Glucose intolerance is a common occurrence in ELBW infants. Hyperglycemia in ELBW infants is treated with human insulin, diluted to a concentration of 0.2 U/ml to deliver infusion rates of 0.01-0.1 U/kg/hr in a minimal fluid volume. In 6 infants treated with low-dose insulin infusions, normalization of blood glucose (<130 mg/dl) did not occur until a median of 17 hours (range 14-25). We hypothesized that this delay was due to binding of insulin to the microbore tubing, rather than to insulin resistance. To test this hypothesis, we used a Microparticle Enzyme Immunoassay (Abbott Laboratories) to measure insulin concentration from stock solution (0.2 U/ml), and from the effluent collected at various time intervals during infusion @ 0.05 ml/hr through microbore tubing (tubing length 152 cm; residual volume 0.28 ml; medexinc). We found that at 1 hour, only 2.3% of the stock concentration was present in the effluent. The concentration delivered (expressed as% of stock concentration) increased as a function of time; 21.8 and 27.2% was delivered at 4 and 8 hours, respectively. It took an average of 18 hours to saturate insulin binding sites and achieve a steady state delivery rate equal to 100% of the stock concentration. This correlates well with the clinical lag time for blood glucose normalization observed in ELBW infants. Prior reports of insulin administration to ELBW infants recommend addition of albumin to IV solutions to reduce adsorption to plastic infusion materials, an approach which incurs added expense and exposure to a human blood product. As an alternative, we filled the tubing with 5 U/ml insulin × 5 minutes, followed by a 3 ml flush with saline, and then with stock insulin solution before resuming the insulin infusion of 0.2 U/ml @ 0.05 ml/hr. We recovered 20% of the stock concentration in the effluent collected at 1 hour and achieved steady state insulin delivery by 4 hours. We conclude that the delay in clinical response to insulin infusions in ELBW infants is due to binding of insulin to infusion tubing and that this problem can be minimized by flushing the tubing with a higher concentration (5 U/ml) of insulin prior to initiation of a low-dose insulin infusion rate.
(Spon by: Steven M. Block)