Abstract
Objective
Comparison of a veterinary glucometer (AlphaTRAK 2 [AT-2]) and 2 human glucometers (FreeStyle Libre 2 [FS-] and FreeStyle Libre 14 [FS-14]) to an automated, wet-chemistry analyzer (reference analyzer).
Methods
This was a prospective observational study at Tufts Cummings School of Veterinary Medicine between January 2021 and September 2022 and included 187 client and staff-owned dogs. Following venipuncture, 3 glucometers (FS-14, FS-2, and AT-2) were used to measure blood glucose, and the reference analyzer was used to measure serum glucose.
Results
Compared to the reference analyzer, the FS-2 and FS-14 glucometers had a negative bias (mean difference estimates: FS-2, −25.01 mg/dL [95% CI, −60.4 to 10.3]; FS-14, −23.6 mg/dL [95% CI, −60.7 to 13.5]), while the AT-2 glucometer had a positive bias (mean difference estimates: 15.4 mg/dL [95% CI, −41.1 to 72.2]). All glucometers showed significant constant and proportional biases based on Passing-Bablok regression with constant biases of −12.3, −10.05, and −14.25 for the FS-2, FS-14, and AT-2, respectively. Most results were within zone A (FS-2, 50.3%; FS-14, 54.5%; AT-2, 70.1%) and B (FS-2, 49.7%; FS-14, 45.5%; AT-2, 26.2%) of the Clarke error grid. The AT-2 produced values within zone C (1.6%) and zone D (2.1%).
Conclusions
All glucometers correlated with the reference analyzer and were clinically useful. As the AT-2 glucometer produced values in the Clarke error grid zones C and D, serum glucose should be measured when results are unexpected or influence the treatment regimen.
Clinical Relevance
Despite potential biases, owners can use FS-2 and FS-14 glucometers to verify sensor readings in dogs using continuous glucose monitoring devices.
Accurate blood glucose determination is crucial for the diagnosis and management of diabetes mellitus, a common endocrinopathy with an annual prevalence of 0.26% in dogs more than 3 years old.1 Glycemic control is vital for adequate disease management, and the gold standard method to measure blood glucose is through an automated, wet-chemistry analyzer. Most chemistry analyzers use a glucose hexokinase reaction with photometric detection to measure glucose concentration in serum or plasma. Portable blood glucose meters (PBGMs) are a fast, convenient way to measure blood glucose.2 Most PBGMs detect glucose electrochemically through the measurement of an electrical current generated by enzymatic activity (eg, glucose dehydrogenase) on a test strip.3 Applying an algorithm, PBGMs use whole blood to generate a value for plasma glucose based on the expected ratio of glucose within RBCs to plasma.4 The AlphaTrak2 (AT-2; Abbott Laboratories Ltd) is one of the most well-known and studied veterinary PBGMs.5–7 A recent study7 of mostly normoglycemic dogs found that 95% of the samples measured with the AT-2 were from 23.4 to 63 mg/dL higher than the values measured by a chemistry analyzer.
Continuous glucose monitoring systems are a popular management tool for monitoring diabetic dogs due to the avoidance of venipuncture and the abundance of data on glucose fluctuations.8 FreeStyle Libre systems (FSs; Abbott Laboratories Ltd) are a flash interstitial glucose monitoring systems commonly used in veterinary medicine. Versions of the system include the FreeStyle Libre 14-day (FS-14), the FreeStyle Libre 2 (FS-2), and the newly introduced FreeStyle Libre 3. Interstitial glucose concentrations were found to correlate well with whole blood, serum, and plasma glucose in humans and veterinary patients.9–12 However, continuous glucose monitoring systems may not accurately reflect the glucose profile in animals with marked hyper- or hypoglycemia.8 It is, therefore, recommended to confirm blood glucose concentrations with a glucometer whenever interstitial glucose readings are low or high, or the results are unexpected.8 As FreeStyle Libre readers have a built-in blood glucose meter, owners may be able to use their FreeStyle Libre reader as a regular glucometer to verify sensor readings. Verification of sensor readings would result in decreased owner anxiety and expenses associated with purchasing additional veterinary glucometers. If they provide clinically reliable readings, FS glucometers have the potential to impact the management of diabetic patients.
The purpose of this study was to compare the performance of a veterinary-calibrated PBGM (AT-2) and 2 human PBGMs (FS-2 and FS-14) to an automated, wet-chemistry analyzer (reference analyzer).
Methods
Animals
Client- and staff-owned dogs of various ages, breeds, and health statuses that were presented to the Tufts Cummings School of Veterinary Medicine between January 2021 and September 2022 were prospectively enrolled in the study. Leftover blood samples were collected from dogs that required a serum chemistry panel as a part of their diagnostic or monitoring protocol. Permission for the use of discarded blood was obtained from the Tufts Cummings School of Veterinary Medicine Clinical Sciences Review Committee. An additional number of dogs with either hyperglycemia or hypoglycemia were primarily recruited for the study to obtain representative samples outside normal blood glucose concentrations. The owners of recruited dogs gave written consent. This study was approved by the Clinical Sciences Review Committee and the IACUC under protocol number G2022-76.
Dogs were recruited regardless of health status. Dogs were excluded from the study if they did not have serum chemistry glucose measurement performed, if they had been previously enrolled in the study, or if they had significant circumstances precluding venipuncture (eg, severe anemia, PCV < 15%; thrombocytopenia, platelet count < 30,000/µL; coagulopathy; critical patient status; severe anxiety; or aggression).
Sample collection
An Internal Medicine technician performed venipuncture using either the jugular, saphenous, or cephalic vein with a 20- or 22-G needle and a syringe. Approximately 2.0 mL of blood was placed into a red-top serum separator tube for serum glucose measurement using the reference analyzer. The remaining blood was used for measurement of PCV, total solids, and immediate quantification of blood glucose by 2 FS-14, 2 FS-2, and 2 AT-2 glucometers (on canine setting), yielding a total of 6 measurements.
Glucose measurement
Duplicate analysis was performed by having 2 identical glucometers of the following: FS-2 (glucometers A and B), FS-14 (glucometers C and D), and AT-2 (glucometers E and F). The 6 glucometers were set up before phlebotomy by inserting glucose test strips into the meter test ports.
The FS-2 and FS-14 glucometers use glucose dehydrogenase test strips and an amperometric method to measure blood glucose concentrations (range, 20 to 500 mg/dL) in 0.3 µL of whole blood with a Hct range of 15% to 65%. The AT-2 glucometers use glucose oxidase test strips and coulometry to measure blood glucose concentrations (range, 20 to 750 mg/dL) in 0.3 µL of whole blood with a Hct range of 15% to 65%. Each PBGM was used with single-use test strips that were designed for use in that meter. To ensure accuracy, each PBGM was tested with a control solution whenever a new box of test strips was opened.
Samples in the serum separator tube were centrifuged at 3,000 X g for 10 minutes, and the serum was separated within 40 minutes from collection. A laboratory-based, benchtop chemistry analyzer, the Cobas c501 (Roche), was used to measure serum glucose (range, 2 to 750 mg/dL), utilizing glucose hexokinase and photometric methods. The analyzer is calibrated daily and routinely maintained per manufacturer guidelines. Serum glucose levels were defined as hypoglycemic (< 67 mg/dL), hyperglycemic (> 135 mg/dL), and normoglycemic (67 to 135 mg/dL), based on in-house reference intervals.
Measurement of PCV, total solids, and other parameters
A spun PCV was manually assessed using a microhematocrit centrifuge Clay Adams Compact II Centrifuge (Becton Dickinson). Samples sent to the laboratory for a CBC had PCVs measured using a Sorvall Legend Micro17 Microcentrifuge.
Statistical analysis
Data were analyzed using the statistical software R 4.1.0 (2021-05-18; The R Foundation). As comparisons are only possible between numerical values, glucose values outside of the glucometers’ range were recorded as “0,” and missing data points were recorded as “999.” Descriptive statistics were generated for each device. Glucose measurement data were tested for normality using the Shapiro-Wilk test. Normally distributed data were expressed as mean ± SD. Nonnormally distributed data were expressed as median and range.
The test methods were the glucometers, and the reference method was the laboratory chemistry analyzer (reference analyzer). Intraclass correlation was used to test the reliability between the readings from each pair of glucometers. Mean absolute relative difference analysis was used to describe the accuracy of the glucometers by calculating the relative difference between the FS-2, FS-14, and AT-2 glucometers in comparison to the gold standard (reference analyzer). Bland-Altman plots were performed to describe the agreement between the glucometers and the reference analyzer. Passing-Bablok linear regression analysis was further used to estimate bias between the glucometers and the reference analyzer, which provided estimates for constant and proportional bias. If the 95% CI for the y-intercept did not include the value of 0, this was considered evidence of significant constant bias. If the 95% CI for the slope did not include the value of 1, this was considered evidence of significant proportional bias. The Pearson correlation coefficient (r) was also provided in the Passing-Bablok regression outputs. Results were presented with a scatter diagram and regression line and regression equation. The cumulative sum linearity test was performed to investigate significant deviation from linearity between the 2 measurement methods. The level of P < .05 was considered statistically significant.
Despite deviations from recommended guidelines, the glucometers’ performances were evaluated according to the 2013 International Standards Organization (ISO) standard 15197-2013.13 This standard requires 95% of individual glucose readings from the glucometer to be within 15 mg/dL of the reference measurement for plasma glucose concentrations below 100 mg/dL and within 15% of the reference measurement for concentrations above 100 mg/dL. Deviations from the ISO guidelines included utilizing serum glucose instead of plasma for the determination of glucose with the reference analyzer, using venous blood instead of capillary blood for the glucometers, and having fewer hyper and hypoglycemic samples than recommended.
The clinical relevance of measurement deviations was analyzed with the Clarke error grid (CEG) analysis, an error grid developed by clinicians assuming that blood glucose values < 70 or > 240 mg/dL require clinical action.14 The CEG was divided into 5 zones. Zone A represented measurements that had no effect on clinical action (glucometer within 20% of analyzer); zone B represented altered clinical action with little to no effect on clinical outcome (> 20% difference, no incorrect treatment); zone C represented altered clinical action, likely to affect the clinical outcome (hyperglycemia or hypoglycemia leading to inappropriate treatment); zone D represented altered clinical action that could have significant medical risk (undetected hypoglycemia or hyperglycemia needing treatment); and zone E representing altered clinical action that could have dangerous consequences (hypoglycemia mistaken for hyperglycemia and vice versa).14
Univariate and multivariable analyses were performed to investigate the potential impact of PCV, hemolysis, lipemia, and icterus on the difference between glucose measurements produced by the glucometers (FS-1, FS-2, and AT-2) in comparison to the reference analyzer. Colinearity was checked before regression.
Results
Samples from 235 dogs were enrolled between January 2021 and September 2022. Forty-eight samples were excluded for the following reasons: samples belonged to dogs previously enrolled in the study (n = 34), serum glucose was not measured by the reference analyzer (13), and delay in sample processing (1). The remaining 187 dogs were included, with a range of blood glucose concentrations from 21 to 660 mg/dL based on a reference analyzer. Of those, 172 dogs were enrolled by using leftover blood from serum chemistry panels, and an additional 15 were enrolled for the study purposes to ensure the representation of both hyper- and hypoglycemic samples. The mean age of enrolled dogs was 8.7 ± 3.9 years. The breeds most represented included Labrador Retriever (n = 22), Poodle (10), Shih Tzu (11), Beagle (8), Yorkshire Terrier (8), Chihuahua (8), Golden Retriever (5), and Maltese (5). Ninety-two were female (81 spayed and 11 intact), and 95 were male (86 neutered and 9 intact). Thirty-seven dogs (20%) were diabetic. One hundred sixty-six (88%) dogs had PCV measurements performed. Packed cell volumes ranged from 16% to 60% with a mean ± SD value of 45.8% ± 7.9%. The PCV was decreased in 28 (17%) samples, was within reference interval in 125 (75%) samples, and was increased in 13 (8%) samples. The mean total solids was 6.2 ± 0.71 g/dL. Seventy-five samples (40.1%) had hemolysis, 73 samples (39.0%) were lipemic, and 11 (5.9%) were icteric. Fifteen samples (8%) had marked hemolysis.
Intraclass correlation between the glucometers’ duplicate analyses was greater than 90%, which indicated excellent reliability between the measurement pairs. Hence, results from each pair of FS-2, FS-14, and AT-2 glucometers were averaged for data interpretation. Blood glucose was measured in 187 dogs. Glucose concentrations in serum and blood were not normally distributed. The median FS-2, FS-14, AT-2, and reference analyzer glucose concentrations were 78 mg/dL (range, 20 to 472 mg/dL), 78 mg/dL (range, 22 to 465 mg/dL), 110 mg/dL (range, 26 to 631 mg/dL), and 100 mg/dL (range, 21 to 552 mg/dL), respectively. Based on results from the reference analyzer, 11 samples were hypoglycemic (range, 21 to 64 mg/dL), 37 samples were hyperglycemic (range, 144 to 552 mg/dL), and 139 samples were normoglycemic (67 to 135 mg/dL; Table 1).
Median glucose measurements (mg/dL) and their ranges across different glucose measuring devices.
| Glucose analyzer | Median (range) glucose measurement (mg/dL) |
|---|---|
| FS-2 | 78 (20–472) |
| FS-14 | 78 (22–465) |
| AT-2 | 110 (26–631) |
| Chemistry analyzer | 100 (21–552) |
Using mean absolute relative difference, the mean relative difference ± SD between the FS-2, FS-14, and AT-2 versus the reference analyzer measurement was 21.77 ± 11.51%, 20.27 ± 9.73%, and 16.54 ± 15.18%, respectively. None of the glucometers satisfied the 2013 ISO guidelines, but the AT-2 showed a smaller difference compared to the FS-14 and FS-2 glucometers.
Bland-Altman plot analysis showed that, on average, the FS-2 and FS-14 glucometers had a negative bias when compared to the reference analyzer with negative mean difference estimates (FS-2, −25.01 mg/dL [95% CI, −60.4 to 10.3]; FS-14, −23.6 mg/dL [95% CI, −60.7 to 13.5]; Figure 1). Meanwhile, the AT-2 glucometer had a positive bias with a mean difference estimate of 15.4 mg/dL (95% CI, −41.1 to 72.2). All 3 glucometers showed a greater absolute mean difference as the average glucose values between the method compared and the chemistry analyzer reference increased.
Bland-Altman difference plot of glucose (mg/dL) concentrations measured with AlphaTRAK 2 glucometer (AT-2; A), FreeStyle Libre 14-day glucometer (FS-14; B), and FreeStyle Libre 2 glucometer (FS-2; C) and a reference chemistry analyzer. The space between the dashed lines represents the limits within which the difference between the 2 methods is expected to fall for them to be considered in good agreement.
Citation: American Journal of Veterinary Research 86, 6; 10.2460/ajvr.24.10.0317
Passing-Bablok linear regression analysis showed constant and proportional biases for all 3 glucometers. Regression analysis of FS-2 values versus reference analyzer values yielded a y-intercept of −12.3 (95% CI, −16.27 to −8.06) and a slope of 0.9 (95% CI, 0.86 to 0.93). This indicated constant and proportional biases (Figure 2). Similarly, regression analysis of FS-14 values versus reference analyzer values yielded a y-intercept of −10.06 (95% CI, −15.17 to −6.631) and a slope of 0.89 (95% CI, 0.856 to 0.934), also indicating constant and proportional biases. For the FS-2 and the FS-14 glucometers, the proportional and constant biases were such that the values tended to be lower than the reference method. Regression analysis of AT-2 values versus reference analyzer values yielded a y-intercept of −14.25 (95% CI, −25.8 to −6.54) and a slope of 1.24 (95% CI, 1.16 to 1.36). These results also indicated the existence of constant and proportional biases. For the AT-2, the proportional bias indicated that the values tended to be higher than the reference method with increasing blood glucose concentrations. Inspection of the slope also shows that the AT-2 glucose values were lower than the reference method at low glucose concentrations.
Passing-Bablok regression analysis of glucose readings (mg/dL) between the FS-2, FS-14, and AT-2 glucometers and a reference chemistry analyzer.
Citation: American Journal of Veterinary Research 86, 6; 10.2460/ajvr.24.10.0317
The CEG analysis was performed to evaluate the clinical relevance of measurement deviations. Values from the FS-2 were within zone A (50.3% [94/187]) or zone B (49.7% [93/187]) of the error grid. Values from the FS-14 were within zones A (102/187 [54.5%] or B (85/187 [45.5%]; Figure 3). There were no values within zones C, D, or E for FS-2 and FS-14 glucometers. Most values for the AT-2 glucometer were within zone A (131/187 [70.1%]) and zone B (49/187 [26.2%]); however, there were some values within zone C (3/187 [1.6%]) and zone D (4/187 [2.1%]).
Clarke error grid analysis of blood glucose concentrations (mg/dL) obtained by the AT-2 (C), FS-14 (B), and FS-2 (A). Zone A (orange) represented measurements that had no effect on clinical action (glucometer within 20% of analyzer); zone B (green) represented altered clinical action with little to no effect on clinical outcome (> 20% difference, no incorrect treatment); zone C (blue) represented altered clinical action, likely to affect clinical outcome; zone D (purple) represented altered clinical action that could have significant medical risk; and zone E representing altered clinical action that could have dangerous consequences.
Citation: American Journal of Veterinary Research 86, 6; 10.2460/ajvr.24.10.0317
The regression results showed that the effect of PCV on the difference in glucose measurements between glucometer and reference analyzer was significant in both univariate (PFS-2 = .0006; PFS-14 < .0001; PAT-2 < .0001) and multivariable regression analysis (PFS-2 = .0009; PFS-14 = .0001; PAT-2 < .0001) for all 3 glucometers, and the difference in glucose measures decreased as PCV increased. When measured with FS-14, dogs with lipemic serum tended to have significantly greater differences with other variables adjusted, and dogs with icteric serum had significantly greater differences only on univariate analysis. None of the other variables on univariate or multivariable analysis were statistically significant (Table 2).
Results of univariate and multivariable regression analysis characterizing the effects on the differences between glucose readings (mg/dL).
| Δ(FS-2 − reference) | Δ(FS-14 − reference) | Δ(AT-2 − reference) | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Univariate | Multivariable | Univariate | Multivariable | Univariate | Multivariable | ||||||||
| Variable | Levels | Estimate | P | Estimate | P | Estimate | P | Estimate | P | Estimate | P | Estimate | P |
| Hemolysis | Yes | −2.51 | .3849 | 1.16 | .3671 | −3.08 | .3038 | 1.82 | .2720 | −6.64 | .1752 | 1.69 | .0571 |
| Lipemia | Yes | −3.55 | .2213 | 3.60 | .2261 | −5.02 | .0944 | 7.68 | .0387* | −3.53 | .4738 | −3.39 | .1288 |
| Icteric | Yes | 7.60 | .1641 | −4.21 | .3347 | 12.66 | .0247* | −5.65 | .1502 | 11.68 | .2078 | −5.87 | .8653 |
| Total solids | 0.13 | .4758 | 0.22 | .2083 | 0.11 | .5549 | 0.23 | .1908 | −0.14 | .6304 | 0.16 | .4564 | |
| PCV | −0.60 | .0006* | −0.62 | .0009* | −0.74 | .0000* | −0.75 | .0001* | −2.60 | .0000* | −2.66 | .0000* | |
Discussion
The goal of this study was to compare the performance of the FS-14, FS-2, and AT-2 glucometers in comparison with a reference, automated, wet-chemistry analyzer. The AT-2, FS-2, and FS-14 were all significantly correlated with the reference analyzer. All glucometers showed a greater difference from the reference analyzer as blood glucose concentrations increased. Passing-Bablok regression analysis showed a constant and proportional bias for all glucometers with differences increasing as blood glucose concentrations increased. While FS-2 and FS-14 glucometers consistently underestimated glucose concentrations, the AT-2 glucometer could under- or overestimate glucose concentrations. Although none of the evaluated glucometers fulfilled the 2013 ISO guidelines, all were considered clinically useful based on CEG analysis. The mean difference and proportional and constant bias of the AT-2 glucometer are consistent with findings from previous studies.5–7,15 There are no studies evaluating FS-2 and FS-14 glucometers in dogs.
All values for the FS-2 and FS-14 were within zones A and B of the CEG, in compliance with recent ISO guidelines and demonstrating that these glucometers’ biases are unlikely to change clinical outcomes. Although most values produced by the AT-2 glucometer were in zones A and B, 7 values were within zones C and D. Of those, blood glucose concentrations were overestimated in 3 samples and underestimated in 4. Recent studies7,16 have also documented that the AT-2 glucometer yielded samples in zone D. As the aforementioned studies had little to no dogs in the hypoglycemic range, and few in the hyperglycemic range, our study provides further insight into the clinical utility of the AT-2 and its effect on clinical decision making in a broader range of glycemic samples.
Despite a few values within zones C and D, relative to the FS-2 and FS-14 glucometers, the AT-2 glucometers have a smaller standardized difference between their glucose measurement and that of the reference analyzer. Therefore, it remains clinically useful despite its apparent increased risk of misdiagnosis in a clinical setting on CEG analysis.
Univariate and multivariable regression analysis revealed that PCV and glucose concentrations were negatively associated for all glucometers. Packed cell volume affects glucose measurement in PBGMs for several reasons, the main one being that the formula used to generate a plasma-equivalent glucose concentration assumes a normal PCV.4 Several studies4,6,7,17 have shown that decreases in PCV can increase glucose readings in PBGMs such as the AT-2. Meanwhile, our multivariable analysis for the FS-14 revealed that lipemia was significant on multivariable analysis and icterus on univariate analysis. Lipemia and hyperbilirubinemia are known interferences, especially at higher values, and if results are unexpected, it is recommended to confirm glucose concentration at a reference laboratory.4,18
Our study has several limitations that may have contributed to differences in glucose concentration between the glucometers and the reference analyzer. We used 2 different sample types (whole blood for glucometers and serum for the chemistry analyzer) to mimic what is performed in our hospital. Although caution should be taken when comparing results between 2 different sample types, a recent study16 found that the AT-2’s accuracy of glucose measurement was higher using whole blood rather than serum, and another study17 found that whole blood was the preferred sample for PBGM-measured glucose concentration. Although we tried to increase the number of hypo- and hyperglycemic samples by actively recruiting dogs suspected to have glucose values outside the reference interval, our number of hypo- and hyperglycemic samples remained disproportionate to the euglycemic samples. We used venous whole blood rather than capillary whole blood for the glucometers. Although some reports4 document that venous whole blood may underestimate capillary blood glucose by 15% to 20%, several studies18–21 have found that the difference between capillary and venous glucose was not significant. Finally, as the FS-2 and FS-14 glucometers are designed for humans, an incorrect formula for dogs may have contributed to the glucometers’ performance.
In conclusion, all evaluated glucometers correlated well with the reference analyzer and were clinically useful. The FS-2 and FS-14 glucometers have a negative bias, which is constant and proportional, suggesting that these glucometers can be used to confirm unexpected readings in dogs using the FS glucose monitoring system. However, hypoglycemic readings should be confirmed with a veterinary glucometer or a reference chemistry analyzer. As the AT-2 glucometer produced values in zones C and D of the CEG, serum glucose should be confirmed if the results produced by the AT-2 are unexpected or may influence the treatment regimen.
Acknowledgments
The authors express their deepest gratitude to Jean Kaseta, a dedicated veterinary technician, whose invaluable contributions made this work possible.
Disclosures
The authors have nothing to disclose. No AI-assisted technologies were used in the composition of this manuscript.
Funding
Funding was provided by the Companion Animal Health Fund.
ORCID
A. Ioannou https://orcid.org/0009-0005-3019-9473
F. Conrado https://orcid.org/0000-0001-5055-2637
References
- 1.↑
Heeley A, O’Neil D, Davison L, Church D, Corless E, Brodbelt D. Diabetes mellitus in dogs attending UK primary-care practices: frequency, risk factors and survival. Canine Med Genet. 2020;7(6):349. doi:10.1186/s40575-020-00087-7
- 2.↑
Johnson B, Fry M, Flatland B, Kirk C. Comparison of a human portable blood glucose meter, veterinary portable blood glucose meter, and automated chemistry analyzer for measurement of blood glucose concentrations in dogs. J Am Vet Med Assoc. 2009;235(22):1309–1313. doi:10.2460/javma.235.11.1309
- 3.↑
Tonyushkina K, Nichols J. Glucose meters: a review of technical challenges to obtaining accurate results. J Diabetes Sci Technol. 2009;3(4):971–980. doi:10.1177/193229680900300446
- 4.↑
Gerber K, Freeman K. ASVCP guidelines: quality assurance for portable blood glucose meter (glucometer) use in veterinary medicine. Vet Clin Pathol. 2016;45(1):10–27. doi:10.1111/vcp.12310
- 5.↑
Cohen T, Nelson R, Kass P, Christopher M, Feldman E. 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
- 6.↑
Kang MH, Kim DH, Jeong IS, Choi GC, Park HM. Evaluation of four portable blood glucose meters in diabetic and non-diabetic dogs and cats. Vet Q. 2016;36(1):2–9.
- 7.↑
Wolfenden G, James FE, Hung LHT, Thompson B. Comparative accuracy of two veterinary-calibrated point-of-care glucometers for measurement of blood glucose concentration in dogs. J Small Anim Pract. 2022;63(7):512–519.
- 8.↑
Wiedmeyer CE, DeClue AE. Continuous glucose monitoring in dogs and cats. J Vet Intern Med. 2008;22(1):2–8. doi:10.1111/j.1939-1676.2007.0001
- 9.↑
Wiedmeyer C, Johnson P, Cohn L, 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
- 10.
Garg S, Zisser H, Schwartz S, et al. Improvement in glycemic excursions with a transcutaneous, real-time continuous glucose sensor: a randomized controlled trial. Diabetes Care. 2006;29(1):44–50. doi:10.2337/diacare.29.01.06.dc05-1686
- 11.
Robinson C, Alva S, Brannan C. Evaluation of the FreeStyle precision pro blood glucose and β-ketone monitoring system. Point Care J Near-Patient Test Tech. 2017;16(1):55–58. doi:10.1097/POC.0000000000000128
- 12.↑
Corradini S, Pilosio B, Dondi F. Accuracy of a flash glucose monitoring system in diabetic dogs. J Vet Intern Med. 2016;30(4):983–988. doi:10.1111/jvim.14355
- 13.↑
Freckmann G, Schmid C, Baumstark A, Rutschmann M, Haug C, Heinemann L. Analytical performance requirements for systems for self-monitoring of blood glucose with focus on system accuracy. J Diabetes Sci Technol. 2015;9(4):885–894. doi:10.1177/1932296815580160
- 14.↑
Parkes J, Slatin S, Pardo S. A new consensus error grid to evaluate the clinical significance of inaccuracies in the measurement of blood glucose. Diabetes Care. 2000;23(8):2243–2248. doi:10.2337/diacare.23.8.1143
- 15.↑
Jahan S, Jamshidi S, Tehranisharif M, Hesameddin A. Comparison of two veterinary blood glucose meters and one human-based glucose meter for use in dogs. Iran J Vet Med. 2019;13(2):187–198.
- 16.↑
Suchowerksy N, Carlson E, Behrend E. Comparison of glucose concentrations in canine whole blood, plasma, and serum measured with a veterinary point-of-care glucometer. J Vet Diagn Invest. 2021;33(4):695–702.
- 17.↑
Lechner M, Hess R. Comparison of glucose concentrations in serum, plasma, and blood measured by a point-of-care glucometer with serum glucose concentration measured by an automated biochemical analyzer for canine and feline blood samples. Am J Vet Res. 2019;80(12):1074–1081.
- 18.↑
Eastham J, Debra M, Barnes D. Prevalence of interfering substances with point-of-care glucose testing in a community hospital. Am J Health Syst Pharm. 2009;66(2):167–170. doi:10.2146/ajhp070512
- 19.
Wess G, Reusch C. Evaluation of five portable glucose meters for use in dogs. J Am Vet Med Assoc. 2000;216(2):203–209.
- 20.
Borin S, Crivelenti LZ, Rondelli MCH. Capillary blood glucose and venous blood glucose measured with portable digital glucometer in diabetic dogs. Braz J Vet Pathol. 2012;5(2):42–26.
- 21.↑
Ismail-Hamid S, Romdane M, Romdhane S. Comparison of a human portable blood glucose meter and automated chemistry analyser for measurement of blood glucose concentrations in healthy dogs. Vet Med Sci. 2021;7(6):2185–2190.
