No abstract

Summary.For the normalisation of blood glucose levels in diabetic patients by feedback controlled insulin delivery, a self-manageable and reliable method for continuous glucose estimation is still not available. By combining a commercially available needle type dialysis probe (molecular cutoff 20,000 Da) with a sensitive glucose sensor, we obtained a device for continuous glucose measurement in dialysate. This device was tested in healthy volunteers during a 75-g oral glucose tolerance test and in Type 2 (non-insulin-dependent) diabetic patients. Venous glucose and subcutaneous sensor signal were followed for 300 min (ten healthy subjects), 21 h (three healthy subjects) or 9 h (seven Type 2 diabetic patients). The recovery of the microdialysis was interindividually different, but after calibration, glucose levels in the dialysate and subcutaneous glucose sensor signal correlated well (r = 0.84-0.95). Under the assumption of a physiologic and technical delay between intravenous and subcutaneous glucose, correlation coefficient between intravenous glucose and subcutaneous sensor signal ranged from 0.60 to 0.93. We conclude that changes in blood glucose could be monitored in the subcutaneous tissue by the microdialysis technique in a continuous on line manner.

The incidence and magnitude of hypoglycemia (i.e., blood glucose values less than 50 mg/dl) were assessed by continuous blood glucose monitoring over 24 h in 10 insulin-dependent diabetic (IDD) patients treated with continuous subcutaneous insulin infusion (CSII) and 9 IDD patients under intensified conventional treatment (ICT). A newly developed, battery-powered blood glucose monitor was employed. Patients were thus enabled to move freely in the hospital premises. Despite similar quality of previous blood glucose control (HbA1: 8.0 +/- 0.05% CSII versus 8.0 +/- 0.3% ICT, mean +/- SEM), the obtained profiles showed better regulation under CSII treatment (mean blood glucose [MBG], 99.6 +/- 10.0 versus 133.1 +/- 7.4 mg/dl; M-value, 12.3 +/- 3.5 versus 26.2 +/- 4.1; mean amplitude of glycemic excursion [MAGE], 71.9 +/- 8.7 versus 132.9 +/- 14.2 mg/dl; CSII versus ICT, mean +/- SEM). The incidence of blood glucose values less than 50 mg/dl was 9/10 patients (CSII) and 5/9 patients (ICT). In both groups, hypoglycemia was most frequent at noon and was related to elevated pre- and postprandial free insulin levels. Patients became aware of hypoglycemia only in 6/23 episodes (CSII) and 6/8 episodes (ICT). Our data indicate that CSII as well as ICT may result in postprandial hyperinsulinemia leading to frequent hypoglycemic episodes of variable length, reassessing the traditional experience of close correlation between aggressive insulin therapy and enhanced hypoglycemic risk.

An amperometric enzyme electrode and a wick technique were used for measurement of glucose in sc. tissue of sheep. When wicks were left implanted long enough to equilibrate with interstitial fluid, sc. glucose could have been reproducibly determined with the necessary accuracy. It was demonstrated that sc. tissue glucose concentrations in sheep are about 30% higher than in whole blood and are on the level of plasma glucose. This allows interpretation of sc. glucose sensor currents since results of in vitro-calibrations cannot be transferred to in vivo conditions. When an enzymatic sensor was implanted in the sc. compartment, the sensor signals were closely related to changes of blood glucose. These in vivo experiments indicate that short term glucose-monitoring with an subcutaneously implantable glucose sensor is feasible and so may provide a possible access to glycemic control. Further experiments will have to show, if glucose-controlled insulin infusions based on the output of a sc. glucose sensor will be able to maintain stable normoglycemia.

The microdialysis technique was used for following the glucose content of the extracellular subcutaneous (SC) fluid under varying blood glucose levels in rats. The glucose content in the microdialysis perfusion fluid was continuously analyzed by means of the measuring flow chamber of an ex vivo glucose monitor. In six ChBB rats blood glucose levels were varied between 40 mg/dl and 575 mg/dl by intravenous (IV) infusion of glucose and by SC injections of insulin, respectively. After a running-in period of about half an hour, the glucose content in the perfusion fluid was closely related to the blood glucose concentration (r > 0.92) up to a time period of 6 hrs. The "relative recovery" rate of glucose by the microdialysis probe in the SC tissue varied within the 6 experimental sessions. The relative recovery rate could be shown to be not dependent on the absolute blood glucose levels in the individual rat within the glucose concentration range tested.