Abstract

Various routes of insulin infusion have been utilized to improve metabolic control in type I (insulin-dependent) diabetes. To determine the relative effectiveness of subcutaneous (SC) and intravenous (IV) systems of insulin-pump therapy, we studied five type I diabetic subjects aged 20–39 yr who were randomly assigned to a 3-mo period of either SC or IV insulin infusion and then crossed over to the alternate route for a similar period. After an initial 7- to 14-day period of in-hospital insulin-dose adjustment, all subjects were similarly followed as outpatients during both infusion periods, with a minimum of four daily preprandial self-measurements of blood glucose. Although the overall mean levels of blood glucose and HbA_1c (116 ± 6 vs. 114 ± 5 mg/dl and 6.1 ± 0.2 vs. 5.9 ± 0.1% for the SC and IV systems, respectively) were similar, there was a greater incidence of low (< 50 mg/dl) and high (> 180 mg/dl) preprandial blood glucose readings during SC- than during IV-pump therapy (P < .05). Furthermore, the frequency of nocturnal and preprandial hypoglycemia was greater during SC-pump therapy (P < .05 and P < .02, respectively). Eight documented technical problems related to the pump and catheter occurred with the SC system, whereas 17 episodes occurred with the IV system. The number of episodes of ketosis (4 vs. 5 for SC and IV, respectively) was similar with both systems. The number of episodes of ketosis (4 vs. 5 for SC and IV, respectively) was similar with both systems. In summary, although similar degrees of near normalization of mean blood glucose levels can be achieved with both SC and IV insulin-pump therapy, less glycemic variability, in particular less hypoglycemia, occurs with the external IV infusion system. These observations provide continued rationale for the development of implantable insulin-delivery devices for insulin replacement in type I diabetes mellitus.