Veterinary Clinical Pathology ISSN 0275-6382
ORIGINAL RESEARCH
Assessment of the accuracy and precision of the i-Smart 30 VET Electrolyte Analyzer in dogs, cats, cattle and pigs Han-Jun Kim1, Hye-Rim Lee1, Yoon-Seo Park2, Soon-Goo Kyung3, Sun Hee Do1 1
Department of Clinical Pathology, College of Veterinary Medicine, Konkuk University, Seoul, Korea; 2Gyeonggi Large Animal Clinic, Gyeonggi-do, Korea; and 3Jangsu Stud Farm, Korea Racing Authority, Jeollabuk-do, Korea
Key Words Chloride, method comparison, point-of-care, potassium, precision, sodium Correspondence S. H. Do, Department of Clinical Pathology, College of Veterinary Medicine, Konkuk University, Seoul 143-701, Korea E-mail: [email protected] DOI:10.1111/vcp.12267
Background: Performance evaluation of point-of-care (POC) electrolyte analyzers is essential for determining their precision and accuracy in clinical practice. Objective: The purpose of this study was to validate the i-Smart 30 VET Electrolyte Analyzer for canine, feline, bovine, and porcine samples in comparison with the ion-selective electrolyte analyzer Roche 9180 electrolyte analyzer. Methods: A total of 400 heparinized whole blood samples were collected and analyzed by both instruments for sodium, potassium, and chloride concentrations. Within-run, between-day, and total imprecision were evaluated. Statistical analyses included tests for correlation, regression, bias, and total error. Results: The coefficients of variation (CV) of both within-run and between-day imprecisions in the i-Smart 30 VET ranged from 0.4–1.6%. In addition, total CV (0.3–1.7%) and total error (0.7–3.7%) of the i-Smart 30 VET were acceptable according to the ASVCP guidelines (< 5%). The correlation between the i-Smart 30 VET and the Roche 9180 was excellent (r > .98). There was no proportional error according to the regression (slope ranges 0.92–1.00, 95% CI includes 1.00), but a constant error was detected for sodium concentration in dogs (interval = 0.5), cattle (interval = 3.0), and pigs (interval = 4.0), and for chloride concentration in cats (interval = 1.0). Most of the bias was within 95% CI, and the total error range (0.8–3.5%) was acceptable according to ASVCP guidelines. Conclusion: The i-Smart 30 VET Electrolyte Analyzer provides precise and accurate measurements of sodium, potassium, and chloride concentrations in whole blood samples from dogs, cats, cattle, and pigs.
Introduction Point-of-care (POC) testing has been developed over several decades, enabling clinicians to initiate appropriate treatment in both human and animal patients.1–6 Use of POC tests is beneficial because the results are quickly obtained, which is especially important in emergency situations. In addition, the modest blood volume requirements reduce patients’ blood loss related to blood sampling. However, POC instruments designed for human specimens may be inaccurate with animal blood and may require minimal technical skills from operators. Therefore, it is important to validate and compare the performance of newly
410
developed POC analyzers to the gold standard analytic methods.7–9 The purpose of this study was to assess the precision and accuracy of the i-Smart 30 VET Electrolyte Analyzer in comparison to a Roche reference instrument, and to determine if the 2 analyzers can be used interchangeably.
Materials and Methods Instruments The reference instrument was an ion-selective electrode analyzer (ISEA) Roche 9180 (Roche Diag-
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Table 1. Reportable ranges of Roche 9180 and i-Smart 30 VET electrolyte analyzers in whole blood, serum, plasma, and aqueous solutions. Analyte Sodium (mmol/L) Potassium (mmol/L) Chloride (mmol/L)
Roche 9180
i-Smart 30 VET
40–205 1.5–15 50–200
20–250 0.5–20.0 20–250
nostics, Indianapolis, IN, USA) operating with 5 different ion-selective electrodes (ISE) and requiring 95 lL sample volume. It is fully automated, including calibration. Sodium, potassium, chloride, ionized calcium, and lithium concentrations can be measured in samples of heparinized whole blood, serum, plasma, urine, dialysate, and aqueous standard solutions. The reportable ranges of each analyte for Roche 9180 electrolyte analyzer are shown in Table 1.10 The i-Smart 30 VET Electrolyte Analyzer (i-SENS, Seoul, Korea) is a portable electrolyte analyzer developed for veterinary medicine operating with 3 ISE. Sodium, potassium, and chloride concentrations can be measured in 60 lL of heparinized whole blood, heparinized plasma, and serum samples. The reportable ranges for each analyte of i-Smart 30 VET Electrolyte Analyzer are shown in Table 1.11 For the actual sample application, there are cartridges (30–200 tests per cartridge) containing the required components for the individual analyses, including ion-specific electrochemical sensor cards, calibration solution, sampler, a waste reservoir, a valve, and tubes. The screen-printed, planar-typed micro-electrode ISE contains a polymeric sensing membrane dispensed over an internal hydrogel layer. The sensor card contains miniaturized ion-specific electrochemical sensors composed of thin, small plastic substrates. The sample is introduced to the sensor and analysis proceeds automatically. The result is obtained in 35 seconds and displayed on the LCD screens, and may be printed out. Cleaning and one-point calibration are performed after each test. Both the Roche and the i-Smart analyzer compare the ion activity of an unknown electrolyte concentration to a reference solution. An ion-selective membrane reacts with the specific electrolyte, and the difference between the membrane potential and reference electrode are calculated. The ion concentration of the sample is determined from the measured electrical potential (voltage) using the Nernst equation by direct potentiometry.12,13 The default reference intervals are provided in both electrolyte analyzers and it is possible to adapt laboratory-specific normal ranges by users. In addition, the correction factors can be applied through their
own programming menu. Users can set correction factors for both proportional (slope) and constant (intercept) errors.10,11 The i-Smart 30 VET Electrolyte Analyzer can store reference intervals and correction factors for up to 6 different species.
Sample collection This study was approved by the Institutional Animal Care and Use Committee (IACUC, Approval No. KU12022, April 15, 2012) of the University of Konkuk. A total of 400 heparinized whole blood samples—100 each from dogs, cats, cattle, and pigs—were tested for performance evaluation of the i-Smart 30 VET Electrolyte Analyzer. Samples were collected in lithiumheparin tubes (BD Vacutainer; Becton Dickinson, Franklin Lakes, NJ, USA) irrespective of sex, age, breed, and health status, from dogs and cats presented to the teaching hospital, College of Veterinary Medicine, Konkuk University, and from cattle and pigs during visits at local farms, with their owners’ consent. The samples were kept on ice or refrigerated at 4°C during transport. Samples were transported to the laboratory within 2 hours of collection and analyzed within another 2 hours.
Sample analysis and determination of accuracy, precision, correlation, and bias Analyses were performed by trained veterinarians under standardized conditions. Only nonhemolytic, nonclotted, and nonlipemic whole blood samples were analyzed. The experimental materials, including cartridges and i-Smart Electrolyte QC multilevel control reagents (i-SENS, Seoul, Korea), were used within quality specifications up to their indicated expiration dates. The 500 lL samples of heparinized whole blood were analyzed in duplicates and on the same day at about the same time on both analyzers for sodium, potassium, and chloride concentrations. A one-point calibration was performed prior to each measurement, and a 2-point calibration was performed every 4 and 6 hours for the Roche 9180 and i-Smart 30 VET, respectively, ensuring precision and consistency. To monitor performance, a daily control run was performed with known sodium, potassium, and chloride concentrations at 3 levels (low, medium, and high). Three levels of quality control materials (i-Smart Electrolyte QC multi-levels) for each analyte were used to determine within-run and between-day imprecision. Two replicates of each sample were measured in 2 runs per day for 20 consecutive days.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
411
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Kim et al
Statistical analyses Means, standard deviation (SD), and coefficient of variation (CV) were calculated from repeat measurements. Within-run (SDwr), within-day (SDwd), day-to-day (SDdd), between-run (SDbr), and between-day (SDbd) precision were calculated.14 Precision data were calculated using the following equations: Within-Run Precision : SDwr vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uP 2 uT P u ðX Xij2 Þ2 ti¼1 j¼1 ij1 ¼ 4T Within-Day Precision : SDwd vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uT u P mean mean 2 u ðXi1 Xi2 Þ ti¼1 ¼ 2T Day-to-Day Precision : SDdd vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uT u P mean u ðXi X mean Þ2 ti¼1 ¼ T 1 Between-run Precision : SDbr qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ SD2wd SD2wr =2
(inaccuracy) in analytic measurement. The total error was calculated using the following equation: TEo % ¼ 2CV þ Bias% The calculated total error was compared to the total allowable error (TEa) according to the ASVCP guidelines for allowable total error.15 To compare the i-Smart 30 VET Electrolyte Analyzer to the Roche analyzer, the Kolmogorov–Smirnov test was used to verify the assumption of normality. The correlation of sodium, potassium, and chloride concentrations between analyzers was determined by Pearson correlation coefficients. Correlations were characterized as excellent (r ≥ 0.95), very good (r = 0.90–0.94), good (r = 0.80–0.89), fair (r = 0.59–0.79), or poor (r < 0.59).16,17 Proportional and constant error between the 2 analyzers were determined using the Passing–Bablok regression analysis.18 A 95% confidence interval (CI) for slope and intercept that did not include a value of 1.0 and zero, respectively, indicated significant proportional and constant error. Bland–Altman plots were constructed to determine mean bias and the SD of the bias.19 TEo was calculated using the bias from the analyzer comparison and the total CV of the precision study according to the equation Bias% ¼
Between-day Precision : SDbd qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ SD2dd SD2wd =2 where T is the total number of days; j the run number mean within-day (1 or 2); i the day; Xi1 the average of the mean the average of all results day i; 2 replicates day i; Xi Xmean the average of all results. The total precision was calculated using the following equation: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SDTotal¼ SD2betweendayþSD2betweenrunþSD2withinrun
Bias Meancomparison study 100 Meanspeciesspecific from comparison study
TEo % ¼ 2 total CV (from precision study) þ Bias% ðfrom method comparisonÞ Comparability was assessed by comparing observed total error to the ASVCP’s TEa, which is 5% for all electrolytes, except for low concentrations of potassium, for which TEa is 10%.20 The study design and statistical analysis are described as a flowchart in Figure 1.
The bias for each analyte was also calculated to assess systemic analytical errors in comparison with the manufacturer’s provided target values.15 The bias (%) and total CV were calculated as follows: Bias% ¼
Biasmanufacturer Meanmeasured 100 Meanmanufacturer CV% ¼
SD 100 Mean
Observed total error (TEo) was calculated according to the inaccuracy (bias) and imprecision (CV) of the overall error that occurred as a combined effect of random error (imprecision) and systemic error
412
Figure 1. Flowchart illustrating the study design for performance evaluation of the i-Smart 30 VET Electrolyte Analyzer.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Table 2. Precision of the i-Smart 30 VET Electrolyte Analyzer for 3 levels of i-Smart Electrolyte quality control samples. Manufacturer’s Target Ranges Analyte Na
Within-Run
Between-Day
Total
Level
N
Mean
Mean (mmol/L)
SD (mmol/L)
CV (%)
SD (mmol/L)
CV (%)
SD (mmol/L)
CV (%)
TE (%)
1 2 3 1 2 3 1 2 3
40 40 40 40 40 40 40 40 40
111–121 (116) 132–142 (137) 152–162 (157) 1.5–2.5 (1.9) 3.5–4.5 (4.0) 5.4–6.4 (5.9) 74–84 (79) 96–106 (101) 121–131 (126)
116.3 136.9 156.1 1.90 3.99 5.91 79.2 101.1 126.4
0.63 0.43 0.44 0.025 0.030 0.050 0.49 0.27 0.32
0.5 0.3 0.3 1.3 0.7 0.8 0.6 0.3 0.3
0.09 0.26 1.48 0.018 0.045 0.040 0.26 0.04 0.89
0.1 0.2 1.0 1.0 1.1 0.7 0.3 0.0 0.7
0.76 0.50 1.56 0.033 0.053 0.064 0.59 0.32 1.05
0.7 0.4 1.0 1.7 1.3 1.1 0.7 0.3 0.8
1.6 0.8 2.6 3.7 2.8 2.4 1.8 0.7 2.0
+
K+
Cl
N indicates the number of runs; CV, coefficient of variation; TE, total error.
Analyses were performed using SPSS version 19.0 (SPSS Inc., Chicago, IL, USA) and Analyse-it version 3.60, Method Validation Edition (Analyse-it Software, Ltd., Leeds, UK), including CV, the Kolmogorov–Smirnov test, Pearson correlation analysis, Passing–Bablok regression, and Bland–Altman plots. For all tests, a P < .001 was considered statistically significant.
Results Precision and accuracy The precision of the i-Smart 30 VET for sodium, potassium, and chloride concentrations are shown in Table 2. The CV for within-run and between-day assays were between 0.1% and 1.3% for level 1, between 0.0 and 1.1% for level 2, and between 0.3% and 1.0% for level 3 controls. Total imprecision and
TEo for all analytes based on quality control materials were < 2% and < 4%, respectively.
Instrument comparison The results obtained using the i-Smart 30 VET Electrolyte Analyzer were compared to the Roche 9180 on the basis of the Pearson correlation coefficients, Passing– Bablok regression analysis, and Bland–Altman difference plots (Table 3, Figures 2–5). The correlation between i-Smart 30 VET Electrolyte Analyzer and the Roche 9180 was excellent for sodium, potassium, and chloride concentrations in canine, feline, bovine, and porcine blood (r ≥ .982, P < .0001). Results of Passing–Bablok regression showed that the slopes for all 3 electrolytes were ranging from 0.92–1.00. The 95% CI of the slopes for all 3 electrolytes included 1.00 indicating that there was
Table 3. Pearson correlation coefficient, Passing–Bablok regression analysis, and Bland–Altman plots data for i-Smart 30 VET and Roche 9180 electrolyte analyzer comparison
Species
Analyte
Range (mmol/L)
Pearson Coefficiency*
Passing–Bablok’s Slope (95% CI)
Passing–Bablok’s Intercept (95% CI)
Bias mean † (95% CI)
Bias SD†
Bias %
TE Observed
Canine
Na+ K+ ClNa+ K+ Cl Na+ K+ Cl Na+ K+ Cl
118–175 2.6–6.7 87–130 126–164 2.4–6.5 94–141 114–159 2.4–7.0 79–112 122–168 3.1–7.4 82–125
0.992 0.996 0.987 0.991 0.997 0.994 0.996 0.997 0.984 0.992 0.983 0.982
1.00 (1.00–1.00) 1.00 (0.96–1.00) 1.00 (0.92–1.00) 1.00 (1.00–1.05) 1.00 (1.00–1.00) 1.00 (0.94–1.00) 1.00 (1.00–1.00) 1.00 (0.96–1.00) 0.94 (0.87–1.00) 1.00 (1.00–1.00) 0.96 (0.89–1.00) 0.92 (0.86–1.00)
0.5 (0.5–0.5) 0.0 (0.0–0.2) 0.0 (0.0–8.8) 2.0 (4.9–2.0) 0.0 (0.0–0.0) 1.0 (1.0–7.9) 3.0 (3.0–3.0) 0.0 (0.0–0.2) 3.3 (3.0–10.5) 4.0 (4.0–4.0) 0.0 (0.2–0.4) 7.3 (1.0–13.3)
0.4 (2.1–2.9) 0.0 (0.3–0.2) 0.3 (2.6–3.2) 1.7 (0.7–4.1) 0.1 (0.2–0.1) 0.6 (1.5–2.6) 3.4 (1.9–4.8) 0.0 (0.2–0.2) 2.6 (5.4–0.2) 3.6 (1.6–5.7) 0.1 (0.5–0.2) 1.4 (4.2–1.5)
1.26 0.12 1.47 1.20 0.08 1.03 0.72 0.09 1.40 1.02 0.18 1.44
0.3 0.0 0.3 1.1 2.2 0.5 2.5 0.0 2.6 2.6 2.1 1.4
1.1 0.8 1.1 2.0 3.0 1.3 3.3 0.8 3.5 3.4 3.0 2.2
Feline
Bovine
Porcine
CI indicates confidence interval. *P < .0001. † Mean and SD value of Bias from Bland–Altman plot.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
413
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Kim et al
Figure 2. Passing–Bablok regression analyses (left) and Bland–Altman difference plots (right) for the i-Smart 30 VET and the Roche 9180 electrolyte analyzer for sodium, potassium, and chloride concentrations in canine blood (n = 100). In the regression plots, the black dotted line indicates the line of identity (y = x), and the red line is the line of best fit. In the difference plots, the black dashed horizontal line represents the mean bias, and the red dotted horizontal line represents the 95% limits of agreement.
no proportional bias between the 2 analyzers for any tested species. The intercept of the regression included 0.0 within 95% CI except for sodium concentration in dogs (95% CI = 0.5–0.5), cattle (95% CI = 3.0–3.0), and pigs (95% CI = 4.0–4.0) as well as
414
chloride concentration in cats (95% CI = 1.0–7.9). These results are indicative of a constant error in sodium concentrations in dogs, cattle, and pigs, and chloride concentrations in cats when measured on the i-Smart 30 VET.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Figure 3. The Passing–Bablok regression analyses (left) and Bland–Altman difference plots (right) for the i-Smart 30 VET and Roche 9180 electrolyte analyzers for sodium, potassium, and chloride in feline blood (n = 100). In the regression plots, the black dotted line indicates the line of identity (y = x), and the red line is the line of best fit. In the difference plots, the black dashed horizontal line represents the mean bias, and the red dotted horizontal line represents the 95% limits of agreement.
The ranges of the mean differences in the Bland– Altman plots in all 4 species were 0.4–3.6 mmol/L (SD ranges = 0.72–1.26) for sodium; 0.1–0.0 mmol/L (SD ranges = 0.08–0.12) for potassium; and 2.6– 0.6 mmol/L (SD ranges = 1.03–1.47) for chloride. The Bland–Altman plots showed that with the exception of
a few values most results were within the 95% CI. In all 4 species, bias and TEo were < 3% (range 0.0–2.6%) and < 5% (range 0.8–3.5%), respectively, for sodium, potassium, and chloride concentrations measured by the i-Smart 30 VET when compared to the Roche 9180.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
415
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Kim et al
Figure 4. The Passing–Bablok regression analyses (left) and Bland–Altman difference plots (right) for the i-Smart 30 VET and Roche 9180 electrolyte analyzers for sodium, potassium, and chloride in bovine blood (n = 100). In the regression plots, the black dotted line indicates the line of identity (y = x), and the red line is the line of best fit. In the difference plots, the black dashed horizontal line represents the mean bias, and the red dotted horizontal line represents the 95% limits of agreement.
Discussion POC analyzers are powerful and life-saving tools used by clinicians for critical care and emergency cases in both human and veterinary medicine.2,3 The ASVCP defines ‘accuracy’ as the closeness of agreement
416
between the results of a measurement and the true concentration of an analyte, and ‘comparability’ as the closeness of results from 2 or more instruments that process the same specimen, ensuring that results from different instruments are similar enough to be used interchangeably without clinical error.15
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Figure 5. The Passing–Bablok regression analyses (left) and Bland–Altman difference plots (right) for the i-Smart 30 VET and Roche 9180 electrolyte analyzers for sodium, potassium, and chloride in porcine blood (n = 100). In the regression plots, the black dotted line indicates the line of identity (y = x), and the red line is the line of best fit. In the difference plots, the black dashed horizontal line represents the mean bias, and the red dotted horizontal line represents the 95% limits of agreement.
Our study shows that the i-Smart 30 VET yielded acceptable CV for the within-run and between-day imprecisions for sodium, potassium, and chloride in quality control materials. The total imprecision CV and TEo were within quality specification that fulfill required
criteria of the ASVCP guidelines for total allowable error (ie, TEa < 5%).15 Based on the results of the method comparison assay, the sodium, potassium, and chloride concentrations in blood samples from dogs, cats, cattle, and pigs
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
417
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Kim et al
measured by the i-Smart 30 VET were comparable with the results from a widely used laboratory analyzer, the Roche 9180. The correlation between the results produced by the i-Smart 30 VET and Roche 9180 was excellent (r > .98), and most of the bias was within the 95% CI. There was agreement between the 2 methods for most measurements, except sodium concentration in dogs, cattle, and pigs, and chloride concentration in cats. Mean differences between the iSmart 30 VET and Roche 9180 were near zero, except for slightly higher sodium concentrations in samples from cats, pigs, and cattle, and chloride concentrations in samples from pigs and cattle. Different methods used by the 2 instruments may explain some of the discrepancies. Both Roche 9180 and i-Smart 30 VET analyze ion concentration by potentiometric measurement using ISE; however, the Roche 9180 uses a conventional macroelectrode. These macro-electrodes must be replaced every 6 or 12 months by trained engineers, resulting in relatively high maintenance cost. In contrast, the i-Smart 30 VET uses a screen-printed planar ISE.10,11 Micro-electrodes with a diameter < 25 lM have several advantages such as small size, minimization of solution resistance effects, and rapid response times.12,13 However, the electrochemical behavior of a micro-electrode is slightly different from the macroelectrode due to diffusion from the edges of the electrodes.21,22 As other POC analyzers, the i-Smart 30 VET Electrolyte Analyzer has several advantages. It operates with an all-in-one cartridge, is easy to operate, and produces results quickly. Moreover, the module is relatively small and requires a relatively small blood volume for analysis. Another advantage of this analyzer is that it is easy to remove potential blood clots simply by replacing the cartridge. In conclusion, the i-Smart 30 VET Electrolyte Analyzer provided acceptable results in comparison with the Roche 9180 for sodium, potassium, and chloride concentrations in blood from dogs, cats, cattle, and pigs. Disclosure: The authors have indicated that they have no affiliations or financial involvement with any organization or entity with a financial interest in, or in financial competition with, the subject matter or materials discussed in this article.
References 1. Plebani M. Does POCT reduce the risk of error in laboratory testing? Clin Chim Acta. 2009;404:59–64.
418
2. Kazmierczak SC. Point-of-care testing quality: some positives but also some negatives. Clin Chem. 2011;57:1219–1220. 3. Lee EJ, Shin SD, Song KJ, et al. A point-of-care chemistry test for reduction of turnaround and clinical decision time. Am J Emerg Med. 2011;29:489–495. 4. Wong DM, Gigu ere S, Wendel MA. Evaluation of a point-of-care portable analyzer for measurement of plasma immunoglobulin G, total protein, and albumin concentrations in ill neonatal foals. J Am Vet Med Assoc. 2013;242:812–819. 5. Dixon-Jimenez AC, Brainard BM, Cathcart CJ, Koenig A. Evaluation of a point-of-care coagulation analyzer (Abaxis VSPro) for identification of coagulopathies in dogs. J Vet Emerg Crit Care (San Antonio). 2013;23:402–407. 6. Montesinos A, Ardiaca M. Acid-base status in the avian patient using a portable point-of-care analyzer. Vet Clin North Am Exot Anim Pract. 2013;16:47–69. 7. O’Kane MJ, McManus P, McGowan N, Lynch PL. Quality Error Rates in Point-of-Care Testing. Clin Chem. 2011;57:1267–1271. 8. Beneteau-Burnat B, Pernet P, Pilon A, et al. Evaluation of the GEM Premier 4000: a compact blood gas CO-Oximeter and electrolyte analyzer for point-of-care and laboratory testing. Clin Chem Lab Med. 2008;46:271–279. 9. Smolcic VS, Bilic-Zulle L, Fisic E. Validation of methods performance for routine biochemistry analytes at Cobas 6000 analyzer series module c501. Biochem Med. 2011;21:182–190. 10. Roche Diagnostics. Instructions for Use 9180 Electrolyte Analyzer V5.0. Mannheim, Germany: Roche Diagnostics GmbH; 2008. 11. i-SENS. i-Smart 30 VET electrolyte analyzer Operator’s Manual. Seoul, Republic of Korea: Medical Technology Promedt Consulting GmbH; 2013. 12. Yan S. Electrochemical cells – New Advances in Fundamental Researches and Applications. IA: InTech; 2012:1–15. 13. Yuzuru I, Masao M. Electrochemical Measurements with Interdigitated Array Microelectrodes. Curr Sep. 1995;14:1. 14. Clinical and Laboratory Standards Institute (CLSI). EP10-A2: Preliminary Evaluation of Quantitative Clinical Laboratory Methods; Approved Guideline. 2nd ed. Wayne, PA: CLSI; 2002:41. 15. Harr KE, Flatland B, Nabity MB, Freeman KP. ASVCP guidelines: allowable total error guidelines for biochemistry. Vet Clin Pathol. 2013;42:424–436. 16. Riond B, Weissenbacher S, Hofmann-Lehmann R, Lutz H. Performance evaluation of the Sysmex pocH-100iV Diff hematology analyzer for analysis of canine, feline, equine, and bovine blood. Vet Clin Pathol. 2011;40: 484–495.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
17. Welles EG, Hall AS, Carpenter DM. Canine complete blood counts: a comparison of four in-office instruments with the ADVIA 120 and manual differential counts. Vet Clin Pathol. 2009;38:20–29. 18. Bablok W, Passing H. Application of statistical procedures in analytical instrument testing. J Autom Chem. 1985;7:74–79. 19. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1:307–310.
Evaluation of i-Smart 30 VET Electrolyte Analyzer
20. Gunn-Christie RG, Flatland B, Friedrichs KR, et al. ASVCP quality assurance guidelines: control of preanalytical, analytical, and postanalytical factors for urinalysis, cytology, and clinical chemistry in veterinary laboratories. Vet Clin Pathol. 2012;41:18–26. 21. Christopher MAB, Ana MOB. Electrochemistry: Principles, Methods, and Applications. IA: Oxford University Press; 1993:82–90. 22. Joseph W. Analytical Electrochemistry. IA: Wiley-VCH; 2000:140–161.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
419
ORIGINAL RESEARCH
Assessment of the accuracy and precision of the i-Smart 30 VET Electrolyte Analyzer in dogs, cats, cattle and pigs Han-Jun Kim1, Hye-Rim Lee1, Yoon-Seo Park2, Soon-Goo Kyung3, Sun Hee Do1 1
Department of Clinical Pathology, College of Veterinary Medicine, Konkuk University, Seoul, Korea; 2Gyeonggi Large Animal Clinic, Gyeonggi-do, Korea; and 3Jangsu Stud Farm, Korea Racing Authority, Jeollabuk-do, Korea
Key Words Chloride, method comparison, point-of-care, potassium, precision, sodium Correspondence S. H. Do, Department of Clinical Pathology, College of Veterinary Medicine, Konkuk University, Seoul 143-701, Korea E-mail: [email protected] DOI:10.1111/vcp.12267
Background: Performance evaluation of point-of-care (POC) electrolyte analyzers is essential for determining their precision and accuracy in clinical practice. Objective: The purpose of this study was to validate the i-Smart 30 VET Electrolyte Analyzer for canine, feline, bovine, and porcine samples in comparison with the ion-selective electrolyte analyzer Roche 9180 electrolyte analyzer. Methods: A total of 400 heparinized whole blood samples were collected and analyzed by both instruments for sodium, potassium, and chloride concentrations. Within-run, between-day, and total imprecision were evaluated. Statistical analyses included tests for correlation, regression, bias, and total error. Results: The coefficients of variation (CV) of both within-run and between-day imprecisions in the i-Smart 30 VET ranged from 0.4–1.6%. In addition, total CV (0.3–1.7%) and total error (0.7–3.7%) of the i-Smart 30 VET were acceptable according to the ASVCP guidelines (< 5%). The correlation between the i-Smart 30 VET and the Roche 9180 was excellent (r > .98). There was no proportional error according to the regression (slope ranges 0.92–1.00, 95% CI includes 1.00), but a constant error was detected for sodium concentration in dogs (interval = 0.5), cattle (interval = 3.0), and pigs (interval = 4.0), and for chloride concentration in cats (interval = 1.0). Most of the bias was within 95% CI, and the total error range (0.8–3.5%) was acceptable according to ASVCP guidelines. Conclusion: The i-Smart 30 VET Electrolyte Analyzer provides precise and accurate measurements of sodium, potassium, and chloride concentrations in whole blood samples from dogs, cats, cattle, and pigs.
Introduction Point-of-care (POC) testing has been developed over several decades, enabling clinicians to initiate appropriate treatment in both human and animal patients.1–6 Use of POC tests is beneficial because the results are quickly obtained, which is especially important in emergency situations. In addition, the modest blood volume requirements reduce patients’ blood loss related to blood sampling. However, POC instruments designed for human specimens may be inaccurate with animal blood and may require minimal technical skills from operators. Therefore, it is important to validate and compare the performance of newly
410
developed POC analyzers to the gold standard analytic methods.7–9 The purpose of this study was to assess the precision and accuracy of the i-Smart 30 VET Electrolyte Analyzer in comparison to a Roche reference instrument, and to determine if the 2 analyzers can be used interchangeably.
Materials and Methods Instruments The reference instrument was an ion-selective electrode analyzer (ISEA) Roche 9180 (Roche Diag-
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Table 1. Reportable ranges of Roche 9180 and i-Smart 30 VET electrolyte analyzers in whole blood, serum, plasma, and aqueous solutions. Analyte Sodium (mmol/L) Potassium (mmol/L) Chloride (mmol/L)
Roche 9180
i-Smart 30 VET
40–205 1.5–15 50–200
20–250 0.5–20.0 20–250
nostics, Indianapolis, IN, USA) operating with 5 different ion-selective electrodes (ISE) and requiring 95 lL sample volume. It is fully automated, including calibration. Sodium, potassium, chloride, ionized calcium, and lithium concentrations can be measured in samples of heparinized whole blood, serum, plasma, urine, dialysate, and aqueous standard solutions. The reportable ranges of each analyte for Roche 9180 electrolyte analyzer are shown in Table 1.10 The i-Smart 30 VET Electrolyte Analyzer (i-SENS, Seoul, Korea) is a portable electrolyte analyzer developed for veterinary medicine operating with 3 ISE. Sodium, potassium, and chloride concentrations can be measured in 60 lL of heparinized whole blood, heparinized plasma, and serum samples. The reportable ranges for each analyte of i-Smart 30 VET Electrolyte Analyzer are shown in Table 1.11 For the actual sample application, there are cartridges (30–200 tests per cartridge) containing the required components for the individual analyses, including ion-specific electrochemical sensor cards, calibration solution, sampler, a waste reservoir, a valve, and tubes. The screen-printed, planar-typed micro-electrode ISE contains a polymeric sensing membrane dispensed over an internal hydrogel layer. The sensor card contains miniaturized ion-specific electrochemical sensors composed of thin, small plastic substrates. The sample is introduced to the sensor and analysis proceeds automatically. The result is obtained in 35 seconds and displayed on the LCD screens, and may be printed out. Cleaning and one-point calibration are performed after each test. Both the Roche and the i-Smart analyzer compare the ion activity of an unknown electrolyte concentration to a reference solution. An ion-selective membrane reacts with the specific electrolyte, and the difference between the membrane potential and reference electrode are calculated. The ion concentration of the sample is determined from the measured electrical potential (voltage) using the Nernst equation by direct potentiometry.12,13 The default reference intervals are provided in both electrolyte analyzers and it is possible to adapt laboratory-specific normal ranges by users. In addition, the correction factors can be applied through their
own programming menu. Users can set correction factors for both proportional (slope) and constant (intercept) errors.10,11 The i-Smart 30 VET Electrolyte Analyzer can store reference intervals and correction factors for up to 6 different species.
Sample collection This study was approved by the Institutional Animal Care and Use Committee (IACUC, Approval No. KU12022, April 15, 2012) of the University of Konkuk. A total of 400 heparinized whole blood samples—100 each from dogs, cats, cattle, and pigs—were tested for performance evaluation of the i-Smart 30 VET Electrolyte Analyzer. Samples were collected in lithiumheparin tubes (BD Vacutainer; Becton Dickinson, Franklin Lakes, NJ, USA) irrespective of sex, age, breed, and health status, from dogs and cats presented to the teaching hospital, College of Veterinary Medicine, Konkuk University, and from cattle and pigs during visits at local farms, with their owners’ consent. The samples were kept on ice or refrigerated at 4°C during transport. Samples were transported to the laboratory within 2 hours of collection and analyzed within another 2 hours.
Sample analysis and determination of accuracy, precision, correlation, and bias Analyses were performed by trained veterinarians under standardized conditions. Only nonhemolytic, nonclotted, and nonlipemic whole blood samples were analyzed. The experimental materials, including cartridges and i-Smart Electrolyte QC multilevel control reagents (i-SENS, Seoul, Korea), were used within quality specifications up to their indicated expiration dates. The 500 lL samples of heparinized whole blood were analyzed in duplicates and on the same day at about the same time on both analyzers for sodium, potassium, and chloride concentrations. A one-point calibration was performed prior to each measurement, and a 2-point calibration was performed every 4 and 6 hours for the Roche 9180 and i-Smart 30 VET, respectively, ensuring precision and consistency. To monitor performance, a daily control run was performed with known sodium, potassium, and chloride concentrations at 3 levels (low, medium, and high). Three levels of quality control materials (i-Smart Electrolyte QC multi-levels) for each analyte were used to determine within-run and between-day imprecision. Two replicates of each sample were measured in 2 runs per day for 20 consecutive days.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
411
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Kim et al
Statistical analyses Means, standard deviation (SD), and coefficient of variation (CV) were calculated from repeat measurements. Within-run (SDwr), within-day (SDwd), day-to-day (SDdd), between-run (SDbr), and between-day (SDbd) precision were calculated.14 Precision data were calculated using the following equations: Within-Run Precision : SDwr vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uP 2 uT P u ðX Xij2 Þ2 ti¼1 j¼1 ij1 ¼ 4T Within-Day Precision : SDwd vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uT u P mean mean 2 u ðXi1 Xi2 Þ ti¼1 ¼ 2T Day-to-Day Precision : SDdd vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi uT u P mean u ðXi X mean Þ2 ti¼1 ¼ T 1 Between-run Precision : SDbr qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ SD2wd SD2wr =2
(inaccuracy) in analytic measurement. The total error was calculated using the following equation: TEo % ¼ 2CV þ Bias% The calculated total error was compared to the total allowable error (TEa) according to the ASVCP guidelines for allowable total error.15 To compare the i-Smart 30 VET Electrolyte Analyzer to the Roche analyzer, the Kolmogorov–Smirnov test was used to verify the assumption of normality. The correlation of sodium, potassium, and chloride concentrations between analyzers was determined by Pearson correlation coefficients. Correlations were characterized as excellent (r ≥ 0.95), very good (r = 0.90–0.94), good (r = 0.80–0.89), fair (r = 0.59–0.79), or poor (r < 0.59).16,17 Proportional and constant error between the 2 analyzers were determined using the Passing–Bablok regression analysis.18 A 95% confidence interval (CI) for slope and intercept that did not include a value of 1.0 and zero, respectively, indicated significant proportional and constant error. Bland–Altman plots were constructed to determine mean bias and the SD of the bias.19 TEo was calculated using the bias from the analyzer comparison and the total CV of the precision study according to the equation Bias% ¼
Between-day Precision : SDbd qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ SD2dd SD2wd =2 where T is the total number of days; j the run number mean within-day (1 or 2); i the day; Xi1 the average of the mean the average of all results day i; 2 replicates day i; Xi Xmean the average of all results. The total precision was calculated using the following equation: qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi SDTotal¼ SD2betweendayþSD2betweenrunþSD2withinrun
Bias Meancomparison study 100 Meanspeciesspecific from comparison study
TEo % ¼ 2 total CV (from precision study) þ Bias% ðfrom method comparisonÞ Comparability was assessed by comparing observed total error to the ASVCP’s TEa, which is 5% for all electrolytes, except for low concentrations of potassium, for which TEa is 10%.20 The study design and statistical analysis are described as a flowchart in Figure 1.
The bias for each analyte was also calculated to assess systemic analytical errors in comparison with the manufacturer’s provided target values.15 The bias (%) and total CV were calculated as follows: Bias% ¼
Biasmanufacturer Meanmeasured 100 Meanmanufacturer CV% ¼
SD 100 Mean
Observed total error (TEo) was calculated according to the inaccuracy (bias) and imprecision (CV) of the overall error that occurred as a combined effect of random error (imprecision) and systemic error
412
Figure 1. Flowchart illustrating the study design for performance evaluation of the i-Smart 30 VET Electrolyte Analyzer.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Table 2. Precision of the i-Smart 30 VET Electrolyte Analyzer for 3 levels of i-Smart Electrolyte quality control samples. Manufacturer’s Target Ranges Analyte Na
Within-Run
Between-Day
Total
Level
N
Mean
Mean (mmol/L)
SD (mmol/L)
CV (%)
SD (mmol/L)
CV (%)
SD (mmol/L)
CV (%)
TE (%)
1 2 3 1 2 3 1 2 3
40 40 40 40 40 40 40 40 40
111–121 (116) 132–142 (137) 152–162 (157) 1.5–2.5 (1.9) 3.5–4.5 (4.0) 5.4–6.4 (5.9) 74–84 (79) 96–106 (101) 121–131 (126)
116.3 136.9 156.1 1.90 3.99 5.91 79.2 101.1 126.4
0.63 0.43 0.44 0.025 0.030 0.050 0.49 0.27 0.32
0.5 0.3 0.3 1.3 0.7 0.8 0.6 0.3 0.3
0.09 0.26 1.48 0.018 0.045 0.040 0.26 0.04 0.89
0.1 0.2 1.0 1.0 1.1 0.7 0.3 0.0 0.7
0.76 0.50 1.56 0.033 0.053 0.064 0.59 0.32 1.05
0.7 0.4 1.0 1.7 1.3 1.1 0.7 0.3 0.8
1.6 0.8 2.6 3.7 2.8 2.4 1.8 0.7 2.0
+
K+
Cl
N indicates the number of runs; CV, coefficient of variation; TE, total error.
Analyses were performed using SPSS version 19.0 (SPSS Inc., Chicago, IL, USA) and Analyse-it version 3.60, Method Validation Edition (Analyse-it Software, Ltd., Leeds, UK), including CV, the Kolmogorov–Smirnov test, Pearson correlation analysis, Passing–Bablok regression, and Bland–Altman plots. For all tests, a P < .001 was considered statistically significant.
Results Precision and accuracy The precision of the i-Smart 30 VET for sodium, potassium, and chloride concentrations are shown in Table 2. The CV for within-run and between-day assays were between 0.1% and 1.3% for level 1, between 0.0 and 1.1% for level 2, and between 0.3% and 1.0% for level 3 controls. Total imprecision and
TEo for all analytes based on quality control materials were < 2% and < 4%, respectively.
Instrument comparison The results obtained using the i-Smart 30 VET Electrolyte Analyzer were compared to the Roche 9180 on the basis of the Pearson correlation coefficients, Passing– Bablok regression analysis, and Bland–Altman difference plots (Table 3, Figures 2–5). The correlation between i-Smart 30 VET Electrolyte Analyzer and the Roche 9180 was excellent for sodium, potassium, and chloride concentrations in canine, feline, bovine, and porcine blood (r ≥ .982, P < .0001). Results of Passing–Bablok regression showed that the slopes for all 3 electrolytes were ranging from 0.92–1.00. The 95% CI of the slopes for all 3 electrolytes included 1.00 indicating that there was
Table 3. Pearson correlation coefficient, Passing–Bablok regression analysis, and Bland–Altman plots data for i-Smart 30 VET and Roche 9180 electrolyte analyzer comparison
Species
Analyte
Range (mmol/L)
Pearson Coefficiency*
Passing–Bablok’s Slope (95% CI)
Passing–Bablok’s Intercept (95% CI)
Bias mean † (95% CI)
Bias SD†
Bias %
TE Observed
Canine
Na+ K+ ClNa+ K+ Cl Na+ K+ Cl Na+ K+ Cl
118–175 2.6–6.7 87–130 126–164 2.4–6.5 94–141 114–159 2.4–7.0 79–112 122–168 3.1–7.4 82–125
0.992 0.996 0.987 0.991 0.997 0.994 0.996 0.997 0.984 0.992 0.983 0.982
1.00 (1.00–1.00) 1.00 (0.96–1.00) 1.00 (0.92–1.00) 1.00 (1.00–1.05) 1.00 (1.00–1.00) 1.00 (0.94–1.00) 1.00 (1.00–1.00) 1.00 (0.96–1.00) 0.94 (0.87–1.00) 1.00 (1.00–1.00) 0.96 (0.89–1.00) 0.92 (0.86–1.00)
0.5 (0.5–0.5) 0.0 (0.0–0.2) 0.0 (0.0–8.8) 2.0 (4.9–2.0) 0.0 (0.0–0.0) 1.0 (1.0–7.9) 3.0 (3.0–3.0) 0.0 (0.0–0.2) 3.3 (3.0–10.5) 4.0 (4.0–4.0) 0.0 (0.2–0.4) 7.3 (1.0–13.3)
0.4 (2.1–2.9) 0.0 (0.3–0.2) 0.3 (2.6–3.2) 1.7 (0.7–4.1) 0.1 (0.2–0.1) 0.6 (1.5–2.6) 3.4 (1.9–4.8) 0.0 (0.2–0.2) 2.6 (5.4–0.2) 3.6 (1.6–5.7) 0.1 (0.5–0.2) 1.4 (4.2–1.5)
1.26 0.12 1.47 1.20 0.08 1.03 0.72 0.09 1.40 1.02 0.18 1.44
0.3 0.0 0.3 1.1 2.2 0.5 2.5 0.0 2.6 2.6 2.1 1.4
1.1 0.8 1.1 2.0 3.0 1.3 3.3 0.8 3.5 3.4 3.0 2.2
Feline
Bovine
Porcine
CI indicates confidence interval. *P < .0001. † Mean and SD value of Bias from Bland–Altman plot.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
413
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Kim et al
Figure 2. Passing–Bablok regression analyses (left) and Bland–Altman difference plots (right) for the i-Smart 30 VET and the Roche 9180 electrolyte analyzer for sodium, potassium, and chloride concentrations in canine blood (n = 100). In the regression plots, the black dotted line indicates the line of identity (y = x), and the red line is the line of best fit. In the difference plots, the black dashed horizontal line represents the mean bias, and the red dotted horizontal line represents the 95% limits of agreement.
no proportional bias between the 2 analyzers for any tested species. The intercept of the regression included 0.0 within 95% CI except for sodium concentration in dogs (95% CI = 0.5–0.5), cattle (95% CI = 3.0–3.0), and pigs (95% CI = 4.0–4.0) as well as
414
chloride concentration in cats (95% CI = 1.0–7.9). These results are indicative of a constant error in sodium concentrations in dogs, cattle, and pigs, and chloride concentrations in cats when measured on the i-Smart 30 VET.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Figure 3. The Passing–Bablok regression analyses (left) and Bland–Altman difference plots (right) for the i-Smart 30 VET and Roche 9180 electrolyte analyzers for sodium, potassium, and chloride in feline blood (n = 100). In the regression plots, the black dotted line indicates the line of identity (y = x), and the red line is the line of best fit. In the difference plots, the black dashed horizontal line represents the mean bias, and the red dotted horizontal line represents the 95% limits of agreement.
The ranges of the mean differences in the Bland– Altman plots in all 4 species were 0.4–3.6 mmol/L (SD ranges = 0.72–1.26) for sodium; 0.1–0.0 mmol/L (SD ranges = 0.08–0.12) for potassium; and 2.6– 0.6 mmol/L (SD ranges = 1.03–1.47) for chloride. The Bland–Altman plots showed that with the exception of
a few values most results were within the 95% CI. In all 4 species, bias and TEo were < 3% (range 0.0–2.6%) and < 5% (range 0.8–3.5%), respectively, for sodium, potassium, and chloride concentrations measured by the i-Smart 30 VET when compared to the Roche 9180.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
415
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Kim et al
Figure 4. The Passing–Bablok regression analyses (left) and Bland–Altman difference plots (right) for the i-Smart 30 VET and Roche 9180 electrolyte analyzers for sodium, potassium, and chloride in bovine blood (n = 100). In the regression plots, the black dotted line indicates the line of identity (y = x), and the red line is the line of best fit. In the difference plots, the black dashed horizontal line represents the mean bias, and the red dotted horizontal line represents the 95% limits of agreement.
Discussion POC analyzers are powerful and life-saving tools used by clinicians for critical care and emergency cases in both human and veterinary medicine.2,3 The ASVCP defines ‘accuracy’ as the closeness of agreement
416
between the results of a measurement and the true concentration of an analyte, and ‘comparability’ as the closeness of results from 2 or more instruments that process the same specimen, ensuring that results from different instruments are similar enough to be used interchangeably without clinical error.15
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Figure 5. The Passing–Bablok regression analyses (left) and Bland–Altman difference plots (right) for the i-Smart 30 VET and Roche 9180 electrolyte analyzers for sodium, potassium, and chloride in porcine blood (n = 100). In the regression plots, the black dotted line indicates the line of identity (y = x), and the red line is the line of best fit. In the difference plots, the black dashed horizontal line represents the mean bias, and the red dotted horizontal line represents the 95% limits of agreement.
Our study shows that the i-Smart 30 VET yielded acceptable CV for the within-run and between-day imprecisions for sodium, potassium, and chloride in quality control materials. The total imprecision CV and TEo were within quality specification that fulfill required
criteria of the ASVCP guidelines for total allowable error (ie, TEa < 5%).15 Based on the results of the method comparison assay, the sodium, potassium, and chloride concentrations in blood samples from dogs, cats, cattle, and pigs
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
417
Evaluation of i-Smart 30 VET Electrolyte Analyzer
Kim et al
measured by the i-Smart 30 VET were comparable with the results from a widely used laboratory analyzer, the Roche 9180. The correlation between the results produced by the i-Smart 30 VET and Roche 9180 was excellent (r > .98), and most of the bias was within the 95% CI. There was agreement between the 2 methods for most measurements, except sodium concentration in dogs, cattle, and pigs, and chloride concentration in cats. Mean differences between the iSmart 30 VET and Roche 9180 were near zero, except for slightly higher sodium concentrations in samples from cats, pigs, and cattle, and chloride concentrations in samples from pigs and cattle. Different methods used by the 2 instruments may explain some of the discrepancies. Both Roche 9180 and i-Smart 30 VET analyze ion concentration by potentiometric measurement using ISE; however, the Roche 9180 uses a conventional macroelectrode. These macro-electrodes must be replaced every 6 or 12 months by trained engineers, resulting in relatively high maintenance cost. In contrast, the i-Smart 30 VET uses a screen-printed planar ISE.10,11 Micro-electrodes with a diameter < 25 lM have several advantages such as small size, minimization of solution resistance effects, and rapid response times.12,13 However, the electrochemical behavior of a micro-electrode is slightly different from the macroelectrode due to diffusion from the edges of the electrodes.21,22 As other POC analyzers, the i-Smart 30 VET Electrolyte Analyzer has several advantages. It operates with an all-in-one cartridge, is easy to operate, and produces results quickly. Moreover, the module is relatively small and requires a relatively small blood volume for analysis. Another advantage of this analyzer is that it is easy to remove potential blood clots simply by replacing the cartridge. In conclusion, the i-Smart 30 VET Electrolyte Analyzer provided acceptable results in comparison with the Roche 9180 for sodium, potassium, and chloride concentrations in blood from dogs, cats, cattle, and pigs. Disclosure: The authors have indicated that they have no affiliations or financial involvement with any organization or entity with a financial interest in, or in financial competition with, the subject matter or materials discussed in this article.
References 1. Plebani M. Does POCT reduce the risk of error in laboratory testing? Clin Chim Acta. 2009;404:59–64.
418
2. Kazmierczak SC. Point-of-care testing quality: some positives but also some negatives. Clin Chem. 2011;57:1219–1220. 3. Lee EJ, Shin SD, Song KJ, et al. A point-of-care chemistry test for reduction of turnaround and clinical decision time. Am J Emerg Med. 2011;29:489–495. 4. Wong DM, Gigu ere S, Wendel MA. Evaluation of a point-of-care portable analyzer for measurement of plasma immunoglobulin G, total protein, and albumin concentrations in ill neonatal foals. J Am Vet Med Assoc. 2013;242:812–819. 5. Dixon-Jimenez AC, Brainard BM, Cathcart CJ, Koenig A. Evaluation of a point-of-care coagulation analyzer (Abaxis VSPro) for identification of coagulopathies in dogs. J Vet Emerg Crit Care (San Antonio). 2013;23:402–407. 6. Montesinos A, Ardiaca M. Acid-base status in the avian patient using a portable point-of-care analyzer. Vet Clin North Am Exot Anim Pract. 2013;16:47–69. 7. O’Kane MJ, McManus P, McGowan N, Lynch PL. Quality Error Rates in Point-of-Care Testing. Clin Chem. 2011;57:1267–1271. 8. Beneteau-Burnat B, Pernet P, Pilon A, et al. Evaluation of the GEM Premier 4000: a compact blood gas CO-Oximeter and electrolyte analyzer for point-of-care and laboratory testing. Clin Chem Lab Med. 2008;46:271–279. 9. Smolcic VS, Bilic-Zulle L, Fisic E. Validation of methods performance for routine biochemistry analytes at Cobas 6000 analyzer series module c501. Biochem Med. 2011;21:182–190. 10. Roche Diagnostics. Instructions for Use 9180 Electrolyte Analyzer V5.0. Mannheim, Germany: Roche Diagnostics GmbH; 2008. 11. i-SENS. i-Smart 30 VET electrolyte analyzer Operator’s Manual. Seoul, Republic of Korea: Medical Technology Promedt Consulting GmbH; 2013. 12. Yan S. Electrochemical cells – New Advances in Fundamental Researches and Applications. IA: InTech; 2012:1–15. 13. Yuzuru I, Masao M. Electrochemical Measurements with Interdigitated Array Microelectrodes. Curr Sep. 1995;14:1. 14. Clinical and Laboratory Standards Institute (CLSI). EP10-A2: Preliminary Evaluation of Quantitative Clinical Laboratory Methods; Approved Guideline. 2nd ed. Wayne, PA: CLSI; 2002:41. 15. Harr KE, Flatland B, Nabity MB, Freeman KP. ASVCP guidelines: allowable total error guidelines for biochemistry. Vet Clin Pathol. 2013;42:424–436. 16. Riond B, Weissenbacher S, Hofmann-Lehmann R, Lutz H. Performance evaluation of the Sysmex pocH-100iV Diff hematology analyzer for analysis of canine, feline, equine, and bovine blood. Vet Clin Pathol. 2011;40: 484–495.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
Kim et al
17. Welles EG, Hall AS, Carpenter DM. Canine complete blood counts: a comparison of four in-office instruments with the ADVIA 120 and manual differential counts. Vet Clin Pathol. 2009;38:20–29. 18. Bablok W, Passing H. Application of statistical procedures in analytical instrument testing. J Autom Chem. 1985;7:74–79. 19. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1:307–310.
Evaluation of i-Smart 30 VET Electrolyte Analyzer
20. Gunn-Christie RG, Flatland B, Friedrichs KR, et al. ASVCP quality assurance guidelines: control of preanalytical, analytical, and postanalytical factors for urinalysis, cytology, and clinical chemistry in veterinary laboratories. Vet Clin Pathol. 2012;41:18–26. 21. Christopher MAB, Ana MOB. Electrochemistry: Principles, Methods, and Applications. IA: Oxford University Press; 1993:82–90. 22. Joseph W. Analytical Electrochemistry. IA: Wiley-VCH; 2000:140–161.
Vet Clin Pathol 44/3 (2015) 410–419 ©2015 American Society for Veterinary Clinical Pathology
419
