Continuous glucose monitoring in diabetic cats

Introduction

Diabetes mellitus (DM) is a common disorder in geriatric cats 1 and appropriate management requires careful monitoring of response to insulin treatment, and in fact good glycemic control can result in diabetic remission in many cases 2,3,4. The recent introduction of technology that allows continuous monitoring of interstitial glucose has led to major improvements in the veterinarian’s ability to supervise and improve glycemic control in affected animals 5,6,7,8,9.

Type II DM is the most common type of DM in cats; this is characterized by abnormal secretion of insulin from the pancreas in combination with peripheral insulin resistance. Diagnosis is made based on the presence of clinical signs (polyuria, polydipsia, polyphagia and weight loss), and documentation of hyperglycemia and glycosuria 2,3. In cats, the diagnosis is complicated by stress hyperglycemia, so it is important not only to document persistent hyperglycemia and glucosuria, but also to rule out other diseases that may cause similar clinical signs, such as hyperthyroidism and gastrointestinal disease. Treatment of feline DM relies on insulin therapy, dietary modification, management of concurrent illness and weight management, and many type II diabetic cats will achieve remission if insulin treatment results in good glycemic control. Factors that influence the likelihood of remission include the severity of pancreatic pathology, the presence of insulin resistance caused by concurrent illness or medications, obesity, and the ability to feed a low carbohydrate diet 10,11. Progressive loss of beta-cells may ultimately result in progression to type 1 DM, therefore good glycemic control is key to a positive outcome in affected cats.

Insulin therapy

Insulin types

There are three insulin products that are appropriate for first line treatment of feline DM (Table 1); protamine zinc insulin (PZI), lente (porcine insulin zinc suspension) insulin, and glargine insulin, an insulin analogue 3. Detemir (another insulin analogue) can also be used, but is not a first-line choice because of its cost. NPH (neutral protamine Hagedorn) insulin tends to have a very short duration of action in cats, and is not recommended.

The starting dose of insulin for a new feline diabetic patient is 1-3 U/cat (0.25-0.5 Unit/kg), and the author recommends the lower end of this dose. Whichever formulation is chosen, twice daily injections are more likely to result in good glycemic control than once a day therapy. If the former is not possible, once daily injections with PZI or glargine can result in effective control of clinical signs in some cats.

Table 1. Insulin products recommended for use in cats.

Goals of insulin treatment

The primary goal of insulin therapy in affected cats is to control the clinical signs of DM while avoiding hypoglycemia. A secondary goal may include achieving diabetic remission. The plan for monitoring should take into account the owner’s lifestyle, any concurrent illness, the age of the patient, and the practicality of tight glucose monitoring. The likelihood of remission in cats is higher with tighter glycemic control; however, severe hypoglycemia can be life-threatening and can lead to permanent neurologic damage. Insulin-induced hypoglycemia also causes secretion of hormones that oppose the action of insulin, such as glucagon, growth hormone, cortisol and epinephrine, which can initiate insulin resistance and worsen diabetic control.

Ideally, blood glucose should be maintained between 80-200 mg/dL (4.4-11.1 mmol/L), but most diabetic cats will sometimes have levels that are above this range. However, most cats are clinically well regulated if the blood glucose concentration is maintained less than 300 mg/dL (16.7 mmol/L) for the majority of the day, because the tubular maximum for resorption of glucose from the feline kidney is approximately 270 mg/dL (15 mmol/L) 12. It is important to remember that it is difficult to assess the duration of insulin action if the glucose nadir is in the hypoglycemic range, because secretion of counter-regulatory hormones such as glucagon will prematurely increase the blood glucose. The ideal monitoring strategy for evaluating response to insulin treatment in diabetic cats should be individualized for the patient and owner.

Traditional monitoring of diabetic patients

Until recently, the primary monitoring tools at the veterinarian’s disposal were assessment of clinical signs and bodyweight, and measurement of serial blood glucose concentrations, urine glucose, and glycosylated proteins.

Clinical signs

The most important treatment goal for any diabetic cat is to control the clinical signs of the disease. Cats with inadequate glycemic control will typically have persistent signs and progressive weight loss, while severe hypoglycemia may cause intermittent signs such as weakness, lethargy, and seizures. Milder hypoglycemia is easily missed, because it may not result in obvious clinical signs, yet will still contribute to poor glycemic control.

Blood glucose curves

Traditional in-hospital or at-home blood glucose curves have been the gold standard for evaluating glycemic control in cats for many years, but this method has numerous limitations. The method is expensive and requires collection of multiple samples, which causes stress to the patient and owner. In addition, blood glucose curves exhibit marked day-to-day variability, even when performed by the owner at home (Figure 1) 13. Misinterpretation of the results can also lead to incorrect treatment decisions.

Glycosylated proteins

Measurement of glycosylated proteins such as fructosamine and hemoglobin A1C (HbA1c) allow assessment of longer-term glycemic control and can aid in interpretation of blood glucose curves. Glucose binds irreversibly to serum proteins and hemoglobin, and these products persist for the life of the protein; the resultant products can be measured in serum or whole blood respectively. Fructosamine indicates adequacy of glycemic control over the previous 1-2 weeks, while HbA1c reflects glycemic control for the previous 4-6 weeks 14,15,16.

Urine glucose

Measurement of urine glucose concentration can also be helpful for assessment of glycemic control, and is particularly useful in cats that are going into remission as well as detecting any relapse of diabetes after remission. Urine glucose should not be used to determine the daily dose of insulin, but glycosuria trends can be very helpful in assessing diabetic control, especially if assessed on a consistent basis and recorded in a diary or log. The presence of urinary ketones can also help indicate impending diabetic ketoacidosis.

Continuous glucose monitoring

CGM (continuous glucose monitoring) systems now allow continuous evaluation of the interstitial blood glucose concentration for up to 14 days via a small flexible catheter placed subcutaneously. The addition of CGM to the veterinarian’s toolbox has improved the ability to accurately monitor insulin-treated diabetic cats, and is more sensitive than traditional glucose curves for detection of hypoglycemia. The method allows the insulin dose to be titrated in real time, and adjusted for differences between day- and night-time requirements. Validation studies in veterinary patients have shown that the interstitial glucose concentration correlates well with the blood glucose concentration in most situations, and the current CGM systems used in veterinary medicine are affordable, easy to apply and use, and well tolerated by patients. They allow an integrated analysis of changes in interstitial glucose in the patient over a 14-day period. The most common device currently in use is the Freestyle Libre^TM (FSL) 14 day interstitial glucose monitor (Figure 2), which has been validated in cats. The Freestyle Libre^TM 2 and 3 models have also been used in cats, but their accuracy has yet to be fully reported in peer-reviewed literature. There are other continuous glucose monitors on the market, including the Dexcom-CGM^TM and Eversense CGM^TM, but again these systems have not yet been evaluated in cats.

In terms of indications for use, the FSL is now an important tool for treating patients with diabetic ketoacidosis, in patients with newly diagnosed DM, and in unstable diabetic patients, where it can be used continuously until better glycemic control is achieved. It is also very useful for routine intermittent monitoring of stable patients.

Accuracy of the FSL monitor

Several studies have investigated the accuracy of the 14-day FSL device in cats 5,6,7,8. In these studies, interstitial glucose measured by the sensor correlated well with both peripheral blood glucose measured by a point-of-care (POC) glucometer and an automated biochemistry analyzer. Most studies indicate that the FSL slightly underestimates the blood glucose compared to the actual value, but evaluation using a surveillance error grid analysis indicates good clinical accuracy (Figure 3) 5. It is important to recognize that there is a lag of up to 30 minutes between changes in the blood glucose and changes in the interstitial glucose, so measurements may differ somewhat 8, and the difference between the two measures is most marked when blood glucose is changing quickly. Most studies have shown a slightly poorer correlation between blood glucose and interstitial glucose in the hypoglycemic range, but this may be due to the smaller number of hypoglycemic samples in published studies, as well as the effect of rapid changes in the blood glucose.

Assessment of the Libre 2 sensor has to date only been published as a single abstract 17, which noted that the sensor slightly underestimated the blood glucose in the mildly hypoglycemic and euglycemic ranges, while it overestimated levels at very low glucose concentrations (< 49 mg/dL/2.7 mmol/L).

Overall, studies suggest that for most diabetic cats, the difference between the interstitial glucose and the blood glucose has little or no impact on clinical decision-making, and that the FSL is sufficiently accurate for monitoring affected cats. The device has not yet been evaluated in cats with diabetic ketoacidosis (DKA), but in the author’s hospital it is very helpful for cats in this category, and it is known that in dogs the device’s performance is not affected by ketosis; however, there is lower accuracy in dehydrated animals 18,19. Skin thickness has also been shown to influence the accuracy of the device in dogs 20, but this has not been evaluated in cats.

Using the FSL monitor

FSL 14 day sensor is a single use disposable device with a diameter of 35 mm and a thickness of 5 mm (Figure 4), allowing interstitial glucose measurements to be accessed in real time by swiping a scanner over the sensor. A dedicated reader can be purchased and used multiple times with sequential sensors, which is advantageous for hospitalized patients, or there are free apps for most Android or iPhones that can be used to scan the sensor. With either option the data can then be uploaded to a computer or the LibreView website using free software. For the Freestyle Libre^TM 14 day sensor, the reader and the app can be used together as long as the reader is initially used to set up the sensor. Note that this is not true for the Libre 2 sensor, where the reader and phone app cannot be used interchangeably. Purchase of the sensor and reader from a retail pharmacy requires a prescription in the United States, but not in most other parts of the world.

To prepare for sensor placement, an area of skin (approximately 5 cm x 5 cm) slightly bigger than the sensor should be shaved and cleaned with an alcohol swab. The sensor pack is loaded into the applicator (Figure 5), and 4-8 drops of tissue adhesive are placed in a clock pattern on the underside of the disc. The applicator is then deployed, making sure that it is held at right angles to the skin surface and avoids bony prominences. As the sensor is deployed, the needle deposits the probe subcutaneously, leaving the disc adhered to the skin surface. A reader or smartphone is then used to start the sensor, which is ready to use 60 minutes later. There are numerous possible sites on an animal that can be used, but the best location is typically the dorso-lateral chest wall or between the shoulders (Figure 6). It is important to avoid any contact with collars or harnesses that may rub on the sensor.

Depending upon the individual patient, the sensor can be left uncovered, or protected by an adhesive patch, a T-shirt, pet sweater, or similar (Figure 7). A covering should certainly be used in active patients or where there are housemates that might attempt to interfere with the sensor; there is no need to remove any covering in order to retrieve data from the reader. Although the sensor is waterproof, it is not recommended to bathe the pet or allow it to swim while the sensor is in place. Once the sensor has expired it can easily be removed by gently peeling it from the skin, if necessary using alcohol or baby oil to remove the glue.

The FSL measures the interstitial glucose every minute and stores this data every 15 minutes on the sensor disc for up to 14 days. The disc can store up to 8 hours of data, but every time the sensor is scanned the data is downloaded onto the reader or mobile phone. The sensor can be scanned at any time, but in order to obtain continuous readings it should be scanned at least every 8 hours to prevent the data from being over-written. The data can then be uploaded to a computer or the LibreView website and viewed on-line or as a pdf file any time during the life of the sensor. The LibreView website allows data from multiple patients to be stored in the cloud, and can be accessed by both the owner and the veterinary care team. The free software allows the user to generate a summary report, which can be viewed on-line or downloaded as a pdf.

Complications of CGM

Although in general there is good correlation between blood glucose and interstitial glucose, problems can arise with the use of the sensor. These include error messages, delays in reporting the measured glucose, persistently high or low readings that do not correlate with the clinical picture, gaps in the data, and rapid fluctuations in the reported interstitial glucose (Figure 8). Total sensor failure may also occasionally occur. Another small point is that although the device measures glucose concentrations between 40 and 500 mg/dL, the graphs generated in the reports do not display glucose concentrations greater than 350 mg/dL (19.4 mmol/L). If there is any doubt about the accuracy of the readings, blood glucose should be checked with either a validated POC glucometer or an automated biochemistry analyzer. Complications with the patient can also sometimes develop; these include erythema at the placement site, and (rarely) abscess formation, so if sequential sensors are to be used in an individual patient, the placement site should be varied to avoid using the same location twice. Note also that although the FSL can measure interstitial glucose for up to 14 days, early detachment occurs in many patients; in cats the average sensor life is approximately 8 days.

Interpretation of data

The FSL summary report, which can be viewed on the manufacturer’s website or via free downloadable software, has a number of different viewing options. The Daily Log and Weekly Summary show curves from individual days, while the Glucose Patterns Insights and Ambulatory Glucose Report display the data averaged over time. This allows assessment of both day-to-day variability and weekly trends. The reports enable the veterinarian to evaluate the insulin dose and duration, and to determine whether there are differences in insulin requirements during the day versus overnight. This also enables accurate assessment of glycemic control in cats given once daily insulin. Another major advantage is the ability to evaluate day-to-day variability in insulin response and to determine the frequency of hypo- and hyper-glycemia events.

Interpretation of individual curves is similar to interpretation of a traditional blood glucose curve, but with the ability to better appreciate day-to-day variability. The glucose nadir, duration of insulin effect and average glucose levels can be easily determined. Ideally, the glucose nadir should fall between 80-150 mg/dL (4.4 to 8.3 mmol/L) and the glucose concentration should remain below 300 mg/dL (16.6 mmol/L) for the majority of the day. Problems that can be detected using the FSL reports include inadequate insulin dose, inadequate duration of insulin action (rapid metabolism), insulin-induced hypoglycemia (Figure 9), and lack of response to insulin; this latter problem suggests either poor client compliance or insulin resistance. Based on the assessment of the curve, a change in insulin dose or formulation can be made if needed, and the response assessed while the sensor is still in place. Because the glucose measurements are available in real time, clinically relevant hypoglycemia can also be spotted and treated immediately, and the insulin dose decreased. When using the FSL to adjust the insulin dose, it is important to wait 5-7 days between alterations, and since the sensor has a 14 day lifespan it is usually possible to make two adjustments to the dose during this period; of course, the dose can be decreased multiple times if necessary.

As mentioned already, although the correlation between the FSL and POC glucometers is usually good, sensor failure or errors can and do occur. If the FSL glucose measurements do not fit with the clinical picture, the blood glucose should be assessed using a POC glucometer or another trusted method. Indications of sensor failure include an error message, a message indicating that the sensor should be scanned again at a later time, gaps in data, and unexpectedly wide swings in the blood glucose that do not fit with clinical signs. In these situations, if the results of the FSL do not correlate with a POC device, the sensor should be replaced.

Conclusion

In summary, continuous glucose monitoring devices can be very valuable in assessment of glycemic control in cats, but having a good understanding of the technology of the chosen sensor (and its limitations and potential errors) allows maximum utilization of the technology. The devices now enable the clinician to undertake more accurate monitoring of diabetic patients, and thus potentially increase the likelihood of diabetic remission, and can now be used in most first-opinion clinics whenever necessary.