Here in the ICU at DoveLewis, we have the opportunity to treat many patients with diabetic ketoacidosis either due to newly diagnosed diabetes or disregulation of a previously diagnosed diabetic patient. Previously diagnosed diabetics can become ketotic because of changing insulin needs or the development of insulin antagonistic conditions. Patients with ketoacidosis who are sick require hospitalization, Regular insulin therapy, nutritional support, acid base and electrolyte monitoring and intensive glucose monitoring to reverse their ketoacidotic state. This often means spending 3-5 days in the ICU. During that time, glucose monitoring every 1-2 hours is required in the initial 1-3 days to guide Regular insulin therapy. This frequent glucose monitoring can be achieved by several methods - each with their advantages and disadvantages.
The first method of serial glucose monitoring is to perform serial peripheral venous blood sampling (saphenous, cephalic, or auricular vein sampling). This method requires no specialized equipment and is easy to perform, but a significant disadvantage of this method is caused by the discomfort of serial and frequent blood sampling for the patient. After 1-2 days of hourly to every other hour sampling, these small veins can become bruised or thrombosed limiting their usefulness for vascular access or further blood sampling. The second method to achieve frequent serial glucose monitoring that eliminates some disadvantages posed by peripheral collections is the placement of a central or peripheral blood sampling catheter. The placement of a sampling catheter requires some specialized equipment and catheters. It also requires the expertise to place them and maintain their patency. It may require heavy sedation or a short anesthesia depending on the type of catheter being placed and depending on the patient’s compliance. Percutaneously placed central jugular catheters usually require heavy sedation to a short anesthesia but are reliable with respect to patency over several days. Peripheral sampling catheters may not require sedation but are less reliably patent for extended days of blood collection. Use of any catheter for sampling requires the withdrawal of a volume of blood (usually 3 cc) prior to collecting the small sample for blood glucose analysis. If the initial withdrawal volume of blood is mixed with heparinized saline, it can be given back to the patient after the diagnostic sample is drawn. Even with this method of return of the initial blood withdrawal volume, many small cats and some small dogs undergoing serial blood collections will become anemic after several days in hospital. Thrombophlebitis and catheter associated infections are rare complications of sampling catheters. Over the last decade, continuous interstitial glucose monitoring systems initially developed for use in human diabetic patients have been evaluated for use in dogs, cats and horses and gives us a final option for frequent serial monitoring of glucose (Weidmeyer 2003).
As with any translational technology from human medicine, the feasibility or practical aspects of use in our veterinary patients, as well as accuracy of the information to guide clinical decision-making, must be critically assessed before it can become the standard of care. With respect to continuous glucose monitoring systems (CGMS) work by Weidemyer, Moretti, Reineke and associates have sufficiently answered these questions about feasibility and accuracy even in diabetic ketotics. The intent of this article is to introduce you to this technology including practical aspects, indications and discuss the challenges or myths that developed as we initially started using this technology here at DoveLewis.
The continuous glucose monitoring system measures interstitial glucose concentration with a subcutaneously placed glucose sensor. This sensor converts glucose concentration in the interstitial space to an electrical signal that is translated into a glucose measurement based on blood glucose calibration points that the user inputs into the machine. The CGMS consists of:
A glucose sensor that is inserted into the patient
A glucose transmitter attached to the glucose sensor and patient
The monitor is the control point for inputting calibration points and viewing glucose history
A com-station (not pictured) that acts as a download port to download information stored in the monitor. This information can be viewed in a graphical or numerical form.
Clinical use and set up for the patient:
- A small region on the lateral thoracic or interscapular region of the patient is clipped (generally the area of clip is about 2x2cm). The Glucose sensor is inserted/injected into the SQ space with the insertion applicator provided with the CGMS sensors. To ensure the glucose sensor is not dislodged, it can be secured with a couple of small drops of tissue glue.
- The glucose transmitter is attached to the glucose sensor. The glucose transmitter is secured to the patient by either suturing with white tape or by securing it within a circumferential chest bandage in such a way that it does not move or dislodge the glucose sensor. The apparatus is left for two hours to initialize. During this initialization time, the CGMS is not taking glucose readings.
- After two hours the monitor will require the input of a calibration blood glucose reading. The user does this by taking a measurement of the patient’s peripheral blood glucose value on any glucometer and inputs that reading into the CGMS monitor. The CGMS should now start reading glucoses every five minutes which can be viewed and scrolled through in real time on the CGMS monitor.
- The next time the CGMS will ask you to calibrate the machine will be in ~12 hours at which point you repeat step three for inputting a calibration blood glucose reading. The CGMS will require calibration glucoses every 12 hours but otherwise will read glucoses every five minutes for up to 72 hours.
- If BG monitoring is required after the initial 72 hours, the glucose sensor will need to replaced and initialization and calibration steps will need to be repeated.
We have had the continuous glucose monitoring system here at DoveLewis since 2008, and with any new technology there can be challenges in implementing an unknown therapy or device. Table 1 addresses some myths that developed with our staff as we started using this machine, and with increasing experience, and in truth a level of trial and error and determination, these have largely been overcome.
Myths and experience/responses with use of CGMS
Myth: Frequent sensor calibration negates the benefit of not having to draw blood from the patient.
Response/Evidence: Previous technology had required four calibration points in the first 12 hours; however, recent papers have indicated that the sensor remains highly accurate with two calibration points requiring only two blood collections per 24 hours to keep the CGMS accurate.
Myth: Interstitial glucose is not accurate for clinical judgments especially in cats.
Response/Evidence: For the response to this, I would cite the evidence of the following references: Weidmeyer JAVMA 2003, Weidemyer JVIM 2009, Reinike, JVECC 2010, Moretti JVIM 2010
Below is a Clark error grid analysis from Reinike, JVECC 2010 indicating that with q12hour calibration, most CGMS readings fall in Zone A and B which are defined as values differing by less than 20% or greater than 20% but not altering treatment or clinical decision making.
Myth: Unable to see real time results. Results can be viewed only in retrospect after download to software.
Response/Evidence: Glucose readings taken every five minutes for eight hours can be viewed directly on the glucose monitor in addition to 72 hour glucose curves that can be downloaded retrospectively.
Myth: Not accurate in DKA patients.
Response/Evidence: Evaluated in DKA patients and it had good clinical accuracy. Not affected by perfusion, degree of ketosis though it is more accurate in patients that are better hydrated. Reinike, JVECCS 2010
Myth: Costs more than q2hr blood glucose collection.
Response/Evidence: In the first day there is a higher cost in setup of the system, thereafter savings will likely be about $80/24 hrs in glucose monitoring if the sensor does not have to be replaced. Most DKA’s will require 2-3 days of monitoring and therefore it is generally more cost effective than q2hr glucose collection and monitoring.
Myth: Sensors have to be replaced frequently
Response/Evidence: Sensors lasted 72 hours even when patients were discharged home to owners. Weidemeyer JVIM 2008
There are some limitations to the usage of CGMS. Notably based on equipment availability, we can only have one patient on CGMS at a time either in the ICU or at home. This may appear to not be an obstacle, but we treat enough DKA patients that sometimes there are several in the ICU at once, at which point we need to select the patient who would benefit from the monitoring the most. Another limitation with respect to the CGMS is that its reading range for blood glucoses is 40-400mg/dl beyond which it will give a LO or HI reading respectively. Most glucometers will not give a numerical value beyond these ranges either so the CGMS readings mirror the clinical technology that we currently use in most patients to monitor glucose. However, in some hyperosmolar ketotics that have very high glucose readings it may not be as useful to titrate insulin therapy. The telemetry of the monitor operates within six feet of the patient, so the monitor will need to be carried outside with dogs that take walks, or must be attached to them so there are not gaps in the data obtained. Because the most fragile and finicky portion of the apparatus is the maintenance of the glucose sensor which is directly attached to the patient, patient selection for use of the CGMS is key to success. A very fractious cat that needs to be wrestled and wrapped in a towel for each treatment or handling has a high likelihood of dislodging the sensor requiring replacement and recalibration - effectively decreasing the utility of the CGMS for frequent, minimally-invasive and accurate measurements of glucose.
Because of patient case load, our use of CGMS is primarily in the ICU and acute hospitalization setting, but it has been used in getting an accurate and longer duration curve to assist in regulating chronic diabetic patients. This is achieved by equipping the patient with the apparatus, returning the patient to their home environment to get a curve while they are getting their regular feeding and insulin routines. After 2-3 days at home, the patient can return to the clinic, the monitor can be removed and the data stored and collected on the monitor can be downloaded into the com-station.
For each of your patients that spend time with us for resolution of their DKA state, we will consider their individual needs in determining which method of frequent glucose monitoring is best for the patient. We still use serial peripheral venous collection, central or peripheral sampling catheter placement and CGMS in the ICU depending on many patient characteristics and the duration of a time they are expected to stay with us. We want to provide the most accurate information, in the least invasive and most cost effective manner.