Volume kinetics of glucose solutions given by intravenous infusion

Abstract: Fluid therapy is often cumbersome to plan since the distribution and the elimination of the administered solution are difficult to analyze in the clinical setting. The knowledge of the adequate rate and dose for each individual patient and under various physiological conditions would be very attractive to obtain in order to avoid fluid overload and hyperglycemia. The overarching objective of this thesis was to develop, validate and employ a novel kinetic model for intravenous glucose solutions. Methods I: In cooperation with mathematical experts, two new kinetic models were developed and designed to include the osmotic fluid shifts that accompany the metabolism of glucose. These models were fitted to data obtained when 21 healthy volunteers received approximately 1 L of Ringer’s acetate (control), glucose 2.5% or glucose 5% solution over 45 min. II: 6 healthy volunteers received four separate infusions of glucose 2.5% solution and the infusion program was chosen to disclose differences in the kinetic parameters by varying the infusion rates and infusion time while aiming at avoiding glycosuria: 10 ml kg-1 and 15 ml kg-1 over 30 min, and 15 ml kg-1 and 25 ml kg-1 over 60 min. III: The volume kinetic model was fitted to data from 12 patients receiving 18.75 ml kg-1 (1.4 L) of glucose 2.5% solution during 60 min while undergoing elective laparoscopic cholecystectomy. IV: 12 unstressed patients with type 2 diabetes received, on two separate occasions, one infusion of an isotonic glucose 2.5% solution at individual rates over 30 min and 60 min, respectively. The fluid dose was individually planned, based on the fasting glucose level before the experiments started by using the volume kinetic model, to reach a predetermined glucose level. V: 12 healthy volunteers received one infusion of glucose 2.5% solution, 19.7 ml kg-1 during 60 min, and on a separate occasion, one infusion of glucose 50% solution with insulin and potassium, 5.0 ml kg-1 during 120 min, while maintaining euglycemia. The dilution-time curves and the concentration-time curves for glucose were compared with the results from previous studies comprising infusions of glucose solutions (papers I-IV and a study on patients post-op abdominal hysterectomy). The aim was to graphically illustrate the risk of hypovolemia after an infusion of a glucose-containing solution. Results I: The two-volume kinetic model was fitted to the data. The volume of distribution (Vd) for the infused fluid was 2.5-3.7 (0.2-0.3) L and for the glucose approximately 12 L. Fluid was accumulated in the cells after the experiments with glucose infusions (0.2-0.4 L) but no expansion of the intermediary volume space could be detected. II: Increased amount of fluid dose resulted in a proportional increase in AUC for both the dilution and the glucose. Predictive performance tests demonstrated a high accuracy for the kinetic model. III: The elimination of fluid and glucose was reduced to app. 1/3 of the normal values, while the Vd for both the fluid and the glucose was within the normal range. IV: The mean deviation of the glucose level from the predetermined level at the end of the infusion was 0.4 mmol/L. The Vd for glucose was increased by 40-60% as compared to healthy subjects. The elimination of glucose was decreased by 37% but a normal value was demonstrated for the fluid. V: A strong relationship was found between the glucose level and the dilution of plasma. The risk of hypovolemia in response to hypoglycemia was prominent in healthy volunteers with rapid elimination of glucose while there was no risk of hypovolemia in connection with surgery or type 2 diabetes. Conclusions: The two-volume kinetic model described the data from studies I-IV well and typical findings were that the Vd for both the water and the glucose compounds of intravenous glucose solutions were relatively stable, with the exception of patients with type 2 diabetes (40-60% larger Vd for glucose). The clearance for both the administered fluid and glucose differed, however, depending on the physiological conditions. During surgery, the clearance for both the fluid and the glucose were reduced to 1/3 of the values seen in healthy unstressed subjects, while in patients with type 2 diabetes, the clearance for the fluid was normal, but, the clearance for glucose was reduced by 37% in this group. The kinetic model was validated and demonstrated to be linear which makes it possible to simulate the effects of fluid therapy not yet conducted.