Veterinary-specific PBGMs are commonly used to monitor systemic blood glucose concentrations of dogs and cats in veterinary hospitals and other settings.1–4 These devices are particularly useful when frequent measurements are required because they provide a rapid assessment of systemic blood glucose concentration, are readily available and cost-effective, and require as little as 0.3 μL of blood.5–7 Studies5,7 in which results for veterinary-specific PBGMs were compared with results for hexokinase reference methods revealed that a number of these machines provide clinically acceptable results when peripheral venous or capillary blood samples are used.
The PBGM that has been used at the authors’ institution is a second-generation veterinary-specific glucometer that involves electrochemical technology and a glucose dehydrogenase reagent to measure blood glucose concentrations in samples obtained from dogs and cats. Studies5,8 conducted to compare results for this device with results for other PBGMs revealed that the PBGM used at our institution had among the lowest bias and greatest accuracy for measurement of glucose concentrations in venous blood and control solutions. The manufacturer recommends the use of blood collected from a peripheral vein or capillary blood collected from a marginal ear vein, non-weight-bearing portion of a paw pad, oral mucosa, or hyperkeratotic skin overlying the elbow joint.9,10 Measurements of blood glucose concentrations in samples obtained from a carpal pad, pinna, and peripheral vein of diabetic dogs were all considered clinically acceptable.11 However, similar comparisons are lacking for healthy dogs and cats as well as dogs and cats with nondiabetic illnesses.
In a number of nondiabetic patients at the authors’ institution, there have been clinically important disparities between the venous and capillary blood glucose concentrations measured with a PBGM. These differences have led to concerns about the effect of sample collection site on the results. One possible cause of these discrepancies may have been the prandial state of patients, which could have affected glucose concentrations in capillary and venous blood. Studies of mice12 and humans13 have revealed asynchronous shifts in blood glucose concentrations in venous and capillary blood samples obtained before and after eating. However, similar studies of dogs are lacking.
The purpose of the study reported here was to compare glucose concentrations measured with a veterinary-specific PBGM in capillary and peripheral venous blood samples obtained from healthy dogs and to determine whether those concentrations were affected by the prandial state of the dogs. We hypothesized that the site of sample collection and prandial state would not cause a significant difference in blood glucose concentrations.
Materials and Methods
Animals
Twelve healthy university-owned dogs (8 Beagles and 4 hound-type dogs; 6 males and 6 females) were used in the study. Mean age of the dogs was 1 year (range, 1 to 2 years), and mean ± SD body weight was 14.2 ± 5.2 kg (range, 9.0 to 21.7 kg). Dogs were included in the study if they were deemed healthy on the basis of results of a physical examination, CBC, and biochemical panel and if they were not receiving medications. Food was withheld from all dogs for 12 hours before the onset of the study, and rectal temperature and blood pressure (measured via Doppler ultrasonographic sphygmomanometry) were evaluated before sample collection to rule out hypothermia (rectal temperature < 37.8°C) or hypotension (systolic blood pressure < 80 mm Hg). The study was approved by the University of Tennessee Institutional Animal Care and Use Committee.
PBGM
A single commercially available PBGMa was used to evaluate all samples via coulometric methods. New vials of test stripsb and sterile lancetsc for the PBGM were used for all measurements. Before testing was conducted, the PBGM was calibrated in accordance with the manufacturer's instructions. Care was taken to ensure that the glucometer was set to the code corresponding to a canine blood sample and that all test strip vials had the same lot number.
Experimental procedures
A serial phase design was used to isolate the effect of sample collection site versus prandial state. All testing was performed on the same day by 2 investigators (JLG and RAS). For each dog, a preprandial venous blood sample (< 1 mL) was obtained from a lateral saphenous vein with a 25-gauge needle. One drop of blood was immediately placed on the end of a test strip for assessment of the blood glucose concentration by use of the PBGM; the remaining blood was placed in a heparinized capillary tube for measurement of PCV to rule out hemoconcentration (PCV > 55%) or anemia (PCV < 35%). Once it was verified that the PCV was within the reference limits, each dog was arbitrarily given a number, and an open-source randomization programd was used to assign each number to a corresponding sequence for capillary blood collection from 3 sites (6 possible sequences for blood collection; 2 dogs/sequence). A capillary blood sample was obtained from each of 3 sites (carpal pad, medial aspect of the pinna, and oral mucosa). For each capillary blood sample, a new test strip was inserted into the PBGM and a new lancet was inserted into the automatic lancing device.e Tissue at the site of sample collection was perforated with the lancet, and 1 drop of blood was allowed to well up at the site. Care was taken to avoid compression of soft tissues during blood collection. A corner of the test strip was placed in contact with the drop of blood, which was absorbed into the test strip, and the glucose concentration was measured with the PBGM and recorded. The lancing process was repeated for the remaining 2 sites.
After capillary blood samples were collected from all 3 sites, dogs were fed a meal. Two hours after the meal was consumed, physical examination, including measurement of rectal temperature and systolic blood pressure, was repeated on each dog to verify normothermia and normotension. Postprandial venous and capillary blood samples were then collected and blood glucose concentration was measured in the same order as the preprandial samples for each dog.
Data analysis
Data were summarized as mean, SD, and range. To determine the effect of collection site (oral mucosa, medial aspect of the pinna, carpal pad, and lateral saphenous vein) and prandial state on glucose concentration, a 2-way repeated-measures ANOVA with 2 within-subject factors was used to compare mean values for the sample collection site, prandial state, and collection site-by-prandial state interaction. Least squares means were calculated and compared with the Bonferroni correction method. Agreement between methods of measurement was evaluated with Bland-Altman scatterplots by graphing the difference between capillary and venous blood glucose concentrations against the mean of the 2 results. Bias, defined as the mean of all differences calculated, was identified for all collection sites to account for differences among locations.14
Concordance between results for samples obtained at each capillary blood collection site and those obtained for the peripheral venous blood sample was evaluated. Blood glucose concentration of each venous sample was classified as hypoglycemic (< 60 mg/dL), euglycemic (60 to 130 mg/dL), or hyperglycemic (> 130 mg/dL). Blood glucose concentrations for the 3 capillary blood collection sites were classified in the same manner, and the percentage of correctly classified results was calculated for each location.
Consensus error grid (also known as Parkes error grid) analysis was performed to evaluate clinical accuracy of the blood glucose concentrations in samples obtained at each site.15 Use of this method involves assigning a level of risk to each measurement error by plotting data points for venous blood glucose concentration and capillary blood glucose concentration on a grid that is divided into 5 zones (A to E). Results in zone A will not have an effect on clinical action. Results in zone B will alter clinical action, but the action will have minimal effect on clinical outcome. Results in zone C will prompt clinical action that is likely to affect clinical outcome. Results in zone D will cause errors in treatment that could pose a substantial medical risk to the patient. Results in zone E will cause severe treatment errors that could have dangerous consequences for the patient.15
All statistical assumptions regarding normality and equality of variances were met. Values of P < 0.05 were considered significant. The analysis was conducted with standard statistical analysis software.f
Results
Animals
None of the 12 dogs were excluded because of results of rectal temperature, blood pressure, or PCV measurements. All dogs consumed the meal provided 2 hours before collection of the postprandial blood sample. Mean ± SD preprandial PCV was 49.5 ± 4.0% (range, 44% to 55%). Mean preprandial rectal temperature was 38.7 ± 0.42°C (range, 38° to 39.4°C), and mean postprandial rectal temperature was 38.2 ± 0.42°C (range, 37.9° to 39.1°C). Mean preprandial systolic blood pressure was 139.2 ± 21.2 mm Hg (range, 110 to 170 mm Hg), and mean postprandial systolic blood pressure was 135.0 ± 14.8 mm Hg (range, 110 to 160 mm Hg).
Effect of prandial state
Results for the 48 preprandial blood samples were compared with results for the 48 postprandial blood samples, and there was no significant (P = 0.846) difference between the mean preprandial and postprandial blood glucose concentrations. The interaction between prandial state and collection site did not have a significant (P = 0.559) effect on blood glucose concentrations.
Effect of collection site
Collection site had a significant (P = 0.001) effect on blood glucose concentrations (Table 1). Glucose concentrations in capillary blood samples collected from the oral mucosa and carpal pad did not differ significantly from each other, but glucose concentrations in both of those samples differed significantly (P = 0.021 and P = 0.002) from the glucose concentration in the venous blood sample (Figure 1). Glucose concentration in the samples obtained from the medial aspect of the pinna did not differ significantly (P = 0.298) from the glucose concentration in the venous blood sample, but they differed significantly from the glucose concentrations in the samples obtained from the oral mucosa (P = 0.003) and the carpal pad (P < 0.001).
Mean ± SD glucose concentrations (mg/dL) measured with a PBGM in blood samples collected from various anatomic locations of 12 dogs after food was withheld for 12 hours (preprandial state) and 2 hours after dogs consumed a meal (postprandial state).
| Preprandial state | Postprandial state | |||
|---|---|---|---|---|
| Site of blood collection* | Blood glucose | Difference† | Blood glucose | Difference† |
| Venous | 88.42 ± 10.56a | — | 87.67 ± 14.26a | — |
| Carpal pad | 65.58 ± 19.68b | −22.8 ± 17.9 | 71.92 ± 29.58b | −15.75 ± 28.7 |
| Medial aspect of pinna | 89.25 ± 11.90a | 0.83 ± 10.38 | 92.42 ± 20.26a | 4.75 ± 14.89 |
| Oral mucosa | 82.58 ± 12.81b | −5.83 ± 11.09 | 78.25 ± 18.72b | −9.42 ± 9.91 |
Venous blood samples were collected from a lateral saphenous vein, and capillary blood samples were collected from the carpal pad, medial aspect of the pinna of an ear, and oral mucosa.
Represents the difference from the glucose concentration of the venous blood sample.
Within a column, values with different superscript letters differ significantly (P < 0.05).
= Not applicable.
Mean ± SD glucose concentrations measured with a PBGM in blood samples collected from a lateral saphenous vein (venous) and capillary blood samples collected from the carpal pad, oral mucosa, and medial aspect of the pinna of an ear of 12 healthy adult dogs.a,b Values with different letters differ significantly (P < 0.05).
Citation: American Journal of Veterinary Research 80, 11; 10.2460/ajvr.80.11.995
Mean ± SD glucose concentrations measured with a PBGM in blood samples collected from a lateral saphenous vein (venous) and capillary blood samples collected from the carpal pad, oral mucosa, and medial aspect of the pinna of an ear of 12 healthy adult dogs.a,b Values with different letters differ significantly (P < 0.05).
Citation: American Journal of Veterinary Research 80, 11; 10.2460/ajvr.80.11.995
Mean ± SD glucose concentrations measured with a PBGM in blood samples collected from a lateral saphenous vein (venous) and capillary blood samples collected from the carpal pad, oral mucosa, and medial aspect of the pinna of an ear of 12 healthy adult dogs.a,b Values with different letters differ significantly (P < 0.05).
Citation: American Journal of Veterinary Research 80, 11; 10.2460/ajvr.80.11.995
Evaluation of Bland-Altman scatterplots revealed bias in glucose concentrations for each collection site. Mean differences between glucose concentrations in venous blood and capillary blood from the medial aspect of the pinna, oral mucosa, and carpal pad were 2.8 mg/dL, −7.6 mg/dL, and −19.3 mg/dL, respectively (Figure 2).
Bland-Altman plots of glucose concentrations in venous blood and capillary blood collected from the oral mucosa (A), medial aspect of the pinna of an ear (B), and carpal pad (C) of 12 healthy adult dogs. Notice that the scale on the y-axis differs among panels. The bias (mean difference) calculated for each collection site is indicated (dotted line). BG = Blood glucose.
Citation: American Journal of Veterinary Research 80, 11; 10.2460/ajvr.80.11.995
Bland-Altman plots of glucose concentrations in venous blood and capillary blood collected from the oral mucosa (A), medial aspect of the pinna of an ear (B), and carpal pad (C) of 12 healthy adult dogs. Notice that the scale on the y-axis differs among panels. The bias (mean difference) calculated for each collection site is indicated (dotted line). BG = Blood glucose.
Citation: American Journal of Veterinary Research 80, 11; 10.2460/ajvr.80.11.995
Bland-Altman plots of glucose concentrations in venous blood and capillary blood collected from the oral mucosa (A), medial aspect of the pinna of an ear (B), and carpal pad (C) of 12 healthy adult dogs. Notice that the scale on the y-axis differs among panels. The bias (mean difference) calculated for each collection site is indicated (dotted line). BG = Blood glucose.
Citation: American Journal of Veterinary Research 80, 11; 10.2460/ajvr.80.11.995
All glucose concentrations for venous blood samples were classified as euglycemic. However, 10 of 24 (41.67%) results obtained for samples from the carpal pad were categorized as hypoglycemic (≤ 60 mg/dL). Mean blood glucose concentration of these samples was 37.4 mg/dL, which was 43.17% less than the venous blood glucose concentration; the lowest blood glucose concentration in samples obtained from the carpal pad was 24 mg/dL. One of 24 (4.17%) results for samples obtained from the oral mucosa and 1 of 24 (4.17%) results for samples obtained from the medial aspect of the pinna were classified as hypoglycemic.
None of the venous or capillary blood samples were classified as hyperglycemic. However, 13 (54.17%) results for samples obtained from the medial aspect of the pinna (mean difference, 12 mg/dL; range, 3 to 31 mg/dL), 4 (16.67%) results for samples obtained from the oral mucosa (mean difference, 6 mg/dL; range, 1 to 13 mg/dL), and 4 (16.67%) results for samples obtained from the carpal pad (mean difference 16 mg/dL; range, 9 to 26 mg/dL) were higher than concurrent glucose concentrations in venous blood samples.
Consensus error grid analysis
Consensus error grid analysis revealed that all measurements were in zone A or B, but glucose concentrations in samples obtained from the carpal pad were more frequently in zone B than were glucose concentrations in samples obtained from the oral mucosa or medial aspect of the pinna. Fifteen of 72 (20.83%) measurements were in zone B, and 10 of those measurements were for blood samples obtained from the carpal pad (10/24 [41.67%] results for samples obtained from the carpal pad). Of the remaining measurements in zone B, 2 (8.33%) were for samples obtained from the oral mucosa, and 3 (12.50%) were for samples obtained from the medial aspect of the pinna (Figure 3).
Consensus error grids for glucose concentrations measured in venous blood samples and capillary blood samples collected from the oral mucosa (A and B), carpal pad (C and D), and medial aspect of the pinna (E and F) of 12 healthy dogs after food was withheld for 12 hours (preprandial state; A, C, and E) and 2 hours after dogs consumed a meal (postprandial state; B, D, and F). Each circle represents results for 1 dog. The grid is divided into 5 zones as follows: zone A, results will not have an effect on clinical action; zone B, results will alter clinical action, but the action will have minimal effect on the clinical outcome; zone C, results will prompt clinical action that is likely to affect clinical outcome; zone D, results will cause errors in treatment that could pose a substantial medical risk to the patient; and zone E, results will cause severe treatment errors that could have dangerous consequences for the patient.
Citation: American Journal of Veterinary Research 80, 11; 10.2460/ajvr.80.11.995
Consensus error grids for glucose concentrations measured in venous blood samples and capillary blood samples collected from the oral mucosa (A and B), carpal pad (C and D), and medial aspect of the pinna (E and F) of 12 healthy dogs after food was withheld for 12 hours (preprandial state; A, C, and E) and 2 hours after dogs consumed a meal (postprandial state; B, D, and F). Each circle represents results for 1 dog. The grid is divided into 5 zones as follows: zone A, results will not have an effect on clinical action; zone B, results will alter clinical action, but the action will have minimal effect on the clinical outcome; zone C, results will prompt clinical action that is likely to affect clinical outcome; zone D, results will cause errors in treatment that could pose a substantial medical risk to the patient; and zone E, results will cause severe treatment errors that could have dangerous consequences for the patient.
Citation: American Journal of Veterinary Research 80, 11; 10.2460/ajvr.80.11.995
Consensus error grids for glucose concentrations measured in venous blood samples and capillary blood samples collected from the oral mucosa (A and B), carpal pad (C and D), and medial aspect of the pinna (E and F) of 12 healthy dogs after food was withheld for 12 hours (preprandial state; A, C, and E) and 2 hours after dogs consumed a meal (postprandial state; B, D, and F). Each circle represents results for 1 dog. The grid is divided into 5 zones as follows: zone A, results will not have an effect on clinical action; zone B, results will alter clinical action, but the action will have minimal effect on the clinical outcome; zone C, results will prompt clinical action that is likely to affect clinical outcome; zone D, results will cause errors in treatment that could pose a substantial medical risk to the patient; and zone E, results will cause severe treatment errors that could have dangerous consequences for the patient.
Citation: American Journal of Veterinary Research 80, 11; 10.2460/ajvr.80.11.995
Discussion
Accurate monitoring of blood glucose concentrations is necessary to determine treatment protocols for patients unable to maintain normoglycemia. A PBGM is a readily available and easy-to-use diagnostic device that provides blood glucose measurements through the analysis of a small amount of arterial, venous, or capillary blood. Accuracy of the PBGM used in the study reported here has been previously evaluated with venous blood samples from diabetic patients,5,7 and the device was found to provide clinically acceptable results. However, the authors are not aware of any studies conducted to evaluate accuracy of this PBGM for capillary blood samples obtained from nondiabetic dogs.
Results of the present study suggested that location for collection of a blood sample had a significant effect on blood glucose concentrations, with greater variability of measurement for capillary blood samples obtained from some collection sites. Evaluation of values plotted on the consensus error grid revealed that all results for the study reported here were in zone A (clinically acceptable) or zone B (benign error).15 Although values in zone B are defined as being of little clinical concern,15 the authors believed that some of the results in this zone were alarmingly low, with 2 blood glucose concentrations < 30 mg/dL, 1 blood glucose concentration < 40 mg/dL, and 5 blood glucose concentrations < 50 mg/dL. Overall, 10 of 24 (41.67%) samples obtained from the carpal pad (5 preprandial samples and 5 postprandial samples) would have been incorrectly classified as hypoglycemic (blood glucose concentration < 60 mg/dL). The mean blood glucose concentration of these samples was 37.4 mg/dL, which was 43.17% less than the blood glucose concentration in venous samples.
Clinicians and pet owners should be aware that analysis of glucose concentrations in capillary blood samples obtained from a non-weight-bearing portion of a paw pad or the oral mucosa by use of this specific PBGM may yield misleading results that could prompt unnecessary interventions. For example, in a diabetic dog, a blood glucose concentration of 24 mg/dL might prompt an owner to decrease or not administer a dose of insulin, whereas the same blood glucose concentration in a clinically ill, nondiabetic dog might prompt a veterinarian to perform diagnostic tests to determine an underlying cause, such as sepsis or insulinoma. Analysis of data for the study reported here supported the use of capillary blood samples obtained from the medial aspect of the pinna, instead of capillary blood samples obtained from the oral mucosa or carpal pad. Additionally, we recommend that all results be interpreted in light of the patient history, results of physical examination, findings of prior diagnostic tests, and comorbidities.
The PBGM used in the study reported here can provide accurate and precise glucose measurements for venous blood samples. However, this PBGM has also been found to provide blood glucose concentrations that are higher than the reference glucose concentration.8 This finding was evident in the present study because 13 (54.17%) glucose concentrations for samples obtained from the medial aspect of the pinna, 4 (16.67%) glucose concentrations for samples obtained from the oral mucosa, and 4 (16.67%) glucose concentrations for samples obtained from the carpal pad were higher (mean difference, 12 mg/dL [range, 3 to 31 mg/dL]; mean difference, 6 mg/dL [range, 1 to 13 mg/dL]; and mean difference, 16 mg/dL [range, 9 to 26 mg/dL], respectively) than the glucose concentration in the venous blood sample. These results were consistent with those of another study8 in which 43% of the samples evaluated by use of this PBGM had blood glucose concentrations higher than the reference value. However, additional caution should be used when interpreting results provided by this PBGM because investigators of another study3 found that PBGMs commonly underestimate, rather than overestimate, blood glucose concentrations.
Another finding for the study reported here was the fact that there was no significant difference in blood glucose concentrations in healthy dogs measured immediately before and 2 hours after a 12-hour period of food withholding. This result is consistent with the pattern of blood glucose homeostasis in humans in whom blood glucose concentrations increase and then insulin is released from pancreatic β cells to activate glucose transporters and restore blood glucose concentrations to baseline values within 2 hours after the final absorption of carbohydrates.16
The present study had several limitations, including the small number of subjects and use of a single PBGM. Although the number of subjects was large enough to detect significant differences, it increased the likelihood of a type II statistical error. This was particularly true for the results for samples obtained from the oral mucosa, for which the concentration of 1 sample was an obvious outlier from results for the remainder of the samples. This value may have been falsely decreased because of sample dilution with saliva, analysis of an insufficient blood volume, or malfunction of a test strip, although no errors were noted on the PBGM display.
For the study reported here, a single PBGM and test strips with the same lot number were used to minimize variability. Although the PBGM was appropriately calibrated before it was used and there were no errors noted during use of the glucometer, it is possible that the device or all test strips of that lot number could have been defective. This scenario is unlikely because all testing materials were purchased directly from a veterinary pharmacy, the PBGM was evaluated and calibrated before the experiment, and the glucometer did not have any complications during the study.
Monitoring blood glucose concentrations with a PBGM that can analyze capillary blood is appealing because a small amount of blood is required, results are rapidly available, there is a minimal amount of trauma to patients, peripheral vasculature is not compromised, and glucometers are easy to use and readily available. Despite the appeal of these features, results of the present study suggested that for healthy dogs, capillary blood samples can yield blood glucose concentrations that differ significantly from blood glucose concentrations of venous blood samples. However, when it is not possible to obtain a venous sample for determination of a blood glucose concentration, capillary blood should be collected from the medial aspect of the pinna. Additional studies are warranted to determine the effect of comorbidities on the accuracy of glucose concentrations in capillary blood samples.
Acknowledgments
Supported by the Companion Animal Fund at the University of Tennessee College of Veterinary Medicine. Funding sources were not involved in the study design, data analysis and interpretation, or writing of the manuscript.
The authors declare that there were no conflicts of interest.
ABBREVIATIONS
| PBGM | Portable blood glucose meter |
Footnotes
AlphaTRAK 2 blood glucose monitoring system, Zoetis, Parsippany-Troy Hills, NJ.
AlphaTRAK 2 blood glucose test strips, Zoetis, Parsippany-Troy Hills, NJ.
AlphaTRAK lancets, Zoetis, Parsippany-Troy Hills, NJ.
Research Randomizer. Available at: www.randomizer.org.
AlphaTRAK lancing device, Zoetis, Parsippany-Troy Hills, NJ.
SAS, version 9.4, SAS Institute Inc, Cary, NC.
References
1. Casella M, Wess G, Reusch CE. Measurement of capillary blood glucose concentrations by pet owners: a new tool in the management of diabetes mellitus. J Am Anim Hosp Assoc 2002;38:239–245.
2. Wess G, Reusch C. Capillary blood sampling from the ear of dogs and cats and use of portable meters to measure glucose concentration. J Small Anim Pract 2000;41:60–66.
3. Wess G, Reusch C. Evaluation of five portable blood glucose meters for use in dogs. J Am Vet Med Assoc 2000;216:203–209.
4. Wess G, Reusch C. Assessment of five portable blood glucose meters for use in cats. Am J Vet Res 2000;61:1587–1592.
5. Kang MH, Kim DH, Jeong IS, et al. Evaluation of four portable blood glucose meters in diabetic and non-diabetic dogs and cats. Vet Q 2016;36:2–9.
6. Mori A, Oda H, Onozawa E, et al. Evaluation of newly developed veterinary portable blood glucose meter with hematocrit correction in dogs and cats. J Vet Med Sci 2017;79:1690–1693.
7. Cohn LA, McCaw DL, Tate DJ, et al. Assessment of five portable blood glucose meters, a point-of-care analyzer, and color test strips for measuring blood glucose concentration in dogs. J Am Vet Med Assoc 2000;216:198–202.
8. Cohen TA, Nelson RW, Kass PH, et al. Evaluation of six portable blood glucose meters for measuring blood glucose concentration in dogs. J Am Vet Med Assoc 2009;235:276–280.
9. Zoetis Inc. AlphaTRAK 2 blood glucose test strips, product information, 2018. Kalamazoo, Mich: Zoetis Inc. Available at: www.zoetisus.com/contact/pages/product_information/msds_pi/pi/alphatrak_2.pdf. Accessed Aug 9, 2018.
10. Zoetis Inc. AlphaTRAK blood glucose monitoring system, quick user's guide, 2018. Kalamazoo, Mich: Zoetis Inc. Available at: www.zoetisus.com/products/dogs/alphatrakmeter/pdf/alphatrak-2-quick-user_s-guide.pdf. Accessed Aug 9, 2018.
11. Borin S, Crivelenti LZ, Hoeppner Rondelli MC, et al. Capillary blood glucose and venous blood glucose measured with portable digital glucometer in diabetic dogs. Braz J Vet Pathol 2012;5:42–46.
12. Togashi Y, Shirakawa J, Okuyama T, et al. Evaluation of the appropriateness of using glucometers for measuring the blood glucose levels in mice. Sci Rep 2016;6:25465.
13. Juneja D, Pandey R, Singh O. Comparison between arterial and capillary blood glucose monitoring in patients with shock. Eur J Intern Med 2011;22:241–244.
14. Bland JM, Altman G. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–310.
15. Parkes JL, Slatin SL, Pardo S, et al. A new consensus error grid to evaluate the clinical significance of inaccuracies in the measurement of blood glucose. Diabetes Care 2000;23:1143–1148.
16. Hall JE. Guyton and Hall textbook of medical physiology. 11th ed. Philadelphia: Elsevier, 2016;971–972.
