Variability of the efficacy of a small subcutaneous insulin bolus in CSII-treated type 1 diabetic subjects.

In order to observe the inter- and intrapatient variability of the effect of a small s.c. insulin bolus, of the same size of pre-prandial boluses commonly used in CSII-treated diabetics, 0.1 U of regular insulin (Actrapid HM) was injected into 5 euglycaemic Type I CSII-treated diabetic subjects, while plasma glucose was maintained at 100 mg/dl using De Fronzo's clamp technique. Each subject was studied 3 times at 2-4 wk intervals. The mean glucose consumption for the 15 tests was 64.7 ± 21 g with a CV of 26 ± 9.1% (within-patient variability). A significant correlation was observed between glucose consumption and basal insulin requirement but no correlation between glucose consumption and age, relative weight, duration of diabetes and glycosylated haemoglobin could be detected. The ANOVA showed a significant inter-individual variability (p = 0.047). In order to study the repercussion of this variability on a glycaemic curve, the same subjects were then infused three times with an i.v. 5-hr glucose infusion (113 g - the highest glucose consumption during clamp studies) at a variable rate adjusted to mimic the absorptive pattern of a carbohydrate meal. A similar intra-patient variability (CV = 30 ± 13.5) was observed and the ANOVA showed a significant inter-subject variability (p = 0.005). There was a significant positive correlation between basal insulin requirement and incremental glucose area and a negative correlation between log incremental area and the glucose infused during clamp studies. These findings document the existence of a great intra-subject variability of the insulin effect, but the existence of an inter-subject variability justifies the efforts to individualize the pre-prandial insulin dose.