Daniel J. Fletcher, PhD, DVM, DAVCECC
Glucose Homeostasis in Health and Illness
In health, plasma glucose concentrations are maintained within a narrow range by homeostatic mechanisms involving several anabolic hormones, including insulin and insulin-like growth factors as well as catabolic, counter-regulatory hormones such as glucagon, cortisol, catecholamines, and growth hormone. The balance between intake (gastrointestinal carbohydrate absorption), tissue utilization (glycolysis, glycogen synthesis, the citric acid cycle, etc.) and endogenous production (glycogenolysis and gluconeogenesis) determines the plasma glucose concentration at any given time.
Patients with derangements in these homeostatic mechanisms may experience significant excursions in blood glucose concentrations. Diseases such as diabetes mellitus and insulinoma are associated with severe glucose excursions due to direct effects on insulin secretion. However, significant derangements in glucose homeostasis due to peripheral insulin resistance have also been documented in other critically ill patients suffering from a multitude of primary diseases, even those without diabetes. High serum levels of insulin-like growth factor-binding protein 1, related to an impaired response of hepatocytes to insulin, have been reported in non-diabetic critically ill patients, and have been associated with an increased risk of death. Several recent controlled clinical trials have demonstrated significant improvement in outcome in heterogeneous populations of critically ill people treated with intensive insulin therapy to maintain normal blood glucose concentrations.1-3
Intermittent Glucose Monitoring
Patients at risk of derangements in blood glucose concentrations have traditionally been monitored using intermittent blood sampling techniques. This can be accomplished by intermittent venipuncture, or via placement of sampling catheters, such as percutaneously inserted peripheral catheters (PICC lines), jugular catheters, or arterial catheters. Only small blood samples are required for each blood glucose determinations using handheld glucometers or blood gas analyzers (typically on the order of 0.1-0.2 ml per sample). Benefits of intermittent blood glucose sampling include high accuracy and relatively low expense. Negative aspects of intermittent blood glucose sampling include increased patient stress, which can itself influence blood glucose concentrations, and significant blood loss with repeated sampling, especially in small or young patients.
Continuous Glucose Monitoring Technologies
Several technologies for continuous monitoring of blood glucose concentrations have been described in the literature, but only interstitial glucose monitors have become available clinically. These devices use sensors placed in the subcutaneous space to measure interstitial glucose concentrations through the use of a glucose oxidase electrode. The electrode generates a current that is proportional to the concentration of glucose in the area of the sensor. The device then uses a mathematical algorithm to estimate the blood glucose concentration from the interstitial glucose concentration. Figure 1 shows a graphical representation of the relationships between blood glucose concentrations ([Glc1]), interstitial glucose concentrations ([Glc2]), and intracellular glucose concentrations ([Glc0]). These concentrations are related by the following equation:
Where V1 and V2 are the volumes of the intravascular and interstitial spaces, respectively.
Under steady state conditions, the diffusion coefficients and volumes are constant, which means that the interstitial and blood glucose concentrations are proportional to one another, and blood glucose concentration can be estimated from interstitial glucose concentration using a simple scaling factor. The continuous interstitial glucose monitors currently on the market all require intermittent calibration using blood glucose measurements, usually via handheld glucometers, to determine the scaling factor for converting interstitial glucose to blood glucose concentrations. Because changes in the scaling factors can occur over time, most of the devices require at least twice daily calibration.
Relationship between blood glucose ([Glc1]), interstitial glucose ([Glc2]) and intracellular glucose ([Glc0]) concentrations, along with their diffusion coefficients.
In patients with rapid changes in blood glucose, such as diabetic patients, patients receiving insulin or corticosteroid therapy, and critically ill patients, estimates of these changes in blood glucose concentration using interstitial glucose measurements may be delayed and dampened due to the time constant associated with the diffusion of glucose from the intravascular to the interstitial space. The time constant can be estimated from the diffusion coefficients, as shown in Figure 2. In experimental studies in dogs, rapid changes in intravascular glucose concentrations resulted in proportional changes in interstitial glucose concentrations approximately 10 minutes after intravenous bolus injection of dextrose solutions. Therefore, data derived from these devices should be interpreted cautiously when rapid changes in blood glucose are occurring, but trends in these changes may be clinically useful.
With rapid changes in blood glucose concentrations ([Glc1]), a delay in reflection of the change in interstitial glucose concentration ([Glc2]) will be noted. The time constant of the delay can be estimated from the diffusion coefficients (see Figure 1).
Accuracy of Continuous Glucose Monitoring Systems
Several veterinary clinical studies have evaluated the use of interstitial glucose monitoring systems in healthy and diabetic dogs, cats, and horses. The Medtronic MinimedTM Continuous Glucose Monitoring System Gold (CGMS GoldTM) was evaluated in a series of studies, and was shown to correlate reasonably well with handheld glucometer systems in each of these studies.4-7 A variety of statistical tests were used in these studies, including Bland-Altman analyses and correlation analyses. In addition, the effect the information gained from the CGMS on treatment recommendations for dogs with diabetes mellitus was evaluated in one study. This information was shown to significantly change insulin dosing recommendations in several cases.4
The author has recently completed a study evaluating the accuracy of this system in 12 dogs and 11 cats with diabetic ketoacidosis (DKA). Patients admitted to the hospital on an emergency basis and diagnosed with DKA were prospectively enrolled in the study. Blood glucose concentrations were measured every 2-4 hours during the course of hospitalization (2-5 days per patient) and were compared to estimates obtained from the CGMS system. A total of 788 pairs of glucoses were obtained from patients and were analyzed using the Consensus Error Grid analysis, a method developed by a group of endocrinologists specifically for evaluation of devices to measure blood glucose concentrations.8 The results of this analysis are shown in Figure 3. Each pair of glucose concentrations is plotted on the error grid, with the "gold standard" glucose concentration on the x-axis (in this case, the glucose obtained from the glucometer) and the CGMS glucose on the y-axis. If the device were to work perfectly, all data points would fall on the diagonal. The grid is divided into 5 regions, based upon the clinical implications of the error in blood glucose concentration measured by the CGMS. Data points falling in grid "A" have no clinical implication, those falling in grid "B" would result in no treatment or benign treatment, those in grid "C" may result in overcorrection of acceptable blood glucose levels, those in grid "D" would result in a failure to detect and treat significant glucose excursions, and those in grid "E" would result in erroneous or contradictory treatment. In this study, 78.6% of the data points fell in grid "A", 20.2% fell in grid "B", 1.2% fell in grid C, and none fell in grids "D" or "E", indicating acceptable clinical accuracy of the device in this group of patients with DKA. In addition, analyses to evaluate the effects of hydration, perfusion, body condition score, and degree of ketosis failed to show any statistically significant effect of these parameters on accuracy of the system. Finally, the effect of calibration frequency was investigated by comparing the accuracy of the device when calibrated every 8 hours with that of the device calibrated every 12 hours. No significant difference in accuracy was noted, suggesting that twice daily calibration of the device was sufficient, even in ill hospitalized patients with DKA. The results of this study provide evidence that the CGMS GoldTM system has clinical utility for in-hospital monitoring of patients with DKA.
Consensus error grid data for 788 pairs of glucose concentrations obtained from 12 dogs and 11 cats with DKA.
Several interstitial continuous glucose monitoring devices are currently on the market. The first FDA approved device on the market was the Medtronic MinimedTM CGMS GoldTM (http://www.minimed.com/products/guardian). This device, consisting of a single use subcutaneous glucose oxidase sensor connected via a cable to a pager-sized data storage device, was able to store blood glucose estimates every 5 minutes, which could later be downloaded to a computer for analysis. It did not provide real time estimates of blood glucose concentrations, and its utility was therefore limited to glucose curve information. The next generation device has recently been released to the market, the Medtronic MinimedTM Guardian RTTM system. This device consists of the same single-use subcutaneous sensor, but is now connected to a small transmitter (approximately the size of a quarter), which wirelessly transmits data to a pager sized unit that stores glucose data every 5 minutes, displays a current estimate of glucose concentration, and can display trend graphs of recent glucose data. Sensors can be left in place for up to 3 days. The author has used this next generation device in a number of dogs, cats, and ferrets, and it has performed well.
The DexcomTM Seven SystemTM (http://dexcom.com/html/dexcom_products.html) is another subcutaneous continuous glucose monitoring system that is very similar to the Guardian RTTM. It consists of a subcutaneous sensor, wireless transmitter, and receiving unit that stores data and displays glucose trends. In addition, Dexcom has demonstrated an implantable subcutaneous device that can be left in place long term (for up to 1 year). Like the Guardian RT, this device must be calibrated daily with blood glucose measurements, which are directly sent to the system via a connected glucometer.
Several other interstitial glucose monitoring systems are currently being evaluated for FDA approval, including the Abbott Freestyle Navigator® system (http://www.abbottdiabetescare.com). It is likely that several options will be available for clinical use in the near future.
Interstitial continuous glucose monitoring systems have the potential to provide useful clinical information for patients with diabetes mellitus and DKA, as well as for monitoring of critically ill hospitalized veterinary patients. They appear to offer good clinical accuracy in a minimally invasive way, and may prove to be useful for implementation of tight glycemic control protocols in patients with diabetes as well as other critically ill patients with insulin resistance.
1. van den Berghe G, et al. N Engl J Med 2001; 345(19):1359.
2. van den Berghe G, et al. Crit Care Med 2003; 31(2):359.
3. Vanhorebeek I, et al. Endocr Pract 2006; 12 Suppl 3:14.
4. Davison LJ, et al. J Small Anim Pract 2003; 44(10):435.
5. DeClue AE, et al. J Am Anim Hosp Assoc 2004; 40(3):171.
6. Wiedmeyer CE, et al. J Am Vet Med Assoc 2003; 223(7):987.
7. Wiedmeyer CE, et al. Diabetes Technol Ther 2005; 7(6):885.
8. Parkes JL, et al. Diabetes Care 2000; 23(8):1143.