Pump-induced insulin aggregation. A problem with the Biostator.

A fall in plasma IRI despite constant C-peptide levels during prolonged insulin euglycemic clamp studies using the Biostator (Ames Division, Miles Laboratories, Elkhart, Indiana) prompted a meticulous evaluation of the Biostator's insulin delivery system. At slow infusion rates, a striking loss of immunoreactive and biologically active insulin was observed after 6 h of the Biostator run. Studies with labeled insulin indicated that the loss of insulin was not due to adsorption of insulin to the tubing since recovery of labeled insulin was close to 100%. A variety of techniques (gel filtration, polyacrylamide gel electrophoresis, centrifugation, and Coomassie Brilliant Blue protein assay) indicated that the loss of insulin activity was due to insulin coming out of solution. The insoluble nature of the immunologically and biologically inactive insulin was confirmed by centrifugation, i.e., 88% 125I-insulin precipitated into the pellet. The dependency of this loss of insulin activity on flow rate was clearly demonstrable with activity (IRI) less than 20% of expected at flow rates of 2.1 ml/h, and 30% at 4.2 ml/h. Full recovery was observed only with flow rates of 16.8 ml/h or greater. At each flow rate, IRI rose only after delivery of the effluent between the pump and exit port, demonstrating that insulin alteration occurs within the pump assembly presumably from heat-induced aggregation. Investigators employing the Biostator should carefully examine their systems for this time- and flow rate-dependent alteration of insulin. The loss of IRI at low flow rates (low-dose insulin clamp or insulin delivery during basal periods) will profoundly influence data generated from the Biostator.