The quantifications of the concentrations of Na+, K+, iCa2+, and tCO, in blood, in addition to the determinations of pH, Pco2, Po2, BE, and Hct, are useful as auxiliary clinical findings for the diagnosis of several diseases in domestic animals. When diagnosed early, alterations in hydration, electrolyte concentrations, and acid-base status allow for fast and adequate clinical intervention and increase the chance of animal survival. Evaluation of tCO, concentration, pH, Pco2, and BE must be performed soon after blood collection. This is not possible in some animals because clinical evaluation is conducted on farms located far away from the laboratories that perform these analyses.
The development of tools for on-farm blood analysis, also known as point-of-care devices, has resulted in many systems that have become fairly commonplace in emergency and critical care settings in both human and veterinary hospitals.1–13 Overall, point-of-care devices have improved patient care by facilitating the stabilization of critically ill patients and assisting clinicians in choosing more specific diagnostic procedures and treatment.9–11,14 In addition to ease of portability and rapid return of results, point-of-care devices are simple to operate and user-friendly for personnel who lack formal training in laboratory techniques.12
Portable clinical analyzers use self-contained cartridges that are programmed to simultaneously analyze combinations of biochemical, blood gas, and hematologic variables.1,2,12,15 Recently, veterinary practitioners have begun to use PCAs to analyze blood samples collected from birds and exotic and laboratory animals; these samples are typically limited to small quantities.12,14 Therefore, PCAs help to eliminate the need for additional sample processing and permit adequate sampling without the need for euthanasia.12 Furthermore, the results from a PCA have been validated for some animal species (ie, cats,16 chickens,14 dogs,7,11,16,17 horses,7,10,16 rodents,12 seals,9 and viscachas18).
Surprisingly, no data are available to evaluate the performance of a PCA in measuring blood gas partial pressures and electrolyte concentrations in cattle and sheep, whereas there are 4 reports7,10,16,19 that include this information for horses. Notably, neither Hb concentration nor So2 variables have been evaluated in horses. The purpose of the study reported here was to compare the results of blood electrolyte concentrations, blood gas partial pressures, and Hct obtained by use of a PCA with those obtained by use of a conventional clinical analyzer (ie, an RA) for blood samples collected from cattle, horses, and sheep.
Materials and Methods
Animals—Blood samples were collected from 24 cattle, 22 horses, and 22 sheep. All animals were considered clinically normal on the basis of results of physical examination and hemogram analysis. The experimental protocol was approved by the Universidade Estadual Paulista-Araçatuba Animal Care and Use Committee.
Sample collection and analyses—Briefly, blood samples from all 3 species were collected from the jugular vein into 1-mL lithium-heparinized plastic syringes containing 50 U of lyophilized calcium-balanced lithium heparina and were analyzed by use of a PCAb with disposable cartridges.c The whole blood sample was introduced into the cartridgec via the lithium-heparinized syringe,a and a portion of the same blood sample was analyzed by use of a conventional RA immediately thereafter.d Analytic variables recorded from both the PCA and RA during analysis included pH, Pco2, Po2, tCO2, So2, HCO3−, BE, Na+, K+, iCa2+, Hct, and Hb. The analyzers were located immediately adjacent to each other. All samples were analyzed in both analyzers within 30 minutes after blood collection. The PCA and the RA were operated by trained personnel (one of the authors [JRP] as well as a laboratory technician).
For quality control of the output results, PCA cartridges were stored in a refrigerator (3° to 8°C) and cartridges were only used prior to the expiration date provided by the manufacturer. The temperature of each cartridge was allowed to equilibrate with the ambient temperature of the room (20°C) before the cartridge pouch was opened. As described elsewhere,19 appropriate measures were taken to avoid touching the contact pads, which could have interfered with data transmission, or exerting pressure over the center of the cartridge, which could have caused premature release of the calibrating solution. The PCA control solutions (aqueous) were used to assess 1 cartridge out of each cartridge batch as recommended by the manufacturer.20
The RAd used was a fully automatic microprocessor analyzer with a fluid calibration system that eliminated the use of expensive calibration gases. The RA was calibrated daily in accordance with the manufacturer's recommendations.21 Equations and derivations of correction factors used by the RA for determination of electrolyte concentrations, pH, and blood gas partial pressures were performed as described elsewhere.7
Statistical analysis—Data were analyzed via a paired Student t test for comparison between mean values of data reported from both analyzers. Deming linear regression and Pearson correlation were determined for all variables obtained from both analyzers. A value of P < 0.05 was considered significant. Data were also analyzed via the Bland-Altman method, in which the difference between 2 measurements obtained by different methods are plotted against their mean.22,23 Visual inspection was used to evaluate agreement after plotting 1 horizontal line that represented bias (mean of the difference), 2 horizontal lines that represented the limits of agreement (mean difference ± 2 SD), and 2 horizontal lines that represented the 95% CI for bias (mean difference ± 2 SEM; lines not shown).23 No bias was indicated when the 95% CI for bias included zero.13,22–25 Agreement between the results of the analyzers was considered good when the bias (mean difference) was small, the 95% CI for bias was narrow, and there were no outliers (ie, no value exceeded the limits of the 95% CI).13 On the basis of application of an objective classification system,13,25 correlation was characterized as poor (r < 0.59), fair (r = 0.59 to 0.79), good (r = 0.80 to 0.92), or excellent (r ≥ 0.93). The data collected during the study reported here appeared to be normally distributed on the basis of review of the reported values for all variables measured and analysis of the Bland-Altman plots.
Results
Cattle—Mean values of the results for analysis of bovine blood samples obtained with the PCA were significantly less than values obtained with the RA for Po2 (P < 0.001), So2 (P < 0.001), Hct (P < 0.001), and concentrations of K+ (P = 0.001) and Hb (P = 0.002; Table 1). Mean values of the results for pH and concentrations of tCO2, HCO3−, BE, and Na+ obtained with the PCA were significantly (P < 0.001) greater than those values obtained with the RA. A significant difference was not detected between results of the analyzers for Pco2 (P = 0.056) and iCa2+ concentration (P = 0.220). Analysis of the variables measured by use of the analyzers revealed an excellent correlation between the results for pH, Pco2, BE, and concentrations of tCO2, Na+, K+, and iCa2+ and a good correlation between the results for Po2, So2, Hct, and concentrations of HCO3− and Hb.
Results for variables evaluated in venous blood samples collected from 24 cattle and analyzed via a PCA and an RA.
| Variable | PCA* | RA* | Mean difference*† | 95% CI | Deming regression‡ | Pearson correlation§ |
|---|---|---|---|---|---|---|
| pH | 7.42 ± 0.04a | 7.40 ± 0.04b | −0.022 ± 0.010 | −0.04 to 0.00 | y = 1.174x − 1.265 | 0.98 |
| Pco2 (mm Hg) | 46.59 ± 4.75 | 47.00 ± 4.81 | 0.413 ± 1.002 | −1.55 to 2.38 | y = 0.9878x + 0.1627 | 0.98 |
| Po2 (mm Hg) | 30.79 ± 3.46a | 34.51 ± 3.31b | 3.721 ± 1.779 | 0.23 to 7.21 | y = 1.056x − 5.642 | 0.86 |
| tCO2 (mmol/L) | 31.58 ± 2.55a | 29.72 ± 2.13b | −1.867 ± 0.859 | −3.55 to −0.18 | y = 1.209x − 4.354 | 0.95 |
| So2 (%) | 59.00 ± 8.33a | 66.04 ± 7.33b | 7.042 ± 3.791 | −0.39 to 14.47 | y = 1.154x − 17.24 | 0.89 |
| HCO3− (mmol/L) | 30.38 ± 2.60a | 28.28 ± 2.06b | −2.096 ± 1.133 | −4.32 to 0.12 | y = 1.291x − 6.120 | 0.91 |
| BE (mmol/L) | 5.67 ± 2.93a | 2.81 ± 2.06b | −2.858 ± 1.116 | −5.05 to −0.67 | y = 1.442x + 1.617 | 0.96 |
| Na+ (mmol/L) | 141.58 ± 4.73a | 138.06 ± 6.57b | −3.525 ± 2.171 | −7.78 to 0.73 | y = 0.7141x + 43.00 | 0.98 |
| K+ (mmol/L) | 4.19 ± 0.33a | 4.25 ± 0.36b | 0.055 ± 0.072 | −0.09 to 0.20 | y = 0.9011x + 0.3652 | 0.98 |
| iCa2+ (mg/dL) | 1.19 ± 0.14 | 1.20 ± 0.14 | 0.013 ± 0.050 | −0.09 to 0.11 | y = 1.017x − 0.03345 | 0.94 |
| Hct (%) | 29.21 ± 5.51a | 31.18 ± 4.47b | 1.975 ± 2.270 | −2.47 to 6.42 | y = 1.256x − 9.952 | 0.92 |
| Hb (g/dL) | 10.04 ± 1.90a | 10.61 ± 1.82b | 0.571 ± 0.775 | −0.95 to 2.09 | y = 1.051x − 1.109 | 0.91 |
Data reported are the mean ± SD.
Mean difference was calculated as follows: mean difference = (mean value obtained via the PCA – the mean value obtained via the RA)/2.
In each equation, y = PCA value and x = RA value.
Correlation was characterized as excellent (r ≥ 0.93), good (r = 0.80 to 0.92), fair (r = 0.59 to 0.79), or poor (r < 0.59). The correlation coefficient for Hb concentration was significantly (P = 0.002) different from 0, and all other correlation coefficients were also significantly (P < 0.001) different from 0.
Within a row, values with different superscript letters differ significantly (P ≤ 0.05; Student t test).
The 95% CI was narrow for the measured values of pH, Pco2, and concentrations of tCO2, HCO3−, K+, Hb, and iCa2+; however, the 95% CI was wide for Po2, So2, BE, Hct, and Na+ concentration. Further analysis of the results revealed a positive bias for Po2, So2, Hct, and K+ and Hb concentrations and a negative bias for pH, BE, and concentrations of tCO2, HCO3−, and Na+. No bias was detected for Pco2 and iCa2+ concentration. No outliers were detected in the Bland-Altman plots for pH, tCO2, and BE, and only 1 outlier was detected in each plot for the remaining 9 variables. Good agreement was detected between the results obtained via the PCA and the results obtained via the RA for pH and tCO2 concentration (Figure 1).
Representative Bland-Altman plots of the results for venous blood samples collected from cattle and analyzed with a PCA and an RA. Differences between results for the PCA and RA are plotted on the y-axis, and the mean value of the measurements for both methods is plotted on the x-axis. The mean difference (ie, bias [solid line]) and 95% limits of agreement (ie, mean bias ± 2 SD [dashed lines]) are indicated. Values located outside of the 95% limits of agreement are considered outliers.
Citation: American Journal of Veterinary Research 71, 5; 10.2460/ajvr.71.5.515
Representative Bland-Altman plots of the results for venous blood samples collected from cattle and analyzed with a PCA and an RA. Differences between results for the PCA and RA are plotted on the y-axis, and the mean value of the measurements for both methods is plotted on the x-axis. The mean difference (ie, bias [solid line]) and 95% limits of agreement (ie, mean bias ± 2 SD [dashed lines]) are indicated. Values located outside of the 95% limits of agreement are considered outliers.
Citation: American Journal of Veterinary Research 71, 5; 10.2460/ajvr.71.5.515
Representative Bland-Altman plots of the results for venous blood samples collected from cattle and analyzed with a PCA and an RA. Differences between results for the PCA and RA are plotted on the y-axis, and the mean value of the measurements for both methods is plotted on the x-axis. The mean difference (ie, bias [solid line]) and 95% limits of agreement (ie, mean bias ± 2 SD [dashed lines]) are indicated. Values located outside of the 95% limits of agreement are considered outliers.
Citation: American Journal of Veterinary Research 71, 5; 10.2460/ajvr.71.5.515
Horses—Mean values of the results for analysis of equine blood samples obtained with the PCA were significantly greater than those values obtained with the RA for Pco2 (P = 0.006), BE (P < 0.001), and concentrations of tCO2 (P < 0.001), HCO3−(P < 0.001), and Na+ (P < 0.001; Table 2). Mean values of the results obtained with the PCA were significantly less than those values obtained with the RA for pH (P = 0.019), So2 (P = 0.004), and K+ concentration (P = 0.011). Mean values were not significantly different between results obtained with the analyzers for Po2 (P = 0.215), Hb (P = 0.822), and Hct (P = 0.526). The concentration of iCa2+ in equine blood was not measured. Analysis of the variables measured by use of the analyzers revealed an excellent correlation between the results for BE and concentrations of tCO2, HCO3−, and K+. A good correlation was detected between the results for pH, Pco2, Po2, and So2. A fair correlation was detected between the results for Hct and Hb concentration.
Results for variables evaluated in venous blood samples collected from 22 horses and analyzed via a PCA and an RA.
| Variable | PCA* | RA* | Mean difference*† | 95% CI | Deming regression‡ | Pearson correlation§ |
|---|---|---|---|---|---|---|
| pH | 7.41 ± 0.03a | 7.42 ± 0.03b | 0.009 ± 0.017 | −0.02 to 0.04 | y = 0.9912x + 0.05568 | 0.87 |
| Pco2 (mm Hg) | 45.14 ± 3.17a | 43.97 ± 3.64b | −1.173 ± 1.817 | −1.17 to 1.82 | y = 0.8518x + 7.688 | 0.87 |
| Po2 (mm Hg) | 38.05 ± 6.02 | 38.80 ± 5.73 | 0.759 ± 2.782 | −4.69 to 6.21 | y = 1.058x − 3.016 | 0.89 |
| tCO2 (mmol/L) | 30.14 ± 3.06a | 25.02 ± 2.32b | −5.113 ± 1.105 | −7.28 to −2.95 | y = 1.338x − 3.356 | 0.95 |
| So2 (%) | 71.05 ± 7.57a | 73.69 ± 7.72b | 2.640 ± 3.856 | −4.92 to 10.20 | y = 0.9774x − 0.9779 | 0.87 |
| HCO3− (mmol/L) | 28.86 ± 2.90a | 27.68 ± 2.46b | −1.182 ± 0.986 | −3.12 to 0.75 | y = 1.192x − 4.124 | 0.95 |
| BE (mmol/L) | 4.23 ± 3.35a | 3.00 ± 2.66b | −1.227 ± 1.117 | −3.42 to 0.96 | y = 1.274x + 0.4055 | 0.96 |
| Na+ (mmol/L) | 137.70 ± 2.19a | 136.10 ± 2.16b | −1.677 ± 1.493 | −1.68 to 1.49 | y = 1.026x − 1.931 | 0.76 |
| K+ (mmol/L) | 4.43 ± 0.68a | 4.47 ± 0.65b | 0.043 ± 0.073 | −0.10 to 0.19 | y = 1.033x − 0.1905 | 0.99 |
| iCa2+ (mg/dL) | — | — | — | — | — | — |
| Hct (%) | 34.05 ± 6.67 | 35.60 ± 13.11 | 1.555 ± 11.317 | −20.63 to 23.73 | y = 0.3129x + 22.91 | 0.51 |
| Hb (g/dL) | 11.68 ± 2.26 | 11.87 ± 4.37 | 0.19 ± 3.82 | −7.31 to 7.68 | y = 0.3089x + 8.016 | 0.49 |
Correlation coefficients for pH, Pco2, So2, and K+ concentration were significantly (P = 0.019, P = 0.006, P = 0.004, and P = 0.011, respectively) different from 0, and all other correlation coefficients were also significantly (P < 0.001) different from 0.
— = Not done.
See Table 1 for remainder of key.
The 95% CI was narrow for the measured values of pH, HCO3− concentration, BE, and K+ concentration; however, the 95% CI was wide for Po2, Pco2, So2, Hct, and concentrations of tCO2, Na+, and Hb. Further analysis of the results revealed a positive bias for pH, So2, and concentrations of K+ and Hb. A negative bias was revealed for Pco2 and BE and concentrations of tCO2, HCO3−, and Na+. Bias was not detected for Po2, Hct, and Hb concentration. Outliers were not detected in the Bland-Altman plots for BE, Na+ concentration, K+ concentration, Hct, and Hb concentration; however, 2 outliers were detected in the plot for Po2 and 1 outlier was detected in each of the 5 remaining plots. Good agreement was detected between the results obtained via the PCA and the results obtained via the RA for BE and K+ concentration.
Sheep—Mean values of the results for the analysis of ovine blood samples obtained with the PCA were significantly higher than those values obtained with the RA for BE (P = 0.002) and concentrations of tCO2 (P < 0.001), HCO− (P = 0.002), and Na+ (P < 0.001; Table 3). Mean values of the results obtained with the PCA were significantly less than those obtained with the RA for Hct (P = 0.001) and concentrations of Hb (P = 0.004) and K+ (P < 0.001). No significant difference was detected between variables measured by use of the analyzers for pH (P = 0.225), Pco2 (P = 0.421), Po2 (P = 0.220), and So2 (P = 0.835). The concentration of iCa2+ in ovine blood was not measured. Analysis of the variables measured by the analyzers revealed an excellent correlation between the results for pH and K+ concentration and a good correlation between the results for Pco2, Po2, So2, tCO2 concentration, BE, and HCO3− concentration. A fair correlation was detected between the results for Na+ concentration, Hct, and Hb concentration.
Results for variables evaluated in venous blood samples collected from 22 sheep and analyzed via a PCA and an RA.
| Variable | PCA* | RA* | Mean difference*† | 95% CI | Deming regression‡ | Pearson correlation§ |
|---|---|---|---|---|---|---|
| pH | 7.41 ± 0.04 | 7.40 ± 0.04 | −0.004 ± 0.014 | −0.03 to 0.02 | y = 1.177x − 1.307 | 0.96 |
| Pco2 (mm Hg) | 40.64 ± 4.53 | 40.23 ± 5.35 | −0.409 ± 2.338 | −4.99 to 4.17 | y = 0.8310x + 7.209 | 0.90 |
| Po2 (mm Hg) | 36.55 ± 4.27 | 37.29 ± 4.60 | 0.745 ± 2.764 | −4.67 to 6.16 | y = 0.9141x + 2.457 | 0.81 |
| tCO2 (mmol/L) | 26.68 ± 2.10a | 22.18 ± 1.68b | −4.505 ± 1.222 | −6.90 to −2.11 | y = 1.313x − 2.429 | 0.81 |
| So2 (%) | 69.45 ± 7.40a | 69.60 ± 7.61b | 0.145 ± 3.235 | −6.19 to 6.49 | y = 0.9693x + 1.988 | 0.91 |
| HCO3− (mmol/L) | 25.50 ± 1.97a | 24.63 ± 1.81b | −0.868 ± 1.119 | −3.06 to 1.32 | y = 1.107x − 1.762 | 0.83 |
| BE (mmol/L) | 0.82 ± 2.30a | −0.08 ± 1.60b | −0.895 ± 1.148 | −3.14 to 1.35 | y = 1.508x + 0.9347 | 0.89 |
| Na+ (mmol/L) | 147.0 ± 1.23a | 145.0 ± 1.91b | −1.991 ± 1.390 | −4.72 to 0.73 | y = 0.5400x + 68.69 | 0.69 |
| K+ (mmol/L) | 4.74 ± 0.44a | 4.84 ± 0.46b | 0.095 ± 0.057 | −0.02 to 0.21 | y = 0.9772x + 0.01585 | 0.99 |
| iCa2+ (mg/dL) | — | — | — | — | — | — |
| Hct (%) | 24.68 ± 5.70a | 28.39 ± 4.42b | 3.705 ± 3.663 | −3.38 to 10.88 | y = 1.392x − 14.82 | 0.77 |
| Hb (g/dL) | 8.50 ± 2.13a | 9.46 ± 1.47b | 0.96 ± 1.39 | −1.75 to 3.68 | y = 1.613x − 6.764 | 0.76 |
The 95% CI was narrow for the measured values of pH, BE, and concentrations of HCO3− and K+, whereas the CI was wide for Pco2, Po2, So2, Hct, and concentrations of tCO2, Na+, and Hb. Further analysis of the results detected a positive bias for K+ concentration, Hct, and Hb concentration, and a negative bias was detected for BE and concentrations of tCO2, HCO3−, and Na+. No bias was detected among results for pH, Pco2, Po2, and So2. Outliers were not detected in the Bland-Altman plots for Na+ concentration and K+ concentration; however, 2 outliers were detected in each plot for Hct and Hb concentration and only 1 outlier was detected in each of the remaining 7 plots. Good agreement was detected between the results obtained via the PCA and results obtained via the RA for K+ concentration.
Reference intervals—Blood samples from 24 cattle, 22 horses, and 22 sheep were analyzed. Reference intervals were calculated from the reported results of the variables analyzed (Table 4).
Reference intervals* for variables calculated from the results of venous blood samples collected from 24 cattle, 22 horses, and 22 sheep and analyzed via a PCA.
| Variable | Cattle | Horses | Sheep |
|---|---|---|---|
| pH | 7.3 to 7.5 | 7.33 to 7.46 | 7.3 to 7.5 |
| Pco2 (mm Hg) | 40.5 to 61.6 | 39.2 to 54.0 | 34.3 to 51.0 |
| Po2 (mm Hg) | 24.0 to 37.0 | 31.0 to 59.0 | 32.0 to 49.0 |
| tCO2 (mmol/L) | 26.0 to 35.0 | 24.0 to 36.0 | 24.0 to 31.0 |
| So2 (%) | 40.0 to 73.0 | 56.0 to 91.0 | 59.0 to 86.0 |
| HCO3− (mmol/L) | 25.0 to 36.0 | 23.0 to 34.0 | 23.0 to 29.0 |
| BE (mmol/L) | −1.0 to 10.0 | −3.0 to 10.0 | −3.0 to 5.0 |
| Na+ (mmol/L) | 137.0 to 162.0 | 133.0 to 142.0 | 145.0 to 149.0 |
| K+ (mmol/L) | 3.5 to 4.9 | 2.7 to 5.9 | 4.0 to 5.6 |
| iCa2+ (mg/dL) | 0.8 to 1.4 | — | — |
| Hct (%) | 20.0 to 38.0 | 17.0 to 43.0 | 13.0 to 32.0 |
| Hb (g/dL) | 7.0 to 13.0 | 6.0 to 15.0 | 4.0 to 11.0 |
Values reported are minimum and maximum values.
See Table 2 for remainder of key.
Discussion
Correlation ranged from good to excellent for most variables in the study reported here, except for Na+ concentration and Hct in horses and sheep. The findings in horses were different from those reported in another study.16 Although popular, the use of correlation to compare results between mean analyses has been criticized.22 High correlation coefficients can result from data that are in poor agreement with the reference values.7,11,13–16,19,22,23 Contrary to correlation analysis, Bland-Altman plots provide additional clinically relevant information when the limits of agreement (95% CI) are inspected. The limits of agreement are a representation of precision and the repeatability of the difference between the values for a single sample.23,26
For pH, the value obtained with the PCA was typically 0.02 U higher or 0.01 U lower than the value obtained with the RA in cattle and horses, respectively, with bias not detected in sheep. These differences in cattle and horses were observed in Bland-Altman plots, in which zero was not included in the range designated by the mean difference ± 2 SEM (also described as the 95% CI for the bias). Different technologies are used for determining pH by use of the PCA and RA; the PCA has sensors that are microfabricated thin-film electrodes, whereas the RA has a glass electrode with a pH-sensitive glass capillary tube.23 Similar to observations made in other studies,15,16,19,23 we judged the differences in pH results to be small and clinically irrelevant because venous pH values were within reference ranges for the 3 species in the study reported here.
The PCA measurements of Pco2 and Po2, in comparison with results for other in-house analyzers, are accurate in dogs11 and humans.15 Results of the study reported here indicated that Pco2 in cattle and sheep was accurately measured via the PCA. Although a significant difference was detected between the analyzers for Pco2 in horses, good agreement between results of the analyzers was evident in the Bland-Altman plot. In addition, Pco2 results were within the reference range in horses and sheep; however, Pco2 was mildly increased in cattle.16 Even though Po2 had “been overestimated by the PCA in cattle, Po2 results obtained with the PCA were significantly lower than results obtained with the RA, whereas good agreement between results was observed in horses and sheep.
In general, when compared with the results of the RA, accuracy of the calculated values (So2, BE, and concentrations of HCO3−, tCO2, and Hb) of the PCA varied from good to excellent depending on the species.11,23 The PCA underestimated So2 in cattle and horses and overestimated tCO2 concentration, HCO3− concentration, and BE in all 3 species when compared with results for the RA. These values were within the reported27 reference range for each species; additionally, similar results were reported in dogs,11 horses,10 and rats.23
The accurate measurement of iCa2+ concentration is essential for the investigation of diseases that alter calcium homeostasis and for clinical monitoring of these diseases during treatment.13 In the study reported here, iCa2+ concentrations obtained with the PCA were similar to those values obtained with the RA in cattle. Also, analysis of the Bland-Altman plot did not reveal a significant difference in results for iCa2+ concentration between analyzers; furthermore, this indicated that the PCA results for iCa2+ concentration in cattle appeared to be reliable and consistent. However, iCa2+ concentrations obtained with both analyzers for cattle were markedly below the reference range.27 In another study,16 iCa2+ concentrations were more variable in horses than in dogs. In the study reported here, it was not possible to measure iCa2+ concentrations in horses or sheep; this is similar to another report28 in humans in which the PCA was used. These discrepancies could be attributed to calibration algorithms16 or high pH values, which increase protein-bound calcium concentration and decrease iCa2+ concentration.13,16 If a high pH does influence iCa2+ concentration, it would not have been possible to measure iCa2+ concentration in cattle because the mean pH was similar to that in sheep and horses. Furthermore, in an attempt to accurately measure iCa2+ concentration as previously suggested,13,29 syringes, which were prefilled with standard quantities of dry calcium-balanced heparin, were used in our study to reduce the dilutional and calcium-binding effects of heparin. Most differences between values obtained with the PCA and RA devices could be accounted for by differences in the type of sample tested (serum vs whole blood14,16 or arterial blood vs venous blood7). However, the performance of the PCA for measuring iCa2+ concentration was judged to be acceptable in chickens,14 cats,16 dogs,7,16 and horses.7,16
Although the results obtained with the PCA for the concentrations of Na+ and K+ were significantly different from those of the RA, Na+ concentration was slightly overestimated and the mean difference for K+ concentration was small when compared with that of the RA, which indicated that concentrations of both electrolytes could be safely estimated by use of the PCA. In mice,12 results obtained with the PCA were low for concentrations of Na+ and Hct, whereas no difference was detected for K+ concentrations between instruments. Although the results obtained with the PCA for Na+ concentration generally did not correlate well with results obtained with the RA in horses and sheep, the SD of the PCA results was less and similar to results reported for mice12 and viscachas.18 Despite some discrepancies, results obtained with the PCA for concentrations of Na+ and K+ were within the reference ranges for all 3 species in the study reported here. Contrary to results of other studies,14,16,18 K+ concentrations obtained from blood analysis by use of the PCA were similar to those values reported from analysis of serum by use of an in-house analyzer and reference ranges used in those studies. In accordance with another report,1 the selection of whole blood for use in our study did not interfere with the accuracy of K+ values obtained with the PCA.
In our study, values obtained with the PCA for Hb concentration in ruminants were decreased when compared with values obtained with the RA. In addition, analysis of the Bland-Altman plot for Hb concentration revealed good agreement between the analyzers for samples obtained from cattle, whereas poor agreement was revealed for mice.12 In the study reported here, all biases were sufficiently infrequent to preclude clinical misinterpretation of venous blood samples collected from horses and ruminants.
The results reported for Hct via the PCA were decreased in cattle and sheep, but Hct results were similar to results obtained with the RA in horses. However, the reported Hct values were within the reference range in cattle and horses, whereas they were slightly less than the lower limit of the reference range in sheep.30 These differences were probably attributable to differences in the calibration procedures used for each system,1,18,31 errors attributable to atypical plasma osmolarity or atypical lipid concentrations,28 or equine erythrocyte characteristics (erythrocyte diameter, rouleaux formation, and rapid sedimentation secondary to rouleaux formation), which in turn would have decreased conductivity through the sample and would result in a subsequent decrease in the reported Hct values of equine samples measured via the PCA.7 Conversely, the range of diameter measurements of bovine and ovine erythrocytes (4 to 8 μm and 3.2 to 6 μm, respectively) is appreciably broader than the diameter measurements in equine erythrocytes; however, rouleaux formation is not typical in cattle and sheep, and the sedimentation rate is slow.31 In the study reported here, analysis of the Bland-Altman plot for Hct results in horses revealed variation of 20% in values between the analyzers; however, all values were within the 95% CI and there was no bias. Thus, results obtained via the PCA should be interpreted with caution in horses, especially when evaluating horses that are anemic.12
In general, the apparent discrepancies in our study might be a function of the type of analyzer, an extension of the algorithm method used,32 or simply because of the small number of samples analyzed. Furthermore, it is necessary to emphasize the importance of validating reference values for each laboratory, the analyzer used, and the species of animal evaluated. As a result of the overall adequate correlation with the RA and fact that the reported results were within the reported reference ranges, analysis of our results indicated that the data reported in the present study for blood gas and electrolyte analyses obtained via the PCA can be used for the evaluation of these variables for cattle, horses, and sheep.
ABBREVIATIONS
| BE | Base excess |
| CI | Confidence interval |
| Hb | Hemoglobin |
| HCO3− | Bicarbonate |
| iCa2+ | Ionized calcium |
| PCA | Portable clinical analyzer |
| RA | Reference analyzer |
| So2 | Oxygen saturation |
| tCO2 | Total carbon dioxide |
Dry heparinized syringe, Becton Dickinson, Franklin Lakes, NJ.
i-STAT, Abbott Laboratories, Abbott Park, Ill.
Eg7+ disposable cartridge, Abbott Laboratories, Abbott Park, Ill.
AVL Omni, 1–9, Roche Diagnostics Corp, Indianapolis, Ind.
References
- 1.↑
Erickson KA, Wilding P. Evaluation of a novel point-of-care system, the i-STAT portable clinical analyzer. Clin Chem 1993;39:283–287.
- 2.
Jacobs E, Vadasdi E, Sarkozi L, et alAnalytical evaluation of i-STAT portable clinical analyzer and use by nonlaboratory health-care professionals. Clin Chem 1993;39:1069–1074.
- 3.
Sibbald WJ, Eberhard JA, Inman KJ, et alNew technologies, critical care, and economic realities. Crit Care Med 1993;21:1777–1780.
- 4.
Woo J, Mccabe JB, Chauncey D, et alThe evaluation of a portable clinical analyzer in the emergency department. Am J Clin Pathol 1993;100:599–605.
- 5.
Wong RJ, Mahoney JJ, Harvey JA, et alStatPal II pH and blood gas analysis system evaluated. Clin Chem 1994;40:124–129.
- 6.
Tortella BJ, Lavery RF, Doran JV, et alPrecision, accuracy, and managed care implications of a hand-held whole blood analyzer in the prehospital setting. Clin Chem 1996;106:124–127.
- 7.↑
Looney AL, Ludders J, Erb HN, et alUse of a handheld device for analysis of blood electrolyte concentrations and blood gas partial pressures in dogs and horses. J Am Vet Med Assoc 1998;213:526–530.
- 8.
Tschudi PR. Evaluation of the portable blood analyzer i-STAT. Schweiz Arch Tierheilkd 1998;140:507–512.
- 9.↑
Larsen RS, Haulena M, Grindem CB, et alBlood values of juvenile northern elephant seals (Mirounga angustirostris) obtained using a portable clinical analyzer. Vet Clin Pathol 2002;31:106–110.
- 10.↑
Silverman SC, Birks EK. Evaluation of i-STAT hand-held chemical analyser during treadmill and endurance exercise. Equine Vet J Suppl 2002;34:551–554.
- 11.↑
Verwaerde P, Malet C, Lagente M, et alThe accuracy of the i-STAT portable analyser for measuring blood gases and pH in whole-blood samples from dogs. Res Vet Sci 2002;73:71–75.
- 12.↑
Tinkey P, Lembo T, Craig S, et alUse of i-STAT portable clinical analyzer in mice. Lab Anim (NY) 2006;35:45–50.
- 13.↑
Tappin S, Rizzo F, Dodkin S, et alMeasurement of ionized calcium in canine blood samples collected in prefilled and self-filled heparinized syringes using the i-STAT point-of-care analyzer. Vet Clin Pathol 2008;37:66–72.
- 14.↑
Steinmetz HW, Vogt R, Kästner S, et alEvaluation of the i-STAT portable clinical analyzer in chickens (Gallus gallus). J Vet Diagn Invest 2007;19:382–388.
- 15.↑
Sediame S, Zerah-Lancner F, d'Ortho MP, et alAccuracy of the i-STAT bedside blood gas analyzer. Eur Respir J 1999;14:214–217.
- 16.↑
Grosenbaugh DA, Gadawski KE, Muir WW. Evaluation of a portable clinical analyzer in a veterinary hospital setting. J Am Vet Med Assoc 1998;213:691–694.
- 17.
Cohn LA, McCaw DL, Tate DJ, et alAssessment 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.
- 18.↑
Wenker CJ, Hunziker D, Lopez J, et alHaematology, blood chemistry and urine parameters of free-ranging plains viscachas (Lagostomus maximus) in Argentina determined by use of a portable blood analyser (i-STAT) and conventional laboratory methods. J Vet Med A Physiol Pathol Clin Med 2007;54:260–264.
- 19.↑
Saulez MN, Cebra CK, Dailey M. Comparative biochemical analyses of venous blood and peritoneal fluid from horses with colic using a portable analyzer and an in-house analyser. Vet Rec 2005;157:217–223.
- 22.↑
Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–310.
- 23.↑
Gray TE, Pratt MC, Cusick PK. Determination of agreement between laboratory instruments. Contemp Top Lab Anim Sci 1999;38:56–59.
- 24.
Jensen AL, Bantz M. Comparing laboratory tests using the difference plot method. Vet Clin Pathol 1993;22:46–48.
- 25.
Papasouliotis K, Cue S, Crawford E, et alComparison of white blood cell differential percentages determined by the in-house LaserCyte hematology analyzer and a manual method. Vet Clin Pathol 2006;35:295–302.
- 26.
Bland JM, Altman DG. Agreement between methods of measurement with multiple observations per individual. J Biopharm Stat 2007;17:571–582.
- 27.↑
Radostitis OM, Gay CC, Hinchcliff KW, et alAppendix 2: reference laboratory values. In: Radostitis OM, Gay CC, Hinchcliff KW, et al, eds. Veterinary medicine. A textbook of the diseases of cattle, horses, sheep, pigs, and goats. 10th ed. Philadelphia: Saunders Elsevier, 2007;2047–2050.
- 28.↑
Daurès MF, Combescure C, Cristol JP. Étude comparative de six analyseurs de gaz du sang. Ann Biol Clin (Paris) 2007;65:505–518.
- 29.
Toffaletti J. Use of novel preparations of heparin to eliminate interference of ionized calcium measurements: have all the problems been solved? Clin Chem 1994;40:508–509.
- 30.↑
Carlson GP. Clinical chemistry tests: table 22–2. In: Smith BP, ed. Large animal internal medicine. 3rd ed. St Louis: Mosby, 2002;393.
- 31.↑
Kramer JW. Normal hematology of cattle, sheep, and goats. In: Feldman BF, Zinkl JG, Jain NC, eds. Schalm's veterinary hematology. 5th ed. Philadelphia: Lippincott Williams & Wilkins, 2000;1075–1084.
- 32.↑
Mock T, Morrison D, Yatscoff R. Evaluation of the i-STAT system: a portable chemistry analyzer for the measurement of sodium, potassium, chloride, urea, glucose, and hematocrit. Clin Biochem 1995;28:187–192.
