Continuous Glucose Monitoring in the Free-Moving Rat V. Thome-Duret, B. Aussedat, G. Reach, M.N. Gangnerau, F. Lemonnier, J.C. Klein, Y. Zhang, Y. Hu, and G.S. Wilson The aim of this work was to set up an experimental model of glycemic fluctuations for assessing in the conscious freely moving rat, the performance of a continuous glucose-monitoring system, using a pocket-calculator-size electronic control unit and a miniaturized subcutaneous glucose sensor. The well-known triphasic glycemic pattern following streptozotocin injection (initial peak and secondary hypoglycemia preceding the establishment of permanent hyperglycemia) was used as a way to obtain spontaneous changes in blood glucose level over a wide concentration range. This report demonstrates that streptozotocin injection produced highly reproducible changes in the current generated by the sensor: an initial peak and a secondary nadir, during which blood sampling provided the evidence of hyperglycemia associated with immunoreactive hypoinsulinemia, and of hypoglycemia associated with hyperinsulinemia, respectively. This reproducible experimental model should be valuable for the assessment of a continuous glucose-monitoring system. Copyright© 1998by W.B. Saunders Company T O PROVIDE continuous monitoring of blood glucose concentration for the management of diabetes therapy, 1,z our laboratories have developed a miniaturized amperometric glucose sensor in the form of a wire? which specifically measures glucose essentially independently of oxygen pres- sure, 4 and in which interferences by ascorbate and acetamino- phen have been eliminated. 5-7 This sensor has been investigated by implantation in the subcutaneous tissue of rats? dogs, 9,1° and human volunteersJ I A wearable electronic control unit (ECU) has been developed to transduce, record, filter, and process the signal generated by the sensor.12 The aim of this work was to set up a reproducible animal model of glycemic fluctuations to assess the performance of the continuous glucose-monitoring system under conditions simulating the changes in blood glucose concentration that can be observed in a diabetic patient. The diabetogenic drug streptozotocin is known to produce in the rat an initial increase in blood glucose, followed by transient hypoglycemia before the establishment of permanent hypergly- cemia 24 hours after injection. These triphasic changes in plasma glucose concentrations are due, respectively, to an initial inhibition of insulin secretion, a massive release of insulin, and finally a definitive insulin deficiency due to [3-cell destruction. 13 We first set up a system for evaluating the glucose-monitoring system in the conscious freely moving rat over a period of several days. Second, we used the variations in blood glucose following streptozotocin injection as an experimental model to evaluate performance of our continuous glucose-monitoring system. MATERIALS AND METHODS Glucose Sensors Preparation of the miniaturized glucose sensor has been described elsewhere) The sensor consists of a platinum anode covered with Teflon, except for a 1.5-ram long cavity near its extremity, where glucose oxidase was layered. A silver/silver chloride cathode was wrapped around the Teflon coating. The sensor was then coated with polyurethane. The external diameter of the sensor was approximately 0.35 mm. Wearable ECU The glucose sensor was connected to a wearable, battery-driven (6-V) ECU (150 × 80 × 30 ram, 250 g; I~cole des Mines, Fontaineb- leau, France), which controls the sensor-applied potential, and acquires, digitizes, and stores the sensor glucose-dependent current. The memory allows the storage of data for up to 6 days (one sample per minute). The ECU continuously displays the values of the current or of its transforma- tion into a glucose concentration (discussed later) on a liquid crystal display (LCD). The digitized signal was processed in real time by an original gradient algorithm. A gradient filter is more efficient than classical filters of similar complexity, because it is simultaneously based on the properties of the signal to be measured (maximum admissible gradient) and on the noise characteristics (duration of perturbations). This smoothing procedure removes all spikes in which the gradient of the signal is greater than the maximum admissible value and which has a width narrower than the filter size. It decreases the amplitude of the spikes that have a width wider than the filter size. This gradient algorithm is described in the Appendix. Surgery and Implantation A 7-cm sterile silicone catheter (Medical Grade Silicone Tubing; 0.64 mm ID, 1.19 mm OD; Sigma Medical, Nanterre, France) was connected to a 21-gauge bent needle. A 1.5-cm connector (0.94 mm ID, 1.56 mm OD; Tygon, Norton, Vermon, OH) was placed between the tip of the needle and a polyethylene catheter (0.58 mm ID, 0.96 mm OD; Diagnostics Merck-Biotrol, France), which was indweller through a metallic tether and was finally connected to a swivel (Harvard reference 568154; Biosciences, Ealing, France). Another piece of silicone cath- eter was connected to the swivel through a 21-gauge needle to a syringe containing 1% heparin (Panpharma, Luitr6, France) in physiological saline solution. The anode and the cathode of a glucose sensor were attached through a connector to two wires running through the tether and attached to the bottom part of the swivel. Finally, the wires (anode and cathode) linked to the top part of the swivel were connected to the electronic control unit. Seven male Wistar rats (250 to 300 g) were anesthetized with halothane. The silicone catheter was indwelled into a jugular vein, its tip being placed in the fight atrium, and tunnelled through the subcutaneous tissue to the interscapular area. Then, the catheter was connected to the From INSERM U341, D@artement de Diab~tologie, H6tel-Dieu, Paris, France; Ecole des Mines, Fontainebleau, France; and the Department of Chemistry, University of Kansas, Lawrence, KS. Submitted July 31, 1997; accepted January 15, 1998. Supported by lnstitut National de la Santd et de la Recherche Mddicale (grant CNAMTS), National Institutes of Health grant no. DK 30718, Aide aux Jeunes Diabdtiques (grant to V.ZD.). Address reprint requests to G. Reach, MD, INSERM U341, Service de Diabgtologie, HOtel-Dieu, 1, place du Parvis Notre-Dame, 75004 Paris, France. Copyright © 1998 by W.B. Saunders Company 0026-0495/98/4707-0006503.00/0 Metabolism, Vol 47, No 7 (July), 1998:pp 799-803 799