| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Brief Communication |
Correspondence: 1 Corresponding Author: Tracy Stokol, BVSc, DACVP, PhD, S1-047A Schurman Hall, College of Veterinary Medicine, Upper Tower Road, Ithaca, NY, 14853, ts23{at}cornell.edu
 |
Abstract |
|---|
 TOPSources and manufacturers Abstract References |
|---|
Key Words: ß-hydroxybutyrate • Bovine • hemolysis • interferences • nonesterified fatty acids
Hemolysis is a preanalytical variable that can have a major effect on the accuracy of biochemical assays. Hemolysis can occur in vivo, as part of an intravascular hemolytic anemia secondary to parasitic (e.g., Babesia bovis) or bacterial (e.g., Clostridial spp.) infections or metabolic disturbances (e.g., hypophosphatemia).1 However, hemolysis in most bovine samples is an artifact of sample collection and/or handling. Traumatic venipuncture, aggressive mixing of blood in collection tubes, exposure of blood to heat or extreme cold, delayed separation of serum from cells, and inadequate removal of erythrocytes after centrifugation may cause in vitro hemolysis.7, 11, 12 Bovine practitioners often collect blood from animals in the field, even during adverse weather, and do not have ready access to centrifuges. Therefore, some degree of hemolysis may be an unavoidable consequence of routine sample collection in this species.
Hemolysis interferes with biochemical assays by: 1) releasing intraerythrocyte constituents, which artifactually increase analyte values if the released constituents are found in higher concentrations in erythrocytes than in serum (e.g., potassium) or if they participate in the enzymatic reactions (e.g., creatine kinase), 2) diluting serum with intraerythrocyte water, which may decrease analyte values, 3) directly inhibiting reactions, and 4) spectral interference.4, 7, 11, 12 Most laboratory analytes are measured with techniques based on enzymatic reactions, where a change in absorbance over time (kinetic assay) or at a defined time point (endpoint assay) is measured spectrophotometrically at specific wavelengths. Free hemoglobin begins to absorb light at 340 nm and has a peak absorbance of 540–590 nm.4, 7, 11, 12 Thus, hemoglobin has the potential to interfere with measurements of enzymatic assays that are performed within hemoglobin's spectral range. The effect of this spectral interference is an increase in absorbance, which artifactually increases analyte values.
Nonesterified fatty acid (NEFA) and ß-hydroxybutyrate (BHB) concentrations are key metabolic analytes used as herd-based indicators of negative energy balance and subclinical ketosis in periparturient dairy cows, respectively.3, 10 NEFA increase secondary to lipolysis, stimulated by negative energy balance and values 
0.40 mEq/L in 10% of an appropriate sample size of dairy cows tested between 2 and 14 days before calving, is indicative of excessive negative energy balance.10 Similarly, BHB values 14 mg/dL in 10% of an appropriate sample size of dairy cows tested within 5–50 days after calving, indicate a herd problem with subclinical ketosis.3, 10 Hemolysis has been reported to artifactually increase ovine NEFA9 and bovine BHB5, 8 results when measured using spectrophotometric techniques. Since some degree of hemolysis is a common feature of many blood samples collected from cattle, particularly those sent by courier or mail to veterinary diagnostic laboratories, it is important to know the degree to which hemolysis will affect assay results and thus impact test interpretation.
The objectives of this study were: 1) to compare different reagents/assay procedures for measuring NEFA and BHB concentrations in bovine serum, 2) to determine the effect of hemolysis on NEFA and BHB concentrations in bovine serum, and 3) to determine if any effect of hemolysis on these tests was reagent/assay-dependent, particularly with respect to inclusion of a sample blank in the assay procedure.
Bovine serum samples (n = 26) submitted to the Animal Health Diagnostic Center at Cornell University for routine NEFA and BHB analysis were selected for inclusion in this study. Serum is an appropriate sample type for measuring these analytes in cattle.13 These samples had minimal to no visible hemolysis. Hemolysis is visible when the hemoglobin concentration exceeds 20 mg/dL.7 A red cell hemolysate was prepared from EDTA-anticoagulated blood collected from 2 dairy cows. The combined blood was centrifuged (3,000 g for 10 minutes) and the supernatant plasma removed. The red cell pellet was washed 3 times by gentle centrifugation (200 g for 5 minutes) in isotonic saline to remove all traces of plasma, the remaining isotonic saline was aspirated, and the red cell pellet was frozen at –80°C. The pellet was lysed by repeat freeze-thaw (37°C) and was centrifuged to remove cellular debris (3,000 g for 10 minutes). The supernatant was used as the hemolysate. The hemolysate was serially diluted in phosphate-buffered saline (PBS), then added to equal volumes of serum from the selected samples to produce mild to marked hemolysis, as assessed visually. The original serum sample was diluted 1:2 in PBS to provide baseline (nonhemolyzed) values.
NEFA and BHB concentrations were measured in all samples and the undiluted hemolysate with an automated chemistry analyzera with appropriate controls. This analyzer uses bichromatic wavelength pairs for determining absorbance of reactions. NEFAs are measured in a 2-stage reaction with an endpoint assay. In the first step, NEFAs are converted to thioesters of coenzyme A (acyl-CoA) by acyl-CoA synthetase in the presence of ATP, magnesium ions, and CoA. In the second step, the acyl-CoA is oxidized by acyl-CoA oxidase to produce hydrogen peroxide, which converts 3-methyl-N-ethyl-N-(ß-hydroxy-ethyl)-aniline and 4-aminoantipyrine in the presence of peroxidase to a colored product, which is detected at 550 nm. Two endpoint assays were evaluated in this study, one with (NEFA-Ab) and one without (NEFA-Cc) a sample blank. BHB was measured with 2 different reagents (BHB-Cd and BHB-Re). Both reagents detect NADH (340 nm primary wavelength) generated from the oxidation of BHB by ß-hydroxybutyrate dehydrogenase in the presence of NAD. BHB-C is a blanked endpoint assay, whereas BHB-R is a 2-point kinetic or rate assay, based on the change in absorbance as NADH is generated over time.
The analyzer also provides a semiquantitative measurement of hemolysis, the hemolytic index (HI). This is accomplished by measuring the absorbance of the sample at 2 different wavelength pairs (570–600 nm and 660–700 nm) and including the results into a formula which accounts for the degree of lipemia and a scaling factor for hemoglobinf. The HI approximates the amount of hemoglobin in mg/dl7 in the sample. Hemolytic indices of 0–20, 20–100, 100–400, and >400 units corresponded to visual assessments of no, mild, moderate, and marked hemolysis, respectively.
The data was not normally distributed, therefore nonparametric statisticsg were used. Results for the 2 NEFA or BHB assays were compared in baseline (nonhemolyzed) samples only by 1) generating and visually inspecting Bland-Altman bias plots6, 2) determining the difference between the assays, 3) comparing their median values with a Wilcoxon rank-sum test, and 4) calculating a Pearson correlation coefficient (r). To determine if hemolysis affected NEFA or BHB results, the median values for each assay at different levels of hemolysis were compared with a 1-way Kruskal-Wallis. Subsequent pairwise comparisons of the medians of the hemolyzed samples compared to baseline were done with a Wilcoxon rank-sum test. Significance was set at P < 0.05, with a Bonferroni adjustment for the number of pairwise comparisons as required.
The median HI of baseline samples was 8 units (range: 0–31 units). Median baseline NEFA concentrations were identical (0.25 mEq/L) for both assays (range for both: 0.06–1.03 mEq/L), P = 0.978. There was no systematic bias based on a median difference of 0 mEq/L (range: –0.01 to 0.02 mEq/L) between the 2 assays (NEFA-A minus NEFA-C) and visual inspection of bias plots (data not shown). The correlation between the NEFA-A and NEFA-C assays was excellent in these baseline samples (r = 1.00, P < 0.0001). Similarly, median BHB concentrations were identical (3 mg/dL) for baseline samples with both reagents (range, BHB-C: 1–24 mg/dL; BHB-R: 1–22 mg/dL), P = 0.991. There was a mild positive bias (median: 1 mg/dl, range: 0–2 mg/dL) with BHB-C compared to BHB-R; visual inspection of the data showed that this bias was independent of the BHB concentration (data not shown). There was an excellent correlation between the 2 assays for BHB on baseline nonhemolyzed samples (r = 1.00, P < 0.0001).
The undiluted hemolysate was brick red (HI: 18,550 units) and had NEFA values of 0 and 6.82 mEq/L with NEFA-A (blanked) and NEFA-C (nonblanked), respectively. The BHB value of the undiluted hemolysate was 0 mg/dl with both reagents. The hemolysate was serially diluted with the serum samples to yield median hemolytic indices of 74 (range, 59–100), 140 (range, 122–164), 258 (range, 241–283), 506 (range, 488–539), 1,014 (range, 979–1,067), and 1,911 (range, 1,810–2,013) units. Increasing hemolysis had opposite effects on NEFA concentrations with the 2 assays, decreasing NEFA-A (blanked) and increasing NEFA-C (nonblanked) values (Fig. 1). For NEFA-A, the decrease in NEFA values was apparent at a median HI of 506 units. Six of the 26 samples had baseline NEFAs at or above a critical medical decision limit of 0.40 mEq/L (range: 0.40–1.03 mEq/L). Result interpretation was affected (i.e., NEFA decreased below this limit) in 17% (1/6) and 33% (2/6) of these samples at a median HI of 506 and 1,911 units, respectively. Both samples had baseline NEFAs that were at or just above the decision limit (range: 0.40–0.47 mEq/L). For NEFA-C, the increase in NEFAs was apparent at a median HI of 140 units and the median NEFA was statistically higher than baseline at a median HI of 506 units. Indeed, at the latter HI, the median NEFA was higher than the medical decision limit (0.41 mEq/L). Twenty of the 26 samples had baseline NEFAs below the decision limit (range: 0.06–0.32 mEq/L). Result interpretation was affected in 5% (1/20), 40% (8/20), and 100% (20/20) at a median HI of 258, 506, and 1,014 units, respectively. In contrast to NEFAs, hemolysis had no effect on BHB concentrations with either reagent source (Fig. 2) and result interpretation was not affected in any individual sample.
|
|
These results indicated that either of the evaluated reagents can be used to measure BHB in bovine serum. The small positive bias in baseline samples with the BHB-C assay did not affect result interpretation and is probably within the limits of assay variability. In contrast to NEFAs, BHB was unaffected by hemolysis regardless of reagent source. These results differ from those of previous reports, in which BHB (measured spectrophotometrically) increased in bovine5,8 or decreased2 in canine serum with higher degrees of hemolysis. The same assay method (ß-hydroxybutyrate dehydrogenase) was used in all studies, although minor differences in reagent concentrations and reaction times are present. For BHB-C, the differences between this and previous studies can probably be attributed to inclusion of a sample blank for the endpoint assay (as previously reported for a different endpoint BHB assay in bovine serum8). For BHB-R, the use of a kinetic, rather than endpoint, assay may have eliminated the interference from hemolysis. A kinetic assay, which measures change in absorbance over unit time, is less likely to be affected by hemolysis, which may increase the total absorbance but will have minimal effect on the rate of change. However, this is unlikely to be the sole explanation, because the previous studies in bovine5 and canine2 serum also used kinetic assays for BHB. Perhaps the use of 2 time points to measure the rate change or the bichromatic wavelength technique (340 nm primary, 700 nm secondary) used with the BHB-R assay in this study contributed to these different results.
In conclusion, similar NEFA and BHB results will be obtained with the evaluated reagents in nonhemolyzed bovine samples. NEFA assays in all species should include a sample blank (NEFA-A), because even mild hemolysis can affect result accuracy when a blank is not included (NEFA-C). Result accuracy for NEFAs, regardless of assay procedure, cannot be assured in moderately to severely hemolyzed samples (HI > 400 units) and such samples should not be analyzed for NEFAs. In contrast, BHB values, as measured with the instrumentation and reagents evaluated in this study, are unaffected by any degree of hemolysis.
|
Acknowledgments |
|---|
|
Sources and manufacturers |
|---|
|
TOPSources and manufacturersAbstractReferences |
|---|
a. Hitachi 917, Roche Diagnostics, Indianapolis, IN.
b. NEFA-A, Wako Chemicals USA, Inc., Richmond, VA.
c. NEFA-C, Wako Chemicals USA, Inc., Richmond, VA.
d. ß-HBA, Catachem Inc, Bridgeport, CT.
e. D-3-hydroxybutyrate, Randox Laboratories Ltd, Ardmore, UK.
f. Hitachi 9–11 Analyzer Operator's Manual, Roche Diagnostics, Indianapolis, IN.
g. Analyse-it for Excel, Analyse-it Software Ltd, Leeds, UK.
h. Reagent specification sheet, NEFA-A, Wako Chemicals USA, Inc., Richmond, VA.
|
References |
|---|
|
TOPSources and manufacturersAbstractReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
