Accuracy of a continuous glucose monitoring system in dogs and cats with diabetic ketoacidosis

Objective – (1) To determine the ability of a continuous interstitial glucose monitoring system (CGMS) to accurately estimate blood glucose (BG) in dogs and cats with diabetic ketoacidosis. (2) To determine the effect of perfusion, hydration, body condition score, severity of ketosis, and frequency of calibration on the accuracy of the CGMS. Design – Prospective study. Setting – University Teaching Hospital. Animals – Thirteen dogs and 11 cats diagnosed with diabetic ketoacidosis were enrolled in the study within 24 hours of presentation. Interventions – Once BG dropped below 22.2 mmol/L (400 mg/dL), a sterile flexible glucose sensor was placed aseptically in the interstitial space and attached to the continuous glucose monitoring device for estimation of the interstitial glucose every 5 minutes. Measurements and Main Results – BG measurements were taken with a portable BG meter every 2–4 hours at the discretion of the primary clinician and compared with CGMS glucose measurements. The CGMS estimates of BG and BG measured on the glucometer were strongly associated regardless of calibration frequency (calibration every 8 h: r=0.86, P<0.001; calibration every 12 h: r=0.85, P<0.001). Evaluation of this data using both the Clarke and Consensus error grids showed that 96.7% and 99% of the CGMS readings, respectively, were deemed clinically acceptable (Zones A and B errors). Interpatient variability in the accuracy of the CGMS glucose measurements was found but was not associated with body condition, perfusion, or degree of ketosis. A weak association between hydration status of the patient as assessed with the visual analog scale and absolute percent error (Spearman's rank correlation, ρ=?0.079, 95% CI=?0.15 to ?0.01, P=0.03) was found, with the device being more accurate in the more hydrated patients. Conclusions – The CGMS provides clinically accurate estimates of BG in patients with diabetic ketoacidosis.     

Endogenous serum insulin concentration in dogs with diabetic ketoacidosis

Objective: To determine endogenous serum insulin concentration in dogs with diabetic ketoacidosis (DKA), and to compare it to endogenous serum insulin concentration in diabetic dogs with ketonuria but no acidosis (KDM), diabetic dogs with uncomplicated diabetes mellitus (DM) that did not have ketonuria or acidosis, and dogs with non‐pancreatic disease (NP). Design: Prospective study. Setting: Veterinary Hospital of the University of Pennsylvania. Animals: Forty‐four client‐owned dogs; 20 dogs with newly diagnosed diabetes mellitus (7 dogs with DKA, 6 dogs with KDM, and 7 dogs with DM) and 24 dogs with non‐pancreatic disease. Interventions: Blood and urine samples were obtained at the time of admission to the hospital. Measurements and main results: Signalment, clinical signs, physical examination findings, and concurrent disease were recorded for all dogs. Blood glucose concentration, venous blood pH, venous blood HCO3^? concentration, urinalysis, and endogenous serum insulin concentration were determined in all dogs. Dogs with DKA have significantly decreased endogenous serum insulin concentrations compared to dogs with DM (P = 0.03) and dogs with non‐pancreatic disease (P = 0.0002), but not compared to dogs with KDM (P = 0.2). Five of 7 dogs with DKA had detectable endogenous serum insulin concentrations, and 2 of these dogs had endogenous serum insulin concentration within the normal range. Conclusions: Diabetic dogs with ketoacidosis have significantly decreased endogenous serum insulin concentration compared to dogs with uncomplicated diabetes mellitus. However, most dogs with DKA have detectable endogenous serum insulin concentrations, and some dogs with DKA have endogenous serum insulin concentrations within the normal range.     

Clinical assessment of blood glucose homeostasis in horses: comparison of a continuous glucose monitoring system with a combined intravenous glucose and insulin test protocol

Background: The combined glucose‐insulin test (CGIT) is helpful for evaluating insulin sensitivity. A continuous glucose monitoring system (CGMS) reports changes in interstitial glucose concentrations as they occur in the blood. Use of the CGMS minimizes animal contact and may be useful when performing a CGIT. Hypothesis: Results obtained using a CGMS are useful for the evaluation of glucose responses during the evaluation of insulin sensitivity in equids. Animals: Seven mature, obese ponies. Methods: Ponies were equipped with CGMS for determination of interstitial glucose concentrations. Glucose (150 mg/kg, IV) and insulin (0.1 U/kg, IV) were administered and blood glucose concentrations determined at (minutes after time zero) 1, 5, 15, 25, 35, 45, 60, 75, 90, 105, and 120 with a hand‐held glucometer. Blood chemistry results were compared with simultaneously obtained results using CGMS. Results: Concordance coefficients determined for comparison of blood glucose concentrations determined by a hand‐held glucometer and those determined by CGMS after the zero time point were 0.623, 0.764, 0.834, 0.854, and 0.818 (for delays of 0, 5, 10, 15, and 20 minutes, respectively). Conclusions and Clinical Importance: Interstitial glucose concentrations obtained by the CGMS compared favorably to blood glucose concentrations. CGMS may be useful for assessment of glucose dynamics in the CGIT.     

Evaluation of a continuous glucose monitoring system in diabetic dogs

The generation of a blood glucose curve is important for assessing the response to insulin therapy in diabetic dogs. Disadvantages of this technique include patient discomfort and the potential for missing transient hypo- or hyperglycaemic episodes. The aim of the current study was to evaluate a continuous glucose monitoring system (CGMS) for use in diabetic dogs. Interstitial fluid glucose concentrations were recorded in 10 diabetic dogs, every five minutes for up to 48 hours, using a subcutaneous sensor attached to the CGMS device. Blood glucose concentrations were measured simultaneously using a glucometer. The correlation between interstitial fluid and blood glucose values was 0.81 (P < 0.01). The largest discrepancies between the two sets of data were seen during the one- to three-hour period following feeding, suggesting that postprandial hyperglycaemia might not be reflected in the interstitial fluid. The authors conclude that the CGMS is a potentially valuable tool in the management of canine diabetic patients.     

Blood-to-saliva glucose time lag in sedated healthy dogs

The management of diabetes mellitus mandates measurement of blood glucose. Saliva offers an alternative to blood sampling, but measurement of the salivary glucose concentration is difficult, and the blood-to-saliva glucose time lag is uncertain. We aimed to determine the serum–saliva glucose time lag in the saliva of healthy dogs. The combined duct of the mandibular and sublingual salivary glands of 6 dogs was cannulated to collect saliva and prevent glucose degradation by oral bacteria. Following a 0.25 g/kg IV bolus of dextrose, paired serum–saliva samples were collected at baseline and in twelve 5-min blocks over 60 min. Serum and salivary glucose levels were analyzed with a linear mixed model for repeated measures with a compound symmetry error structure. Mean (±SD) saliva production was 10.3 ± 2.9 µL/kg/min, and the area under the curve (AUC_glucose)_saliva/serum ratio was 0.006, which highlights the magnitude of the large difference in glucose concentration between the 2 compartments. The serum–saliva glucose time lag was 30–40 min.     

Accuracy of serum beta-hydroxybutyrate measurements for the diagnosis of diabetic ketoacidosis in 116 dogs

The aim of this study was to evaluate the accuracy of serum beta-hydroxybutyrate (beta-OHB) measurements for the diagnosis of diabetic ketoacidosis (DKA) in dogs. One hundred sixteen diabetic dogs were prospectively enrolled in the study: 18 insulin-treated (IT) diabetic dogs that had a positive urine ketone test and 88 untreated, newly diagnosed diabetic dogs. Venous blood gas tensions and pH, serum glucose and urea nitrogen (SUN), and electrolyte (Na+, Cl-, and K+) and urine acetoacetate (AA) concentrations were measured concurrently with serum beta-OHB concentrations. On the basis of laboratory findings, the patients were assigned to I of 3 groups: diabetic ketoacidosis (n = 43); diabetic ketosis (DK, n = 41); and nonketotic diabetes (NDK, n = 31). Serum beta-OHB concentrations differed significantly (P < .001) among the study groups. Although marked differences in beta-OHB concentrations were found, a considerable overlap exists between the distributions of dogs with DK and those with DKA. The overall accuracy of beta-OHB determination as a diagnostic test for DKA, determined by the area under the receiver operating characteristic (ROC) curve, was 0.92. In the 1.9- to 4.8-mmol/L range, serum beta-OHB determination sensitivity varied from 100 to 35.7%, whereas specificity varied from 39 to 100%. The cutoff value of 3.8 mmol/L showed the best equilibrium between specificity (95%), sensitivity (72%), and likelihood ratio (14.8). We concluded that the quantitative measurement of serum beta-OHB may be a potential tool for diagnosing and monitoring ketosis and ketoacidosis in diabetic dogs.     

Comparison of two insulin protocols for diabetic dogs undergoing cataract surgery

Objective To evaluate the effectiveness of two insulin doses to maintain an acceptable range of blood glucose concentrations (70–200 mg dL^?1) in the peri‐operative period in diabetic dogs. Animals Twenty‐four diabetic dogs with a median weight of 20.6 kg and a median age of 8 years old. Methods The dogs were randomly assigned to receive either 25 or 100% of their normal insulin dose subcutaneously on the morning of surgery. The anesthetic and feeding protocols were standardized. On the day before surgery, venous blood was collected for measurement of β‐hydroxybutyrate, cholesterol, glucose, glycosylated hemoglobin, hematocrit, total plasma protein and urea nitrogen. On the day of surgery, blood glucose concentrations were measured prior to anesthesia, prior to the start of surgery, 1 and 2 hours after beginning of surgery, 1 hour after extubation, at 16 : 00 hours and at 20 : 00 hours. β‐hydroxybutyrate concentrations were measured at 20 : 00 hours that day. At 08 : 00 hours the following day, β‐hydroxybutyrate and glucose concentrations were measured. The significance of differences between groups was tested with Wilcoxon's two‐tailed rank‐sum test, Chi‐square test and Fisher's exact test. Results There were no differences in insulin treatments, clinical signs, concurrent diseases and most clinicopathological parameters between the two groups of dogs at entry to the study. The 25% dose group had blood glucose values of 296 (102–601) mg dL^?1 at 16 : 00 hours and 429 (97–595) mg dL^?1 at 20 : 00 hours on the day of surgery. The 100% insulin dose group had lower corresponding values of 130 (55–375) mg dL^?1 (p = 0.04) and 185 (51–440) mg dL^?1 (p = 0.004). No other differences (p < 0.05) were detected between the two groups. Conclusions The administration of a full dose of insulin is only marginally advantageous for reducing glucose to normal (70–120 mg dL^?1) after anesthesia but neither dose consistently induced glycemic values in an acceptable range (70–200 mg dL^?1) or normoketonemia. Clinical relevance Blood glucose should be measured immediately before anesthesia and periodically throughout the peri‐operative period in all diabetic dogs because presurgical subcutaneous administration of 25 or 100% of the normal insulin dose resulted in unpredictable blood glucose concentrations.     

Relationships between glucose, sodium and effective osmolality in diabetic dogs and cats

Objective: To examine the relative contributions of sodium and glucose to serum effective osmolality and the presence of abnormalities of sodium and osmolality in diabetic dogs and cats. Design: Retrospective study. Setting: A university‐based referral hospital. Animals: Diabetic dogs (n=14) and cats (n=13) consecutively admitted to the hospital over a 6‐month period. Interventions: None Measurements: Serum biochemistry assessments. Main results: The mean glucose concentration was higher in diabetic dogs than in diabetic cats. Total osmolality (Osm_T), effective osmolality (Osm_E), and the concentrations of sodium, potassium, blood urea notrogen, bicarbonate, and creatinine did not differ between species. Sodium abnormalities and hyperosmolality affected 44% and 81%, respectively, of the study group. However, marked hyperosmolality (Osm_E>330 mOsm/L) was found in only 33% of the study group. Serum sodium correlated closely with Osm_E in dogs and cats but serum glucose did not correlate with the Osm_E in either species. Subsets of dogs (n=10) and cats (n=7) with diabetic ketosis (DK) were examined separately. DK dogs had significantly lower sodium concentrations than DK cats and the proportion of DK dogs with hyponatremia was nearly 3 times greater than DK cats. Severe hyperosmolality (Osm_E>330 mOsm/L) was more common in DK cats than DK dogs. Conclusions: In diabetic dogs and cats, sodium, not glucose, was correlated with serum Osm_E and marked elevation in pretreatment Osm_E is uncommon. Compensatory reduction in serum sodium may be 1 mechanism for blunting changes in Osm_E in the presence of marked hyperglycemia.     

Serial haemostatic monitoring of dogs with multicentric lymphoma

Lymphoma is the most common haematopoietic malignancy in dogs and it has been associated with hypercoagulability and subsequent thromboembolism. The objectives of this study were to serially characterize the haemostatic status of dogs with multicentric lymphoma. Thromboelastography, thrombin–antithrombin complex concentration and routine haematology and coagulation panels were measured. Twenty‐seven dogs were included in the study and 15 completed the study in remission. At presentation, 81% (22/27) of dogs with multicentric lymphoma had altered haemostatic profiles consistent with hypercoagulability. Laboratory evidence of hypercoagulability did not resolve during treatment or for up to 1 month following attainment of clinical remission. Accelerated rate of clot formation at the time of chemotherapeutic protocol completion was associated with decreased survival time. We concluded that dogs with multicentric lymphoma were frequently hypercoagulable from presentation through 4 weeks after the completion of chemotherapy. Increased angle and shortened K in dogs that have successfully completed their chemotherapeutic protocol may be associated with shorter survival times.     

Outcome of dogs with diabetic ketoacidosis: 127 dogs (1993-2003)

The aim of this study was to retrospectively describe the outcome of 127 dogs with naturally occurring diabetic ketoacidosis (DKA) and to examine the association between outcome of canine DKA and clinical and clinicopathologic findings. Eighty-two (65%) dogs were diagnosed with DKA at the time of initial diagnosis of diabetes mellitus (DM). Eighty-seven dogs (69%) had one or more concurrent disorders diagnosed at the time of hospitalization. Commonly identified concurrent conditions included acute pancreatitis (52, 41%), urinary tract infection (21, 20%), and hyperadrenocorticism (19, 15%). Dogs with coexisting hyperadrenocorticism were less likely to be discharged from the hospital (P = .029). Of 121 treated dogs, 89 dogs (70%) survived to be discharged from the hospital, with a median hospitalization of 6 days. Nonsurvivors had lower ionized calcium concentration (P < .001), lower hematocrit (P = .036), lower venous pH (P = .0058), and larger base deficit (P = .0066) than did survivors. Time from admission to initiation of subcutaneous insulin therapy was correlated with lower serum potassium concentration (P = .0056), lower serum phosphorus concentration (P = .0043), abnormally high white blood cell count (P = .0060), large base deficit (P = .0015), and low venous pH (P < .001). Multivariate analysis showed that base deficit was associated with outcome (P = .021). For each unit increase in the base deficit, there was a 9%) greater likelihood of discharge from the hospital. In conclusion, the majority of dogs with DKA were not previously diagnosed with DM. Concurrent conditions and electrolyte abnormalities are common in DKA and are associated with length of hospitalization. Survival was correlated to degree of anemia, hypocalcemia, and acidosis.     

Evaluation of four portable blood glucose meters in diabetic and non-diabetic dogs and cats

Background: Monitoring of an animal's blood glucose concentration is critical for diagnostic and therapeutic decisions. Over the past few decades, portable blood glucose meters (PBGMs) have been used to monitor blood glucose concentrations in animals. Recently, new and improved PBGMs have been made available on the market.

Objective: The purpose of this study was to evaluate four PBGMs for use in dogs and cats.

Animals and methods: A total of 155 venous blood samples of dogs and 85 venous blood samples of cats were tested using four PBGMs. Control solutions from manufacturers were used to determine the precision of each meter. The coefficient of variation was calculated to determine precision during a set of replicates. Pearson's correlation analysis, Passing–Bablok regression, and Bland–Altman analysis were used to determine the accuracy of four PBGMs against the hexokinase reference method. Error grid analysis was used to evaluate clinical relevance.

Results: All PBGMs, except CERA-PET®, were clinically acceptable for monitoring blood glucose concentrations; AlphaTrak® and VetMate® appeared to be the most accurate ones, demonstrating that to use PBGMs for glucose monitoring, it is important to understand the strengths or limitations of each meter. The difference in results between the PBGMs and the reference method increased at high glucose concentration ranges, which were also affected by the hematocrit.

Conclusions: Although readings of the PBGMs and the reference method varied across glycemic ranges (low, normal, and high glucose concentrations), most PBGMs were clinically acceptable for monitoring blood glucose concentrations in dogs and cats.     

Evaluation of the tongue for oscillometric measurement of arterial pressure in anesthetized Beagle dogs

ObjectiveTo evaluate the agreement between oscillometric blood pressure (OBP) measured from the tongue and invasive blood pressure (IBP), and to compare OBPs measured from the tongue with OBPs measured from the pelvic limb and tail.Study designProspective experimental study.AnimalsA total of eight adult Beagle dogs weighing 11.1 ± 1.2 kg.MethodsAnimals were premedicated with intravenous (IV) acepromazine (0.005 mg kg^–1). Anesthesia was induced with alfaxalone (3 mg kg^–1) IV and maintained with isoflurane. The dorsal pedal artery was catheterized for IBP measurements. Systolic (SAP), diastolic (DAP) and mean (MAP) arterial pressure were simultaneously measured from the tongue, pelvic limb and tail. Based on invasive SAP, hypertension (>140 mmHg), normotension (90–140 mmHg) and hypotension (<90 mmHg) were induced by controlling end-tidal isoflurane concentrations and/or dobutamine/dopamine administration. Agreement between paired IBP and OBP measurements was analyzed with reference standards for noninvasive blood pressure devices used in small animals and humans.ResultsRegardless of cuff placement, the mean bias ± standard deviation between IBP and OBP met veterinary (≤10 ± 15 mmHg) and human (<5 ± 8 mmHg) standards for MAP and DAP. SAP measurements provided by the OBP device showed unacceptable agreement with IBP, and the bias between methods increased at higher blood pressures, regardless of cuff site. During hypotension, tongue OBP showed the largest percentage of absolute difference <10 mmHg in relation to IBP for SAP (90%), MAP (97%), and DAP (93%), compared with pelvic limb (60%, 97% and 82%, respectively) and tail OBP (54%, 92% and 77%, respectively).Conclusions and clinical relevanceThe tongue is a clinically useful site for measuring OBP in anesthetized Beagle dogs, providing reliable estimates of MAP and DAP. The tongue could replace other cuff placement sites and may be a relatively suitable site for assessing hypotension.     

Association of hyponatremia and hyperglycemia with outcome in dogs with congestive heart failure

Objective: To evaluate plasma sodium and glucose concentrations in dogs with congestive heart failure (CHF) prior to treatment and evaluate the differences between survivors and non‐survivors. Design: Retrospective study. Animals: Fifty‐nine dogs with CHF prior to receiving cardiac medication. Interventions: None. Measurements and main results: The mean plasma sodium concentration in dogs with CHF was below the reference range (144–156 mmol/L) and significantly lower (P=0.009) in non‐survivors (141±6 mmol/L) compared with survivors (147±4 mmol/L). The mean plasma glucose concentration was above the reference range (76–117 mg/dL) and significantly higher (P=0.004) in non‐survivors (128±52 mg/dL) compared with survivors (100±13 mg/dL). Forty‐four percent of non‐survivors had concurrent low plasma sodium and high plasma glucose concentrations, whereas no survivors had both abnormalities (P<0.0001). Conclusions: Lower plasma sodium and higher plasma glucose are associated with a worse outcome in dogs with CHF.