The impact of the bolus dose and the basal infusion during insulin pump treatment on pre- and post-prandial hyperinsulinemia was investigated in 8 insulin-dependent diabetic patients. Frequent determinations of free insulin in blood and simultaneous determination of glucose consumption with euglycaemic clamp technique were made after a bolus of 5 U with or without basal infusion and after 10 U with basal infusion. The maximum free insulin level was reached after 75, 65 and 105 min respectively (median; NS), whereas 50% of the maximum was reached already after 15, 10, and 12.5 min (NS). The fall in free insulin after the peak was slower in all protocols compared to the breakfast induced peak in healthy subjects and it was slower with than without basal infusion (p less than 0.03). Glucose consumption followed the free insulin profiles. We conclude, that administration of a subcutaneous bolus with an insulin pump gives a dose-related free insulin peak that is unphysiological in comparison with the normal meal-related insulin secretion. Preprandial hyperinsulinemia, found after a prescribed interval of 30 min between injection and meal, as well as postprandial hyperinsulinemia, increased with a greater bolus indicating that adjustments of timing of meals and snacks may be necessary when the dose is changed. Furthermore, the insulin substitution with external pumps would probably be more physiological if the basal supply was interrupted after a bolus.