Glycemic monitoring and control are important in the management of some canine patients under general anesthesia, including diabetics, neonates, and those with significant hepatic disease or sepsis. Clinical signs recognized in conscious hyper- or hypoglycemic patients, such as altered mentation, vocalization, visual impairment, gastrointestinal signs, restlessness, or appetite changes, are not appreciated while a patient is under general anesthesia.1 The current clinical recommendations for blood glucose monitoring of at-risk patients is to perform a blood glucose measurement every 30 to 60 minutes during the intraoperative period.2 Veterinary point-of-care glucose monitors (POCGM), validated for use in canine and feline patients using venous or capillary blood, are frequently utilized in the clinical setting.3 During general anesthesia, obtaining a blood sample for this method of glucose measurement may be limited by visibility of and access to sampling sites and anesthetic-induced physiologic changes causing decreased perfusion to peripheral tissues. Such physiologic changes may include decreased cardiac output, peripheral vasoconstriction, and hypothermia. Inhalant anesthetics induce dose-dependent vasodilation and cardiac depression, resulting in hypotension.4 Conversely, pain, surgical manipulations, or other stressors may lead to vasoconstriction and hypertension. These blood pressure changes may affect tissue perfusion, altering the ability for glucose to diffuse from capillaries into interstitial tissues.
A flash glucose monitoring system (FGMS; FreeStyle Libre 14 Day; Abbott) developed for humans has been validated for use in dogs with diabetes mellitus.5–9 The FGMS measures interstitial glucose (IG) concentrations, which have been found to correlate well with blood glucose, with a slight lag time of 5 to 12 minutes for blood glucose to equilibrate with the interstitium when blood glucose concentrations are changing.5,7,9–11 The FGMS can be worn for up to 14 days and measures IG concentrations every 60 seconds via a disposable sensor with a small catheter that is inserted under the skin.12 Interstitial glucose concentrations, ranging from 40 to 500 mg/dL, are displayed on a hand-held reader or smartphone application, and no calibration is required.12 The use of an FGMS has become attractive in veterinary medicine for its minimally invasive nature, and it is not uncommon to have diabetic animals present for anesthesia with an FGMS already in place. This system has many advantages, including reduction of direct blood sampling, ease of sampling, and continued ability to monitor glucose, with little to no stress during the pre- and postanesthetic period by the veterinary care team and pet owner. Localized erythema around the sensor, bending of the catheter, and possible ingestion are of the minimal risks reported with use of the FMGS.5,6 A newer version of the FGMS, the FreeStyle Libre 2 (Abbott), was released in 2020 with the same characteristics as its predecessor, adding an alarm feature. It was also determined to be more accurate in adults and children.13 However, the accuracy of this version of the FGMS in dogs has not been reported.
The accuracy of any FGMS during general anesthesia in dogs has not been established, and there are no studies describing the influence of different hemodynamic states that commonly occur during anesthesia on IG measurements. The primary aim of this study was to assess the accuracy of IG measurements obtained from an FGMS (FreeStyle Libre 2; Abbott) in healthy dogs during isoflurane anesthesia compared to a common point-of-care blood glucose monitor (blood glucose point-of-care; BGPOC) and reference plasma glucose measurements using the glucose oxidase-peroxidase method (blood glucose gold standard; BGGS). The secondary aim was to determine whether anesthesia-induced changes in mean arterial blood pressure (MAP) affected FGMS accuracy in dogs. We hypothesized that the FGMS would provide clinically accurate IG measurements in healthy dogs under inhalant anesthesia during periods of normotension with a decrease in accuracy during periods of hypo- and hypertension.
Methods
This study was reviewed and approved by the University of Wisconsin-Madison School of Veterinary Medicine IACUC (protocol V06612).
Animals
Six purpose-bred intact adult 1-to-2-year-old 8.9 ± 1.8 kg (mean ± SD) Beagle dogs (n = 3 female, n = 3 male) were utilized for this study. Dogs were housed in individual climate-controlled kennels with a 12-hour light cycle. A balanced commercial canine diet was fed twice daily with ad libitum access to water. Behavioral and environmental enrichment was provided. All dogs underwent a minimum 1-week acclimation period prior to the study. The dogs were determined to be healthy on arrival based on physical examination, CBC, serum biochemical analyses, PCV, and total protein. All bloodwork performed was within acceptable limits. An American Society of Anesthesiologists physical status of I was assigned to all dogs prior to each anesthetic event based on a preanesthetic physical examination including vital signs and hydration status.14 At the end of general anesthesia, the FGMS sensors were removed, and the dogs were monitored until fully recovered. The dogs were returned to individual housing, and a normal feeding schedule was resumed. Following completion of this study, all dogs were transferred to a different study protocol.
Study design
This study was a blinded, randomized crossover design. Each dog underwent 3 randomized treatments (www.randomizer.org) under general anesthesia, with a minimum 6-day washout period between each treatment, for a total study duration of 16 days. Treatment groups consisted of normotension (MAP, 80 to 100 mm Hg), hypotension (MAP, 45 to 60 mm Hg), and hypertension (MAP, 120 to 160 mm Hg). Dogs were fasted for 12 hours before general anesthesia with free access to water. The evening before each anesthetic event, a new FGMS sensor (FreeStyle Libre 2; Abbott) was applied to the assigned dogs. This was to account for the manufacturer's recommended 12-hour acclimation period. A 5 X 5-inch area was clipped on the right or left side of the neck, just cranial to the scapula, and prepared aseptically. The initial application site was determined randomly via coin flip and then alternated for each subsequent sensor application. A small amount of tissue adhesive (Vetbond; 3 M) was applied to the outermost edge of the sensor prior to application to ensure secure attachment of the sensor to the skin. The sensor was activated using the FreeStyle Libre 2 iOS application (Abbott) or the FreeStyle Libre 2 hand-held reader (Abbott). A light bandage (Vetrap 4 inch; 3 M) was applied around the neck to cover the sensor and prevent accidental removal overnight.
On the day of each treatment, FGMS sensors were scanned to verify functionality prior to premedication, and the IG was recorded as baseline. Blood samples were obtained from a lateral saphenous vein for baseline BGGS and from the lip mucosa for baseline BGPOC. Dogs were premedicated with butorphanol (0.3 mg/kg; Torbugesic, 10 mg/mL, IM; Zoetis) for sedation to allow IV catheter placement. Butorphanol was chosen as a sedative due to its ability to achieve moderate sedation with minimal adverse effects on the cardiovascular system and no effects on blood glucose concentrations.15 After 20 minutes, a 20-gauge over-the-needle IV catheter was aseptically inserted into the left or right cephalic vein. Animals were preoxygenated with a face mask for 5 minutes, and anesthesia was induced with propofol IV, titrated to effect until orotracheal intubation was possible (≤ 5 mg/kg; Diprivan, 10 mg/mL; Fresenius Kabi). Following orotracheal intubation, each dog was connected to a circle breathing system, and anesthesia was maintained with isoflurane in 100% oxygen. Volume-controlled positive pressure ventilation with an adult respiratory circuit was utilized to maintain a target end-tidal partial pressure of carbon dioxide between 35 and 45 mm Hg, a respiratory rate of 8 to 12 breaths per minute, and a peak inspiratory pressure of 8 to 10 cm H2O. Isoflurane percentage was adjusted to maintain parameters set by the study design for each treatment group. Dogs were monitored with a pulse oximeter, ECG, capnometer, esophageal thermometer, and invasive arterial blood pressure connected to a multiparameter monitor (Mindray Datascope Spectrum). Heart rate, respiratory rate, peripheral oxygen saturation, systolic arterial blood pressure, MAP, diastolic arterial blood pressure, end-tidal partial pressure of carbon dioxide, end-tidal isoflurane, and oxygen flow rate were recorded every 5 minutes. Heat support was provided to maintain a temperature of ≥ 37.2 °C as needed with a forced-air warming device (Bair Hugger; 3 M).
After induction, all dogs were positioned in lateral recumbency with the FGMS facing upwards. A 22-gauge over-the-needle IV catheter was aseptically placed in a dorsal pedal or coccygeal artery, and a pressure transducer attached to a noncompliant fluid line filled with 0.9% sodium chloride was attached to the arterial catheter for invasive arterial blood pressure measurement. The transducer was positioned and zeroed to atmospheric pressure at the level of the right atrium. All dogs received lactated Ringer solution at 5 mL/kg/h, IV, throughout anesthesia. Instrumentation and sample collection were performed by an anesthesia resident (EAP) who was blinded to the treatment group.
Blood samples and IG reading were collected prior to the induction of anesthesia (baseline). The start of data collection (T0) was determined once the animal was instrumented and treatment parameters for blood pressure were met. Samples and IG readings were collected at T0 followed by 5, 10, 20, 30, and 40 minutes. Whole blood for plasma BGGS measurement (1 mL) was collected from the lateral saphenous vein using a 22-gauge needle attached to a 3-mL syringe and placed into a sodium fluoride potassium oxalate vacutainer tube. Samples were submitted to the clinical pathology laboratory within 30 minutes of collection for immediate processing and analysis using the glucose oxidase-peroxidase method (VITROS 4600 Chemistry System; Ortho Clinical Diagnostics; calibrated according to manufacturer recommendations). A superficial puncture with a 25-gauge hypodermic needle at the marginal lip mucosa was used to obtain 0.3 μL of whole blood for immediate glucose measurement (BGPOC) via a POCGM (AlphaTrak2; Zoetis) that was calibrated in accordance with manufacturer recommendations when a new vial of test strips was opened. Interstitial glucose was measured from the FGMS sensor using either the cellphone application or hand-held sensor.
Anesthesia was monitored, and treatments were administered by a single Diplomate of the American College of Veterinary Anesthesia and Analgesia (AMS). Normotension was maintained by adjusting the isoflurane concentration and administration of IV fluid boluses (5 mL/kg administered over 15 minutes; lactated Ringer solution) or an anticholinergic drug (glycopyrrolate; 0.01 mg/kg, IV; 0.2 mg/mL; Somerset) as needed to maintain a MAP of 80 to 100 mm Hg. Hypotension was induced and maintained by adjusting the isoflurane concentration until the MAP was < 60 mm Hg. Hypertension was induced and maintained with a constant rate infusion of norepinephrine (0.1 to 0.5 μg/kg/min, IV; 1 mg/mL; Amneal). Norepinephrine was diluted to a concentration of 0.01 mg/mL with 0.9% NaCl to a total of 20 mL and administered at a rate of 0.6 to 3 mL/kg/h to maintain a MAP ≥ 120 mm Hg. The hypotensive and normotensive treatment groups received 0.9% NaCl at the rate of 1 mL/kg/h.
Statistical analysis
Data were assessed for normality using the Shapiro-Wilk test. Parametric data is reported as mean ± SD and nonparametric data as median (IQR). A linear mixed-model analysis for repeated measures was used to analyze the influence of MAP on IG measurements with time and treatment as fixed effects while adjusting for weight. Least square means and pairwise comparisons followed by Tukey-Kramer post hoc analysis were used to test for differences between treatments and time points. Shrout-Fleiss intraclass correlation (ICC) estimates between glucose measurement methods were calculated based on a fixed set to evaluate the reliability of the 3 methods. Extended Bland-Altman analysis was used to compare the FGMS to the reference methods for each measurement, and regression lines with 95% limits of agreement were calculated for each plot.
Clinical accuracy was assessed according to the human International Organization for Standardization (ISO) 15197:2013 criteria for an FGMS, with acceptable accuracy defined as ≥ 95% of IG results measuring within ± 15 mg/dL when BGGS ≤ 100 mg/dL and within ± 15% of BGGS when it is > 100 mg/dL.16 Additionally, ≥ 99% of measured IG values must fall within zones A and/or B of the Parkes consensus error grid analysis (EGA) for type 1 diabetes.16,17 The EGA is used to compare the IG and BGGS measurements and divides them into 5 “zones,” depending on the degree of difference, and expresses a potential clinical outcome based on a given value should treatment be elected. The zones are defined as: A, < 20% deviation from BGGS or both IG and BGGS ≤ 70 mg/dL (treatment would have no effect on clinical action); B, > 20% deviation from BGGS but leads to no treatment or benign treatment; C, overcorrection of acceptable blood glucose concentrations; D, failure to detect blood glucose errors and treatment may cause substantial risk to patient; and E, erroneous treatment that may be harmful to patient.17 Statistical analyses were conducted using SAS (version 9.4; SAS Institute Inc) and Microsoft Excel for Mac (version 16.86; Microsoft Corp). All reported P values are 2-sided, and < .05 was used to define statistical significance.
Results
To detect a 15-mg/mL difference in glucose concentrations between the FGMS and the reference method with a power of 0.8 and a level of significance of 5%, a sample size of 10 paired measurements for each treatment group was required. The calculation assumed an SD of ± 11.4 mg/dL. For Bland-Altman analysis, it is recommended to have an overall total of at least 100 data points for narrower CIs; therefore, 6 dogs were chosen to provide 36 paired IG measurements to both BGGS and BGPOC for each treatment group, resulting in 108 total paired measurements for each comparison. All dogs completed the study without complications. A total of 21 FGMS sensors were placed. Two dogs removed their sensors overnight on 1 occasion each (n = 2), and 1 sensor read “sensor error” the morning of the procedure. All 3 sensors were replaced the morning of general anesthesia with a 60-minute sensor startup time. Data were collected from a total of 18 different sensors (6 per treatment group).
A total of 378 measurements (126 for each glucose measurement method) were collected. Baseline measurements prior to anesthesia (54 total) were not included in the analysis, resulting in 324 individual glucose measurements (108 for each method) used for analysis. Overall medians (IQR) for IG, BGGS, and BGPOC were 119 (105 to 149) mg/dL, 112 (100 to 119) mg/dL, and 146 (126 to 168) mg/dL, respectively. Interstitial glucose measured higher than BGGS (P = .0004), with a median difference of 9 (−2 to 20) mg/dL, and measured lower than BGPOC (P < .001), with a difference of −27 (−37 to −11) mg/dL. There was an overall significant increase in the difference between IG and BGPOC (IG – BGPOC at 20, 30, and 40 minutes compared to T0 (P = .0232, .0309, and .0045, respectively). There were no other differences in IG – BGPOC or the difference in IG and BGGS (IG – BGGS at any other time points. Median (IQR) glucose measurements during normo-, hypo-, and hypertension are summarized in Table 1. The percentages of underestimated and overestimated IG readings in the 3 blood pressure ranges are presented in Table 2.
Median (IQR) interstitial glucose (mg/dL) obtained by the flash glucose monitoring system (IG), plasma glucose by oxidase-peroxidase method (BGGS), and blood glucose by point-of-care monitor (BGPOC) and the differences between IG and BGGS (IG – BGGS) and BGPOC (IG – BGPOC) during normotension (mean arterial pressure [MAP], 45 to 60 mm Hg) and induced hypotension (MAP, 45 to 60 mm Hg) and hypertension (MAP, 120 to 160 mm Hg) in healthy dogs under isoflurane anesthesia.
| Measurement (mg/dL) | Normotension | Hypotension | Hypertension |
|---|---|---|---|
| IG | 115 (99 to 137) | 137 (111 to 149) | 114 (106 to 152) |
| BGGS | 111 (100 to 119) | 113 (99 to 118) | 112 (106 to 145) |
| BGPOC | 144 (123 to 156) | 158 (126 to 173) | 143 (130 to 173) |
| IG – BGGS | 6 (−3 to 18) | 20 (12 to 30)a | −0.5 (−3 to 11)b |
| IG – BGPOC | −29 (−47 to −12) | −26 (−33 to −8)a | −28 (−37 to −15) |
Percentage of IG measurements overestimating, underestimating, or identical to BGGS for all measurements combined and during periods of normotension (MAP, 45 to 60 mm Hg) and induced hypotension (MAP, 45 to 60 mm Hg) and hypertension (MAP, 120 to 160 mm Hg) in healthy dogs under isoflurane anesthesia.
| Combined glucose values | Normotension | Hypotension | Hypertension | |
|---|---|---|---|---|
| Percentage of overestimated glucose values | 67.6 | 63.9 | 91.7 | 44.4 |
| Percentage of underestimated glucose values | 28.7 | 27.8 | 8.3 | 50.0 |
| Percentage of identical glucose values | 3.7 | 8.3 | 0 | 5.6 |
A high degree of reliability was found between IG measurements and both BGGS (ICC, 0.962) and BGPOC (ICC, 0.953). Extended Bland-Altman plots including regression lines and 95% limits of agreement can be found in Figure 1. The bias (mean difference) for IG – BGGS was 17.3 ± 26.0 mg/dL and −17.8 ± 27.7 mg/dL for IG – BGPOC. The slopes of the IG – BGGS and IG – BGPOC regression lines were 0.5 and 0.53, respectively.
Extended Bland-Altman plots of (A) the difference between interstitial glucose (IG) obtained by the flash glucose monitoring system (FGMS) and blood glucose obtained by the oxidase-peroxidase method (BGGS) and (B) the difference between IG obtained by the FGMS and blood glucose obtained by the point-of-care monitor (BGPOC) in healthy dogs under general anesthesia (n = 6). The short-long dashed line represents the line of regression, and the dashed lines represent the 95% limits of agreement (LoAs).
Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.24.08.0242
The percentages of all IG measurements within ± 15 mg/dL of BGGS when BGGS was > 100 mg/dL and within ± 15% of BGGS when BGGS was > 100 mg/dL were 62.1% (18 of 29) and 63.3% (50 of 79), respectively. Parkes consensus EGA demonstrated that 100% of IG measurements fell within zones A and B (Figure 2).
Parkes consensus error grid analysis of IG measurements obtained by the FGMS and the percentage of measurements within zones A through E. IG was plotted against plasma glucose values obtained by the BGGS from healthy dogs (n = 6) under general anesthesia. The zones delineate the magnitude of risk: < 20% deviation from BGGS or both IG and BGGS ≤ 70 mg/dL (treatment would have no effect on clinical action; zone A), > 20% deviation from BGGS but leads to no treatment or benign treatment (zone B), overcorrection of acceptable blood glucose concentrations (zone C), failure to detect blood glucose errors and treatment may cause substantial risk to patient (zone D), and erroneous treatment that may be harmful to patient (zone E).16
Citation: American Journal of Veterinary Research 2025; 10.2460/ajvr.24.08.0242
Discussion
The results of this study indicate that the FGMS demonstrated acceptable clinical accuracy in dogs undergoing isoflurane anesthesia for monitoring IG despite noncompliance with the human ISO standards. This is consistent with previous studies5,6,8,11 evaluating the FGMS in conscious diabetic dogs. Perianesthetic blood glucose monitoring is important in the prevention of adverse outcomes in certain anesthetized animals. Standard protocols for reductions in insulin administration for diabetics on the morning of a procedure and fasting of any high-risk animal can result in unpredictable blood glucose concentrations under anesthesia. In humans, it is well documented that hyperglycemia during anesthetic events is associated with adverse clinical outcomes, such as an increased risk of infections, acute renal failure, length of hospital stay, and mortality.18,19 Gandhi et al20 reported that for each incremental change in intraoperative serum glucose by 20 mg/dL above 100 mg/dL, there was a 30% increase in adverse events relating to pulmonary and renal injuries and death. Other human studies21 have shown that intraoperative glucose concentrations correlate with risk of infection and cardiovascular and neurologic complications. Hypoglycemia may go unrecognized under anesthesia and is most commonly associated with insulin administration.22 The incidence of anesthetic complications in diabetic dogs undergoing general anesthesia is lacking in the published literature. However, the incidence of moderate-to-severe hypotension in diabetic dogs relating to their hyperglycemic and dehydration status has been described.23 Optimal glucose range and management during the perianesthetic period in veterinary patients is widely debated, and no specific standards or guidelines have been established. In humans, management of patients has moved toward moderate and individualized glycemic targets rather than tight regulation.18
The current recommendations in veterinary medicine for monitoring diabetics and other animals prone to extreme glycemic changes under general anesthesia are to check a blood glucose sample every 30 to 60 minutes for the entire duration of the procedure.24,25 The most common method used to measure blood glucose in anesthetized veterinary patients is a POCGM that requires a very small sample of whole blood obtained by a venous blood draw or by pricking an area with a high density of superficial capillaries, such as the lip mucosa, distal ear pinnal margin, carpal pad, or metacarpal/metatarsal pad. During the intraoperative period, access to these sites may be limited, or decreased peripheral perfusion from hypothermia, vasoconstriction, or hypotension could create a challenge for acquiring even a small amount of blood. The lateral aspect of the neck was chosen as the primary site for the application of the FGMS sensor because it was easily accessible, is a location commonly used in practice for application, and allowed for the placement of a protective bandage.5 Alternative acceptable sensor sites described in the literature using previous models of the FGMS include over the lumbar musculature, between the scapulae, the lateral thorax, and over the hip.26,27 The studies26,27 assessing these locations demonstrated there was no significant clinical difference between the sensor locations. Assuming that the most common surgical positioning of a dog is in dorsal recumbency with the head toward the anesthetist, placement on the lateral neck may also provide easy access for readings. However, alternative sites may be needed for those procedures requiring different positioning. Sensor location should also be considered as prolonged pressure to the device may result in unusually low IG readings due to compression of tissues from reduced blood flow.28 Other potential sources of sensor interference or failure under general anesthesia could include patient warming devices, electrocautery, and excessive contact with fluids, but the incidence of these has not been reported.
The FGMS investigated in this study provided a noninvasive, continuous method for monitoring glucose concentrations in anesthetized dogs. An older method of continuous glucose monitoring has previously been evaluated in dogs under anesthesia, but poor agreement was found between the IG and blood glucose measurements.29 The FGMS used in the current study has been evaluated in humans under anesthesia with good correlation to a POCGM and was beneficial in characterizing glycemic trends compared to individual blood glucose readings.30
Interstitial glucose as measured by the FGMS in this study was found to correlate well with the reference method for measuring plasma glucose and the veterinary POCGM. Similar to other studies6–9,11,31 utilizing a previous version of the FGMS, IG regularly measured lower than BGPOC in the dogs. However, IG was more likely to measure higher than BGGS, especially during periods of hypotension (Table 2). This is inconsistent with a study by Silva et al,11 where the previous version of the FGMS was compared in dogs with diabetic ketoacidosis to the same POCGM used in the current study, and reference plasma glucose was measured by the hexokinase method.32,33 The FGMS measured lower than both the POCGM and reference methods. Conversely, Corradini et al5 found that IG (measured by the previous version of the same FGMS) in healthy diabetic dogs measured higher than plasma glucose (hexokinase method).
Several factors can interfere with IG concentrations due to the constant diffusion of glucose across capillary walls into the interstitium. Glucose concentrations in plasma and the interstitium are typically balanced; however, diffusion is not immediate, and in humans there is an observed lag time of 4 to 10 minutes for equilibration.34,35 A study10 in dogs demonstrated a similar delay of 5 to 12 minutes, and another in cats indicated a lag time of up to 30 minutes.7 This delay is unlikely to result in significant discrepancies between interstitial and blood glucose unless glucose concentrations are rising or falling rapidly.34–36 Various physiologic changes could alter the glucose diffusion rate. In dogs with diabetic ketoacidosis, there was a significant difference in the accuracy of the FGMS readings from those that were dehydrated compared with those that were euhydrated.11 In a study performed by Del Baldo et al,31 skin thickness had an effect on FGMS measurements, with increased accuracy in dogs with thicker skin (> 5 mm); in dogs with thin skin (< 5 mm), the clinical accuracy was low, and results should be interpreted with caution. The method of plasma glucose measurement may also affect comparisons between FGMS and plasma glucose since most studies used the hexokinase method, whereas the oxidase-peroxidase method was used in the current study. The hexokinase method for glucose measurement is considered more specific than the oxidase-peroxidase method; however, it is more costly, and it was not available at the investigators’ university veterinary clinical pathology laboratory. Nevertheless, studies32,33 have shown a high correlation between the 2 methods.
Commonly used anesthetic agents in veterinary medicine, including dexmedetomidine and isoflurane, have been shown to inhibit insulin secretion and increase hepatic glucose production in various species, which could lead to increasing plasma glucose concentrations over time.37,38 Additionally, stress associated with hospitalization, anesthesia, and surgical procedures could trigger higher concentrations of circulating catecholamines, cortisol, and glucagon, resulting in acute changes in plasma glucose, which may not be identified by the FGMS immediately.39 Hemodynamic alterations induced by anesthetic drugs may also lead to discrepancies between IG and plasma glucose. Hypotension is a common anesthetic complication, often caused or exacerbated by inhalant anesthetics.4 With newer-generation volatile anesthetics, the dose-related decrease in arterial blood pressure in animals is due to both a decrease in systemic vascular resistance and from a decrease in cardiac output, unlike in humans, where vasodilation is the primary mechanism.4 Regional blood flow to certain tissues may be affected by hypotension, and thus delivery and diffusion of glucose into and out of the interstitium may be delayed or restricted. This could explain the difference in IG – BGGS between dogs during induced hypotension compared to when they were normo- or hypertensive. Given that most IG measurements overestimated plasma glucose during hypotension, the authors speculate that glucose may be slow to leave the interstitium, and cellular usage of glucose decreases when tissue perfusion is reduced.
The difference in IG and BGGS was smallest during induced hypertension (−0.5 [−3 to 11] mg/dL), with the FGMS more likely to give a slightly lower reading compared to plasma glucose concentrations. The lower measurement is likely due to the delay in IG change previously discussed. Hypertension was induced in this study using a constant rate infusion of norepinephrine, a mixed α1-, α2-, and β1-adrenergic agonist. Despite β1 stimulation increasing cardiac contractility, the effects of norepinephrine on blood pressure are primarily attributed to α1 stimulation, which results in arterial and venous vasoconstriction in dogs and humans.37,40 Norepinephrine closely mimics hypertension from sympathetic nervous system stimulation and both stimulates glucagon production and decreases insulin secretion, helping to maintain glucose homeostasis.40 Sympathetically induced changes in blood and IG concentrations that could occur during anesthesia may be more closely represented by the effects of norepinephrine compared to the vasoactive abilities of an α1-selective agonist drug, such as phenylephrine.40
The FGMS failed to comply with human ISO 15197:2013 standards for analytical accuracy when compared to the oxidase-peroxidase glucose measurement method in this study.16 Only 62.1% of IG measurements were within ± 15 mg/dL when BGGS was ≤ 100 mg/dL, and 63.3% were within ± 15% of BGGS when it was > 100 mg/dL. This is consistent with other studies5–8,11 evaluating the accuracy of a previous version of the FGMS in veterinary patients. Nevertheless, these same studies recommend using the FGMS for glucose monitoring in diabetic animals because they demonstrated good clinical accuracy with EGA.5–8,11,17 In the current study, Parkes consensus EGA also showed good clinical accuracy for the FGMS, with 100% of the readings falling in zones A and B, which are considered the clinically acceptable zones. Agreement between the FGMS and both the reference method and the point-of-care method was strong, with excellent reliability found between IG and BGGS (ICC, 0.962) and between IG and BGPOC (ICC, 0.953). However, as plasma glucose concentrations increase, the difference between IG and BGGS increases (Figure 1). Correlation between a previous version of the FGMS and plasma glucose or POCGMs in other veterinary studies has been strong; however, when separated into low, normal, and high blood glucose ranges, correlation decreases considerably.5–9,11 Possible reasons given for the decrease in correlation included a smaller sample size in each subgroup, rapid changes in blood glucose concentration, or subclinical changes in hydration and perfusion.5–9,11 We did not evaluate the accuracy of the FGMS within different glucose ranges due to the small number of measurements outside of the normal glucose range. In a study performed by Howard et al,41 the FreeStyle Libre 14 Day sensor was assessed for accuracy in healthy conscious dogs after a bolus of insulin to induce hypoglycemia. There was limited agreement between the FGMS and the reference method (oxidase-peroxidase method), especially within the hypoglycemic range.41 This was likely due to the rapid change in blood glucose induced by the insulin bolus and a lag in the change of IG. Verification of low or elevated IG measurements with a POCGM or reference method should be considered.
Application of the FGMS was quick and well tolerated by all dogs in the study. Obtaining a blood sample via venipuncture or mucosal puncture was difficult on several occasions, whereas the FGMS provided rapid glucose measurements. Due to the mandatory 60-minute warmup time of the FGMS, we recommend that it be placed at the time of patient intake or the day before a procedure. While the FGMS will continuously monitor IG for up to 14 days, we did not assess whether the accuracy of the sensor could be affected at various times within that 2-week period. The reported accuracy of a previous version of the FGMS over time has been conflicting, with a slight increase to no change.5,11,42,43 In conscious animals, accidental or intentional removal of an FGMS sensor may occur, especially if placed in an area with high locomotion. Two dogs removed their sensor overnight, requiring a new one to be placed the morning of anesthesia. An additional sensor was not functioning the morning of the study. These instances required the mandatory 60-minute warm up period, delaying the start of anesthesia. After removal of the sensor, mild erythema was noted at the site of application in a few of the dogs, likely secondary to the tissue adhesive used to ensure the sensor remained in place. This did not appear to have any negative effect on the dogs and resolved within 24 hours.
An important clinical limitation of utilizing the FreeStyle Libre 2 is the lack of interchangeability between the smartphone application and the handheld reader.12 If the sensor is placed at home by a client prior to anesthesia and activated with the smartphone application, the hospital would be required to continue monitoring with the client's phone or place a new sensor. And if the handheld reader is used to activate the sensor in the hospital, the client would need to borrow the handheld reader to continue to monitor their pet at home. Additionally, newer versions of the FreeStyle Libre sensor cannot be used with previous versions of the handheld reader and smartphone application.
There were additional limitations to this study. A small sample size and the use of young, clinically normal beagle dogs may not fully represent the hemodynamic consequences induced by general anesthesia or surgical procedures on the population of dogs that require regular glucose monitoring. One female completed the study in heat, and another came into heat during the study; however, the phase of the estrous cycle does not appear to have a clinical effect on blood glucose in female beagles.44 We did not assess the accuracy of the FGMS across a wide range of glucose concentrations. Most of the BGGS measurements (85%) were within the normal glucose range (80 to 120 mg/dL) for a dog, with no measurements < 80 mg/dL or > 200 mg/dL.
Although the FGMS did not fully meet the ISO requirements in this study, clinical accuracy was demonstrated for the measurement of IG in dogs under general anesthesia. Nevertheless, for patients that are at a higher risk of developing hypo- or hyperglycemia during the perianesthetic period, application of the FreeStyle Libre 2 FGMS before anesthesia may provide a convenient and less invasive method for monitoring glucose. We recommend that in periods of extreme hypo- or hyperglycemia, or during periods of prolonged hypotension, a secondary method to the FGMS should be utilized to confirm accuracy. Future studies evaluating its accuracy in anesthetized diabetic patients and at lower and higher glucose ranges are advised.
Acknowledgments
The authors would like to thank the University of Wisconsin Research Animal Resources Center veterinary technicians for their continued care of the dogs and Dr. W. E. Swanton.
Disclosures
The authors have nothing to disclose. No AI-assisted technologies were used in the generation of this manuscript.
Funding
Funded by the University of Wisconsin-Madison School of Veterinary Medicine Companion Animal Grant Fund and faculty startup funds.
ORCID
A. M. Sage https://orcid.org/0000-0003-0539-0370
References
- 1.↑
Koenig A. Hypoglycemia. In: Silverstein D, Hopper K, eds. Small Animal Critical Care Medicine. 2nd ed. Elsevier; 2015:352–357.
- 2.↑
Fischer B. Endocrine disease. In: Johnson R, Snyder L, Schroeder C, eds. Canine and Feline Anesthesia and Co-Existing Disease. 2nd ed. Wiley-Blackwell; 2022:324–328.
- 3.↑
Cohen TA, Nelson RW, Kass PH, Christopher MM, Feldman EC. Evaluation of six portable blood glucose meters for measuring blood glucose concentration in dogs. J Am Vet Med Assoc. 2009;235(3):276–280. doi:10.2460/javma.235.3.276
- 4.↑
Steffey E, Mama K, Brosnan R. Inhalational anesthetics. In: Grimm K, Lamont L, Tranquilli W, Greene S, Robretson S, eds. Veterinary Anesthesia and Analgesia. 5th ed. Wiley-Blackwell; 2022:297–331.
- 5.↑
Corradini S, Pilosio B, Dondi F, et al. Accuracy of a flash glucose monitoring system in diabetic dogs. J Vet Intern Med. 2016;30(4):983–988. doi:10.1111/jvim.14355
- 6.↑
Del Baldo F, Canton C, Testa S, et al. Comparison between a flash glucose monitoring system and a portable blood glucose meter for monitoring dogs with diabetes mellitus. J Vet Intern Med. 2020;34(6):2296–2305. doi:10.1111/jvim.15930
- 7.↑
Del Baldo F, Fracassi F, Pires J, et al. Accuracy of a flash glucose monitoring system in cats and determination of the time lag between blood glucose and interstitial glucose concentrations. J Vet Intern Med. 2021;35(3):1279–1287. doi:10.1111/jvim.16122
- 8.↑
Malerba E, Cattani C, Del Baldo F, et al. Accuracy of a flash glucose monitoring system in dogs with diabetic ketoacidosis. J Vet Intern Med. 2020;34(1):83–91. doi:10.1111/jvim.15657
- 9.↑
Shea EK, Hess RS. Validation of a flash glucose monitoring system in outpatient diabetic cats. J Vet Intern Med. 2021;35(4):1703–1712. doi:10.1111/jvim.16216
- 10.↑
Rebrin K, Steil GM, van Antwerp WP, Mastrototaro JJ. Subcutaneous glucose predicts plasma glucose independent of insulin: implications for continuous monitoring. Am J Physiol. 1999;277(3):E561–E571. doi:10.1152/ajpendo.1999.277.3.E561
- 11.↑
Silva DD, Cecci GRM, Biz G, Chiaro FN, Zanutto MS. Evaluation of a flash glucose monitoring system in dogs with diabetic ketoacidosis. Domest Anim Endocrinol. 2021;74:106525. doi:10.1016/j.domaniend.2020.106525
- 12.↑
FreeStyle libre 2 system. Abbott. Accessed February 13, 2024. https://www.freestyle.abbott/us-en/products/freestyle-libre-2.html
- 13.↑
Kompala T, Neinstein AB. Analysis of "accuracy of a 14-day factory calibrated continuous glucose monitoring system with advanced algorithm in pediatric and adult population with diabetes. J Diabetes Sci Technol. 2022;16(1):78–80. doi:10.1177/1932296820967004
- 14.↑
Mayhew D, Mendonca V, Murthy BVS. A review of ASA physical status – historical perspectives and modern developments. Anaesthesia. 2019;74(3):373–379. doi:10.1111/anae.14569
- 15.↑
KuKanich B, Wiese A. Opioids. In: Grimm K, Lamong L, Tranquilli W, Green S, Robertson S, eds. Veterinary Anesthesia and Analgesia. 5th ed. Wiley-Blackwell; 2022:207–222.
- 16.↑
ISO 15197:2013: in vitro diagnostic test systems—requirements for blood-glucose monitoring systems for self-testing in managing diabetes mellitus. ISO. Published May 2013. Accessed February 13, 2024. https://www.iso.org/standard/54976.html
- 17.↑
Parkes JL, Slatin SL, Pardo S, Ginsberg BH. A new consensus error grid to evaluate the clinical significance of inaccuracies in the measurement of blood glucose. Diabetes Care. 2000;23(8):1143–1148. doi:10.2337/diacare.23.8.1143
- 18.↑
Duggan EW, Carlson K, Umpierrez GE. Perioperative hyperglycemia management: an update. Anesthesiology. 2017;126(3):547–560. doi:10.1097/ALN.0000000000001515
- 19.↑
Frisch A, Chandra P, Smiley D, et al. Prevalence and clinical outcome of hyperglycemia in the perioperative period in noncardiac surgery. Diabetes Care. 2010;33(8):1783–1788. doi:10.2337/dc10-0304
- 20.↑
Gandhi GY, Nuttall GA, Abel MD, et al. Intraoperative hyperglycemia and perioperative outcomes in cardiac surgery patients. Mayo Clin Proc. 2005;80(7):862–866. doi:10.4065/80.7.862
- 21.↑
Fish LH, Weaver TW, Moore AL, Steel LG. Value of postoperative blood glucose in predicting complications and length of stay after coronary artery bypass grafting. Am J Cardiol. 2003;92(1):74–76. doi:10.1016/S0002-9149(03)00472-7
- 22.↑
Schwenk ES, Mraovic B, Maxwell RP, Kim GS, Ehrenfeld JM, Epstein RH. Root causes of intraoperative hypoglycemia: a case series. J Clin Anesth. 2012;24(8):625–630. doi:10.1016/j.jclinane.2012.04.009
- 23.↑
Kronen P, Moon-Massat PF, Ludders JW, et al. Comparison of two insulin protocols for diabetic dogs undergoing cataract surgery. Vet Anaesth Analg. 2001;28(3):146–155. doi:10.1046/j.1467-2987.2001.00053.x
- 24.↑
Nelson R. Canine diabetes mellitus. In: Feldman EC, Nelson RW, Reusch C, Scott-Moncrieff JC, eds. Canine and Feline Endocrinology. 4th ed. Elsevier; 2014:213–257.
- 25.↑
Reusch C. Feline diabetes mellitus. Feldman EC, Nelson RW, Reusch C, Scott-Moncrieff JC, eds. Canine and Feline Endocrinology. 4th ed. Elsevier; 2014:258–314.
- 26.↑
Evans JB, Lindbury JA, Jeffery N, Washburn SE, Patterson CA. Hip sensor location for flash interstitial glucose monitoring system provides adequate clinical accuracy in non-diabetic hypoglycemic dogs. Am J Vet Res. 2016;77(8):805–817.
- 27.↑
Koenig A, Hoenig ME, Jimenez DA. Effect of sensor location in dogs on performance of an interstitial glucose monitor. Am J Vet Res. 2016;7(8):805–817. doi:10.2460/ajvr.77.8.805
- 28.↑
Mensh BD, Wisniewski NA, Neil BM, Burnett DR. Susceptibility of interstitial continuous glucose monitor performance to sleeping position. J Diabetes Sci Technol. 2013;7(4):863–870. doi:10.1177/193229681300700408
- 29.↑
Bilicki KL, Schermerhorn T, Klocke EE, McMurphy RM, Roush JK. Evaluation of a real-time, continuous monitor of glucose concentration in healthy dogs during anesthesia. Am J Vet Res. 2010;71(1):11–16. doi:10.2460/ajvr.71.1.11
- 30.↑
Price CE, Fanelli JE, Aloi JA, et al. Feasibility of intraoperative continuous glucose monitoring: an observational study in general surgery patients. J Clin Anesth. 2023;87:111090. doi:10.1016/j.jclinane.2023.111090
- 31.↑
Del Baldo F, Diana A, Canton C, Linta N, Chiocchetti R, Fracassi F. The influence of skin thickness on flash glucose monitoring system accuracy in dogs with diabetes mellitus. Animals (Basel). 2021;11(2):408. doi:10.3390/ani11020408
- 32.↑
Sonowane M, Savory J, Cross RE, Heintges MG, Chester B. Kinetic measurement of glucose with a centrifugal analyzer; hexokinase and glucose oxidase procedures compared. Clin Chem. 1976;22(7):1100–1101. doi:10.1093/clinchem/22.7.1100
- 33.↑
Kumar K, Vijayakumari K, Lakshmi K. A prospective comparative study of glucose estimation by hexokinase and glucose oxidase-peroxidase methods. Int J Clin Biochem Res. 2019;6(3):356–362. doi:10.18231/j.ijcbr.2019.078
- 34.↑
Basu A, Dube S, Slama M, et al. Time lag of glucose from intravascular to interstitial compartment in humans. Diabetes. 2013;62(12):4083–4087. doi:10.2337/db13-1132
- 35.↑
Boyne MS, Silver DM, Kaplan J, Saudek CD. Timing of changes in interstitial and venous blood glucose measured with a continuous subcutaneous glucose sensor. Diabetes. 2003;52(11):2790–2794. doi:10.2337/diabetes.52.11.2790
- 36.↑
Wiedmeyer CE, Johnson PJ, Cohn LA, et al. Evaluation of a continuous glucose monitoring system for use in veterinary medicine. Diabetes Technol Ther. 2005;7(6):885–895. doi:10.1089/dia.2005.7.885
- 37.↑
Kadoi Y. Anesthetic considerations in diabetic patients. Part II: intraoperative and postoperative management of patients with diabetes mellitus. J Anesth. 2010;24(5):748–756. doi:10.1007/s00540-010-0988-0
- 38.↑
Lattermann R, Schricker T, Wachter U, Georgieff M, Goertz A. Understanding the mechanisms by which isoflurane modifies the hyperglycemic response to surgery. Anesth Analg. 2001;93(1):121–127. doi:10.1097/00000539-200107000-00026
- 39.↑
Robertshaw HJ, Hall GM. Diabetes mellitus: anaesthetic management. Anaesthesia. 2006;61(12):1187–1190. doi:10.1111/j.1365-2044.2006.04834.x
- 40.↑
Lerche P. Anticholinergics. In: Grimm K, Lamont L, Tranquilli W, Greene S, Robertson S, eds. Veterinary Anesthesia and Analgesia. 5th ed. Wiley-Blackwell; 2015:185.
- 41.↑
Howard LA, Lidbury JA, Jeffery N, Washburn SE, Patterson CA. Evaluation of a flash glucose monitoring system in nondiabetic dogs with rapidly changing blood glucose concentrations. J Vet Intern Med. 2021;35(6):2628–2635. doi:10.1111/jvim.16273
- 42.↑
Hoss U, Budiman ES. Factory-calibrated continuous glucose sensors: the science behind the technology. Diabetes Technol Ther. 2017;19(suppl 2):S44–S50. doi:10.1089/dia.2017.0025
- 43.↑
Bailey T, Bode BW, Christiansen MP, Klaff LJ, Alva S. The performance and usability of a factory-calibrated flash glucose monitoring system. Diabetes Technol Ther. 2015;17(11):787–794. doi:10.1089/dia.2014.0378
- 44.↑
Willson CJ, Chandra SA, Kimbrough CL, Jordan HL. Effect of estrous cycle phase on clinical pathology values in beagle dogs. Vet Clin Pathol. 2012;41(1):71–76. doi:10.1111/j.1939-165X.2011.00392.x
