Accuracy and precision of low-dose insulin administration.

Abstract

To determine the lowest dose of concentrated (U100) insulin that can be reproducibly delivered. A telephone survey was used to determine current practices in major pediatric hospitals regarding the administration of low doses of concentrated insulin. A sensitive gravitometric technique was used to determine the error in measurement of low doses of U100 insulin by pediatric nurses and parents of diabetic children. A telephone survey revealed that doses as low as 0.5 or 1.0 U (5 to 10 microL) are routinely administered in pediatric hospitals. In our study of pediatric nurses, attempts to deliver 0.5, 1.0, and 2.0 U resulted in delivered doses of 0.975 +/- 0.315, 1.638 +/- 0.376, and 2.153 +/- 0.435 U (mean +/- standard deviation of the mean), respectively. The use of 0.3-mL syringes compared to 0.5-mL syringes did not improve accuracy or precision. Taken as a group, parents of children with diabetes were more accurate (mean delivered dose of 1.063 +/- 0.276 for the 1-U dose), but the individual means ranged from 0.641 to 1.300 and coefficients of variation ranged from 5% to 33%. Only three of the seven parents could deliver 1.0 U with acceptable precision and accuracy. When currently available syringes are used, insulin injections of less than 20 microL (2 U of U100) have an unacceptably large error. It is recommended that, in the inpatient setting, diluted insulin be used if the prescribed dose is less than 2.0 U.