Comparison between a flash glucose monitoring system and a portable blood glucose meter for monitoring dogs with diabetes mellitus

Abstract Background Flash glucose monitoring system (FGMS; FreeStyle Libre) was recently validated for use in diabetic dogs (DD). It is not known if this system is clinically useful in monitoring DD. Objective To compare the clinical utility of FGMS against blood glucose curves (BGCs) obtained with a portable blood glucose meter (PBGM) in monitoring DD. Animals Twenty dogs with diabetes mellitus. Methods Prospective study. Dogs with diabetes mellitus on insulin treatment for at least 1 month were included. Comparisons of insulin dose recommendations based on the in‐hospital GCs acquired using FGMS and a PBGM, consecutive‐day interstitial GCs (IGCs) acquired at home using the FGMS, and consecutive‐day, home vs hospital IGCs acquired using the FGMS were made using concordance analysis. Results There was good concordance between insulin dose recommendations based on FGMS and PBGM generated GCs and IGCs obtained in the 2 different environments on 2 consecutive days, but almost absent concordance between IGCs obtained on 2 consecutive days at home. Glucose nadirs were detected in 34/43 (79%) of Ambulatory Glucose Profile (AGP) reports of the FGMS. In comparison, concordant glucose nadirs were identified in 14/34 (41%) BGCs using PBGM. The individual FGMS scans and PBGM identified 60% and 9% of low IG/hypoglycemic episodes, respectively. Conclusions and Clinical Importance Insulin dose adjustments based on BGCs can be suboptimal. The FGMS allows a more accurate identification of the glucose nadirs and hypoglycemic episodes compared to the use of a PBGM and assessment of day‐to‐day variations in glycemic control.


| INTRODUCTION
Diabetes mellitus (DM) is a common endocrine disease of dogs characterized by an absolute or relative deficiency of insulin. 1 Dogs with DM are treated with exogenous insulin and require regular monitoring to ensure appropriate dosing. Tools available to veterinarians for monitoring the response of diabetic dogs to treatment include clinical signs, body weight, glycated proteins levels, and blood glucose curves (BGCs) among others. 1 Typically, BGCs are conducted in a hospital setting or at home and involve 1-to-2 hourly blood sampling with a portable blood glucose meter (PBGM) over an 8 to 12 hour period. Evaluation of BGCs allows clinicians to determine glucose nadir, time to nadir, mean blood glucose concentration as well as assessing the degree of variation in blood glucose concentration. 1 This method has some disadvantages such as the need for repeated venipuncture, that can be stressful and painful for the animal, but also carries the risk of missing the blood glucose peak or nadir if they fall between 2 sampling times. 2 Additionally, in-hospital BGCs are time consuming, expensive and do not allow the assessment of glycemia on consecutive days. This last aspect represents an important limitation because of the variability of serial blood glucose curves in dogs and humans. 3 Twenty client-owned dogs with diabetes mellitus and admitted to the   Veterinary University Hospital between May 2015 and March 2018 were prospectively enrolled into the study.

| Dogs with diabetes mellitus
DM was diagnosed based on consistent clinical signs including polyuria, polydipsia, weakness and weight loss alongside a blood glucose concentration >180 mg/dL (>11 mmol/L) after food had been withheld for at least 10 hours, glucosuria and serum fructosamine concentration >340 μmol/L (reference interval: 222-382 μmol/L). 12 All dogs had been treated with insulin for at least 4 weeks prior to enrolment in the study. Owners provided written informed consent for inclusion of their dogs in the study. The study was approved by the local Scientific Ethics Committee for Animal Testing.

| Flash glucose monitoring system
The FreeStyle Libre Flash Glucose Monitoring System (FGMS) was used in this study and is composed of a small, lightweight disc-shaped sensor (35 mm × 5 mm). The sensor measures the IG concentration through a small, subcutaneous catheter (0.4 mm × 5 mm). Glucose detection is based on Wired Enzyme Technology, that consists of both enzymatic (glucose oxidase) and amperometric (electrodes) systems. 13 Reduction of glucose by glucose oxidase results in generation of an electric current, the intensity of which is proportional to the IG concentration.
The detection limits of the sensor are between 20 and 500 mg/ dL and measurements outside of this range are recorded as "LO" and "HI," respectively. The system is factory-calibrated and consequently does not require calibration before or during the wearing period. The sensor begins recording data 1 hour after its application and automatically measures the IG concentration every minute. IG concentrations are transferred from the sensor to a reader when the user brings the hand-held reader into close proximity to the sensor. The hand-held reader then displays the current sensor IG concentration, an IG trend arrow, as well as IG concentrations over the preceding 8 hours. Scanning can be performed as often as is needed for current IG concentration, otherwise the measurements are automatically recorded and stored on the sensor (every 15 minutes) and displayed on the reader when scanned. The reader stores data for 90 days. Data can be uploaded from the reader, using the Abbott FreeStyle Libre software to generate summary glucose reports (Ambulatory Glucose Profile, AGP). Among these, the daily log report shows IG fluctuations between 20 and 350 mg/dL during a 24-hour period ( Figure 1) and, as such, it was used in this study to provide a rough estimation of the glucose nadirs as well as the number of low glucose episodes. At the end of recording period, the sensor is fully disposable, but the reader can be reused for a new sensor. The sensor was applied as previously described 11

| Timing
Seven separate GCs were acquired for each dog during the recording period of their respective FGMS (Table 1). During the recording period, each day was divided into 2-time intervals: day-time and night-time. Day-time was approximately defined as the time interval between the morning insulin administration and the evening insulin administration (08:00-20:00). Night-time was approximately defined as the time interval between the evening insulin administration and the next morning insulin administration (20:00-08:00). On days 1, 7, and 14, paired, in-hospital GCs were acquired using the FGMS and PBGM devices. On days 5, 6, 12, and 13, home IGCs were acquired using only the FGMS device by the dogs' respective owners. On day 1 of the study, dogs were hospitalized and the sensor was applied.
For a total of 10 to 12 hours, IG glucose measurements were recorded using the FGMS on a 2-hourly basis. Capillary blood glucose was obtained from the pinna every 2 hours using the PBGM during the same period. On days 7 and 14, food and insulin were given at home and the paired GCs were started after the dog arrived at the clinic (≤1 hour after insulin administration) using the same protocol.
On the remaining days, owners acquired IGCs every 1 to 2 hours during the day-time using the FGMS-recording displayed values in a diary.
At the end of the recording period, the sensor was removed and the FGMS data were downloaded onto a personal computer using the Abbott FreeStyle Libre Software.

| Assessment of IGCs and BGCs
Based on assessment of the GCs, for each GC, 2 hypothetical insulin dose recommendations were made with the aim to maintain either more than 50% of BG/IG values between 90 and 250 mg/dL or BG/IG nadir between 90 and 180 mg/dL 15 (

| Assessment of nadirs
Glucose nadir was defined as the lowest glucose result during the day-time period. Nadirs extracted from the "daily log report" of the AGP, those scanned by the FGMS reader and those detected by the PBGM on days 1, 7, and 14 were compared (study aim 3) ( Figure 1).
The nadirs extracted from the AGP software were considered concordant with those obtained by the FGMS scans and by the PBGM if they had the same time interval (±60 minutes), from the morning injection of insulin.

| Comparison between day-time and nighttime nadirs
The day-time and night-time nadirs, extracted from the daily log report of the AGP were compared (study aim 4) and were considered concordant if they fell within the same glycemic range: <90, 90-180, and >180 mg/dL.

| Assessment of hypoglycemic episodes
The number and duration of low IG values (<70 mg/dL) were recorded from the AGP software and were compared with the FGMS and PBGM recordings (study aim 3).

| Data analysis
Statistical analysis was performed with the aid of 2 commercially available software (GraphPad Prism 7, Cran R statistical package).
Normality was assessed using D'Agostino and Pearson tests and parametric or nonparametric tests were used accordingly. Nonnormal data were reported as median and ranges while normal data were expressed as mean ± SD.
In order to compare results of different combinations of factors (in-hospital GCs acquired using FGMS and PBGM; consecutive-day IGCs acquired at home using the FGMS; home vs hospital IGCs acquired using the FGMS) and considering the impossibility to define a "gold standard" the so-called concordance analysis was used. Such a concordance has been defined in terms of decision about the insulin dose, coded as −1, 0, 1 (decrease, steady, increase). Even if a "gold standard" is not defined, this study aimed to investigate the reliability of an alternative (and much easier to use) tool of analysis, so we adopted the FGMS as the reference for PBGM performances when evaluating hypoglycemic episodes and glucose nadirs. Considered in this framework, our study is a "reliability" study, in the sense of concordance of diagnoses. A lot of measures of concordance are available in statistics, but considering our experimental framework consisting in multimodal qualitative variables, and the above statement about "reliability," following the bio-medical 16 and the statistical literature, 17  Of these, 13 dogs were neutered females, 5 neutered males, and 2 entire males. The median age was 11 years and 1 month (7 years and 2 months-13 years and 8 months), the median body weight was 6.5 kg (range, 6-64.1 kg) and median BCS was 5/9 (3/9-8/9).
Median time from the diagnosis of DM was 7.5 months At the end of the wearing period, 3/20 dogs showed mild erythema at the site of application of the sensor but was self-limiting and did not require specific treatment.

| Assessment of nadirs
The AGP was used as a "gold-standard," with glucose nadirs identified  All dogs tolerated the use of a bandage to maintain the position of the FGMS and mild erythema was noted at the site of the sensor in only 3/20 (15%) dogs. The incidence of erythema at the site of application was much lower than the 50% reported in a previous study of dogs. 11 Erythema occurs in 4% to 44% human patients with diabetes mellitus. [18][19][20][21] One potential cause for the cutaneous erythema is a dermatological reaction to the isobornyl acrylate that is contained in the sensor itself and that can migrate into the adhesive part of the device and come into contact with the skin. 22 The mild erythema does not seem to create discomfort for the animal and can be considered a mild and acceptable side effect in the use of the FGMS.

| Assessments of low IG/hypoglycemic episodes
One of the aims of this study was to compare the GCs generated using the FGMS with those simultaneously generated using a PBGM.
Following the common interpretation of the indices, the correlation showed an optimal concordance between FGMS and PBGM gener- respectively, although individual insulin dose and meal were kept constant. 3 Between-day GV is also reported in 93% of human patients with DM 4 and has been associated with daily fluctuations in the postprandial glycemic response to a standard meal, 29 variable sensitivity to insulin 30  However, the dogs were maintained in a constant hospital environment. Differences in the feeding schedules and in the amount of exercise of the diabetic pet, as well as stress due to unfamiliar environment or repeated vein punctures could have contributed to that findings. In our study, insulin dose recommended from the FGMS profiles obtained in the hospital setting were higher than those obtained at home in approximately 50% of cases. These results are in line with a previous study looking at the treatment decision based on home and hospital generated BGCs in dogs with DM. In that study, the mean and maximum glucose concentration of the hospital curves were significantly lower than those of the home curves. 39 A possible explanation for this finding is that blood glucose concentration in the clinic might be lower than those at home because of reduce appetite 39  The AGP report was used as a "gold standard" and identified the glycemic nadir in 79% of IGCs. In 12% of cases, the glucose nadirs could not be extrapolated from the AGP because the IG readings were higher than 350 mg/dL and were thus not showed by the generated graphs ( Figure 2). Additionally, in 9% of cases it was not possible to analyze the graph because it was not correctly generated owing gaps in the graph which could be the result of intermittent sensor dysfunc- are not reported (Figure 2) unless the sensor is scanned by the reader.
This aspect can limit the utility of the sensor in dogs with poorly controlled DM. Finally, to obtain continuous IG measurements the sensor needs to be scanned at least every 8 hours, which cannot always be logistically possible.
There were a number of limitations in the present study.
First, only a small number of dogs were included in the study.
Second, in the majority of the dogs the sensor lasted less than 14 days, thus limiting the number of data available for the analysis. Third, we did not evaluate if the accuracy of the system varies during the entire wearing period. However, some studies from human medicine demonstrate that accuracy of the device remain stable through 14 days of use. 18,20,54,55 Also, in the previous study in which FGMS was used in diabetic dogs mean absolute relative difference (MARD) on 14th day was only slightly higher compared to the MARD on 1st day. 11 Finally, considering the absence of a gold standard as a monitoring tool for diabetic dogs, in this study FGMS has been used as the "gold standard" to compare its ability in detecting low glucose measurements and glucose nadirs compared to the traditional BGCs because it allows a 24 hours recording time, even if in the hypoglycemic range it is less accurate.

ACKNOWLEDGMENTS
No funding provided for this study. Some of the current results were presented as an oral abstract at the 29 th ECVIM-CA Congress, September 19-21, 2019, Congress Centre Mico, Milano, Italy.