indicate that 1) hepatic insulin extraction is increased in critical illness. This is also confirmed by the observation that steady-state plasma insulin concentrations in this study were relatively lower when compared to steady-state measurements in normal subjects; 2) blood glucose drives pancreatic secretion moderately; 3) there is substantial variation in pancreatic secretion between patients that cannot be explained from the blood glucose variation, but could be related to patient age and diabetic state; 4) there was also substantial variation in insulin sensitivity between patients, since similar levels of insulin sensitivity would have predicted a negative correlation between blood glucose and plasma insulin; 5) insulin infusions are used to compensate for inadequate pancreatic insulin secretion.