for peer review - university of wisconsin–madison2013/11/01 · for peer review 4 48 precision...
TRANSCRIPT
For Peer Review
TECHNICAL NOTE: Validation of BHBCheck blood β-
hydroxybutyrate meter as a diagnostic tool for hyperketonemia.
Journal: Journal of Dairy Science
Manuscript ID JDS-17-13583.R1
Article Type: Technical Notes
Date Submitted by the Author: 09-Oct-2017
Complete List of Authors: Sailer, Kayla; University of Wisconsin-Madison, Dairy Science
Pralle, Ryan; University of Wisconsin Madison, Department of Dairy Science OLIVEIRA, RAFAEL; University of Wisconsin, Dairy Science Erb, Sophia; University of Wisconsin-Madison, Dairy Science Oetzel, Garrett; School of Veterinary Medicine, University of Wisconsin-Madison 53706, Department of Medical Sciences White, Heather; University of Wisconsin-Madison, Dairy Science
Key Words: ketosis, cowside diagnostic tool, transition cow
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
1
Interpretive Summary 1
2
TECHNICAL NOTE: Validation of BHBCheck blood β-hydroxybutyrate meter as a 3
diagnostic tool for hyperketonemia in dairy cows 4
By Sailer et al. 5
Accurate cowside blood β-hydroxybutyrate detection meters are valuable tools for 6
evaluating postpartum hyperketonemia. Optimal diagnostic tools must be sensitive and 7
specific, as well as cost and labor effective to be implemented into routine testing 8
protocols. We evaluated the use of the BHBCheck blood meter for use in ketosis 9
detection compared to a laboratory assay and a previously validated meter. The 10
BHBCheck meter performed similarly to previously validated meters and had adequate 11
diagnostic accuracy for determining hyperketonemia in dairy cows. 12
Page 1 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
2
BHBCHECK METER FOR HYPERKETONEMIA DETECTION 13
14
TECHNICAL NOTE: Validation of the BHBCheck blood β-hydroxybutyrate meter 15
as a diagnostic tool for hyperketonemia in dairy cows 16
17
K. J. Sailer*, R. S. Pralle*, R. C. Oliveira*, S. J. Erb*, G. R. Oetzel†, and H. M. White*1 18
19
*Department of Dairy Science, University of Wisconsin-Madison, Madison, WI 53706 20
†School of Veterinary Medicine, University of Wisconsin-Madison, Madison, WI 53706 21
1Corresponding author: Heather M. White, Department of Dairy Science, University of 22
Wisconsin-Madison, 1675 Observatory Drive, Rm 934B, Madison WI, 53706, Office: 23
608-263-7786, Fax: 608-263-7786, email: [email protected] 24
Page 2 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
3
ABSTRACT 25
Accurate cowside blood β-hydroxybutyrate (BHB) detection meters are valuable 26
tools for rapid diagnosis of hyperketonemia. The main objective of this study was to 27
compare the blood BHB measured in whole blood by the BHBCheck meter (PortaCheck, 28
Moorestown, NJ) to a previously validated meter, Precision Xtra meter (Abbott 29
Laboratories, Abbott Park, Illinois), and a colorimetric laboratory assay. Samples (n=426) 30
were collected from postpartum primiparous and multiparous Holstein cows (n=79 cows) 31
enrolled in one of two experiments (Exp) with different sampling schedules (Exp 1: n=39 32
cows, 58 samples; Exp 2: n=40 cows, 368 samples). In both Exp, whole blood samples 33
were collected from the coccygeal vessels after morning milking, prior to morning 34
feeding. Blood samples were used immediately for BHB quantification via the 35
BHBCheck meter and the Precision Xtra meter. Blood was also collected into evacuated 36
tubes containing no additive (Exp 1) or Potassium Oxalate/Sodium Fluoride (Exp 2), 37
which were centrifuged for serum or plasma separation and stored at -20˚C for 38
subsequent analysis. Laboratory quantification of BHB concentration was done by the 39
BHB LiquiColor Assay (EKF Diagnostics - Stanbio; Boerne, TX; certified for serum and 40
plasma). Data were analyzed by UNIVARIATE, CORR, FREQ, REG, and LOGISTIC 41
procedures of SAS 9.4. Within this sample set, average parity was 3.3 lactations and DIM 42
was 14 d. The proportion of samples classified as hyperketonemia (BHB ≥ 1.2 mmol/L) 43
was 25, 28, and 31% as determined by the colorimetric assay, BHBCheck meter, and 44
Precision Xtra meter, respectively. The correlation for BHBCheck meter BHB 45
concentration compared to the colorimetric assay concentrations was r = 0.96 with a 46
sensitivity of 91% and specificity of 93%. Correlation, sensitivity, and specificity of the 47
Page 3 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
4
Precision Xtra meter concentrations were 0.97, 98%, and 92%, respectively. Bland 48
Altman plots demonstrated minimal bias for both meters. Area under the ROC curve 49
suggests adequate diagnostic accuracy of both meters. Overall, accuracy, sensitivity, and 50
specificity of the BHBCheck meter was similar to the Precision Xtra meter and 51
laboratory assay and is appropriate for use as a cowside diagnostic test for 52
hyperketonemia in dairy cows. 53
Keywords: Ketosis, cowside diagnostic tool, transition cow 54
Page 4 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
5
TECHNICAL NOTE 55
During the postpartum period, dairy cows are at increased risk of developing 56
metabolic disorders, largely due to negative energy balance (NEB) and subsequent 57
mobilization of triglyceride (TG) from adipose tissue (Grummer, 1993; Drackley, 1999; 58
Duffield, 2000). Fates of mobilized TG within the liver include complete oxidation to 59
energy, incomplete oxidation to ketone bodies, or storage as TG (Grummer, 1993). 60
Although some peripheral tissues can use ketone bodies as an energy source, production 61
of ketone bodies beyond tissue usage results hyperketonemia (HYK). When untreated, 62
HYK is associated with negative impacts on animal health, production, and profitability 63
(Herdt, 2000; McArt et al., 2015). Hyperketonemia is defined as blood β-hydroxybutyrate 64
(BHB) ≥ 1.2 mmol/L (Iwersen et al., 2009; McArt et al., 2012b; Gordon et al., 2013; 65
2017) and the average postpartum prevalence of HYK worldwide ranges from 15 to 22%, 66
although it is highly variable by farm (Suthar et al., 2013; Chandler et al., 2015; Santschi 67
et al., 2016). The average cost per case is $289 (McArt et al., 2015); however, early 68
treatment can reduce the costs and negative outcomes (McArt et al., 2012a), making 69
detection protocols that are comprised of labor- and cost-effective diagnostics essential to 70
managing HYK. 71
Previous validation of a handheld blood ketone meter, the Precision Xtra meter 72
(Abbott Laboratories, Abbott Park, IL), provided a valuable quantitative tool for 73
diagnosing HYK using a minimally invasive blood sample, that also had greater 74
sensitivity and specificity than previously used milk and urine tests (Iwersen et al., 2009; 75
Oetzel, 2004). Recently, several additional meters have been developed and some have 76
been evaluated for potential use in bovine HYK diagnostics (Bach et al., 2016). 77
Page 5 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
6
Validation of these diagnostic tests in a controlled setting, prior to field application, is 78
essential. A relatively new meter that has not yet been validated for use in bovine HYK 79
diagnostics is the BHBCheck meter (PortaCheck, Moorestown, NJ). This meter 80
quantifies BHB in a 0.7 µL whole blood sample, which is less blood than other meters 81
which require between 0.8 to 1.5 µL of blood, and quantification is completed in under 5 82
sec, compared to other meters that average 10 sec. We hypothesized that similar to other 83
hand-held BHB diagnostic meters, the BHBCheck meter would provide adequate 84
accuracy for HYK diagnostics in dairy cattle. The objectives of this study were 1) to 85
determine the efficacy of the BHBCheck meter as a cowside BHB monitoring device by 86
validating it against a colorimetric laboratory assay and 2) to compare the BHBCheck 87
meter to an industry standard cowside meter (Precision Xtra). 88
We conducted two animal experiments that were approved by the University of 89
Wisconsin-Madison College of Agricultural and Life Sciences Animal Care and Use 90
Committee (protocol numbers A05467 and A01569). For both experiments (Exp), blood 91
was collected from Holstein cows immediately following morning milking and prior to 92
feeding. The samples from Exp 1 were collected from 39 primiparous and multiparous 93
cows between 4 and 18 DIM during twice weekly sampling on a privately-owned dairy in 94
south-central Wisconsin from May to June 2016. Experiment 1 was intended to provide 95
a balanced dataset of an equal number of samples that were above and below the 1.2 96
mmol/L HYK threshold. This was accomplished by collecting whole blood from cows 97
between 4 and 18 DIM into evacuated collection tubes without additives and immediately 98
analyzing blood using the Precision Xtra test. When a cow with HYK was identified, she 99
was included in the study along with the next cow that was negative for HYK. Whole 100
Page 6 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
7
blood samples were also analyzed using the BHBCheck meter. Samples were collected 101
over several farm visits and some cows contributed more than one sample to the dataset. 102
Although the balanced design in Exp 1 resulted in a nonrepresentative sample set relative 103
to the population prevalence, it avoided having a sample set with a minimal proportion of 104
samples above the 1.2 mmol/L threshold. Whole blood tubes were then centrifuged 105
(2,500 x g, 15°C, 15 minutes) and serum was aliquoted and stored into 1.5 mL 106
microtubes at -20°C until subsequent analysis. Experiment 2 was conducted at the 107
University of Wisconsin-Madison Dairy Cattle Center from May to October 2016. Blood 108
samples were collected from 40 multiparous cows between 1 and 45 DIM into an 109
evacuated tube without additives, and into an evacuated tube containing potassium 110
oxalate and sodium fluoride. Whole blood from the tube without additives was 111
immediately analyzed on the BHBCheck and Precision Xtra meter. The tube containing 112
potassium oxalate and sodium fluoride was centrifuged (2,000 x g, 4°C, 15 minutes) and 113
plasma was aliquoted into 1.5 mL microtubes, and stored at -20°C for subsequent 114
analysis. 115
Serum samples from Exp 1 and plasma samples from Exp 2 were analyzed using the 116
LiquiColor colorimetric assay (EKF Diagnostics - Stanbio, Boerne, TX; certified for 117
serum and plasma collected with EDTA, heparin, or sodium fluoride) per the 118
manufacturer’s protocol. Samples were quantified in duplicate and samples with intra-119
assay CV greater than 10% were reanalyzed. The inter-assay CV of the laboratory assay 120
was 6.7%. Sample BHB concentration determined by the laboratory assay served as the 121
“gold standard” BHB concentration for statistical analysis. 122
Page 7 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
8
Data analysis was completed in SAS 9.4 (Cary, NC). Descriptive statistics were 123
analyzed by PROC UNIVARIATE. Residuals were produced by PROC REG with 124
colorimetric laboratory assay BHB as the dependent variable and BHBCheck or Precision 125
Xtra meter quantified BHB as the independent variable. Pearson correlations between the 126
BHB concentration determined by both meters and by the colorimetric assay were 127
analyzed by PROC CORR. In order to account for correlation among residual errors due 128
to the repeated measures on the experimental unit (cow) across time, different error 129
structures (i.e. compound symmetry, first order auto-regressive, and spatial power) were 130
tested using PROC MIXED. A log-likelihood ratio tests was performed between 131
the repeated measure models and the null model and indicated accounting for repeated 132
structure of the data did not significantly alter model fit. Therefore, analysis of the data 133
continued without accounting for repeated measures. Sensitivity and specificity were 134
calculated by PROC FREQ. Receiver operating characteristic (ROC) curves and 135
associated area under the curve were determined by PROC LOGISTIC. Results for Exp 1 136
and 2 are provided both separately, and combined, in order to provide the diagnostic 137
accuracy of the meters in two sample sets with different proportion of samples greater 138
than or equal to 1.2 mmol/L. 139
Demographics of the samples from Exp 1, 2, and Overall are provided in Table 1. 140
Samples were collected from a total of 79 Holstein cows with an average parity of 3.3 ± 141
0.1 and average DIM of 14.3 ± 0.6 across the two experiments (Table 1). The mean BHB 142
was 1.3 ± 0.12, 1.0 ± 0.03, and 1.0 ± 0.03 mmol/L for samples from Exp 1, 2, and 143
Overall, respectively. The range of BHB was from 0.3 to 5.4 mmol/L across all samples, 144
as determined by colorimetric laboratory assay. 145
Page 8 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
9
Proportion of samples with BHB ≥ 1.2 mmol/L, as determined by the colorimetric 146
assay were 50% for Exp 1, 21% for Exp 2, and 25% for the Overall dataset. Use of either 147
meter to calculate the proportion of samples with BHB ≥ 1.2 mmol/L resulted in slight 148
overestimations in Exp 2 and Overall for the BHBCheck (50%, 25%, and 28% for Exp 1, 149
Exp 2, and Overall) and in Exp 1, Exp 2, and Overall for the Precision Xtra (55%, 27%, 150
and 31% for Exp 1, Exp 2, and Overall). This overestimation highlights the possibility of 151
false positives when using a cowside BHB meter to diagnose HYK, and would result in 152
treating a small number of cows that do not have blood BHB ≥ 1.2 mmol/L per the 153
laboratory assay. Treating false positives does have a cost associated with it dependent on 154
the disease and treatment; however, in the case of HYK, the cost of treatment is relatively 155
low compared with the cost of an untreated HYK case (McArt et al., 2015). 156
Concentrations of BHB as determined by the BHBCheck or Precision Xtra meter were 157
correlated (P<0.001) with concentrations determined by the colorimetric laboratory 158
assay. Correlation coefficients were 0.98 for both meters compared with the laboratory 159
assay in Exp 1. The correlation coefficients were 0.96 and 0.97, for the BHBCheck and 160
Precision Xtra meter, respectively, in Exp 2 and the Overall dataset. Interestingly, the 161
correlation between the two meters was 0.98 for Exp 1, 2, and Overall. Fit statistics for 162
linear regression analysis of the BHBCheck and laboratory assay resulted in an r2 and 163
RMSE of 0.96 and 0.17 for Exp 1, 0.92 and 0.19 for Exp 2, and 0.93 and 0.19 Overall 164
(Figure 1). 165
Linear regression analysis of the Precision Xtra meter compared with the laboratory 166
assay resulted in an r2 and RMSE of 0.97 and 0.17 for Exp 1, 0.94 and 0.17 for Exp 2, 167
and 0.94 and 0.17 Overall (Figure 1). Bland Altman plots (Bland and Altman, 1986) 168
Page 9 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
10
show test agreement (Figure 2) and demonstrate minimal bias for both meters (-0.05 and 169
-0.09 for the BHBCheck and Precision Xtra meter, respectively, compared with the 170
laboratory assay). Bias of -0.05 and -0.09 would not influence diagnostic ability of the 171
test cowside and is consistent with minimal bias previously observed with cowside BHB 172
diagnostic tests (Bach et al., 2016). 173
Although correlation and regression methods are valuable indicators of a diagnostic 174
tool’s quantitative accuracy, it is equally important to evaluate the sensitivity and 175
specificity of the meters at relevant diagnostic cutoff concentrations. Sensitivity and 176
specificity of the Precision Xtra meter was 98 and 92% at 1.2 mmol/L for the Overall 177
dataset in the current study (Table 2), which is consistent with previous validation of this 178
meter (Iwersen et al., 2009; Voyvoda and Erdogan, 2010; Bach et al., 2016). Similarly, 179
the BHBCheck meter had a sensitivity and specificity of 91 and 93% at the 1.2 mmol/L 180
threshold. Sensitivity and specificity were greater in Exp 2 than Exp 1, likely due to the 181
difference in sample size. Although both sensitivity and specificity are important 182
indicators of a diagnostic tool, sensitivity is arguably more valuable for subclinical 183
disorders with low-cost and low-risk treatments. Subclinical animals would fail to be 184
diagnosed if missed by the diagnostic test, given the lack of obvious disease symptoms. 185
In the case of HYK, the treatment cost of treatment is relatively inexpensive compared to 186
the cost of an untreated case (McArt et al., 2015); thus, adequate sensitivity of a potential 187
cowside diagnostic tool is important to consider (Bach et al., 2016). 188
Determination of the area under the ROC curve is an objective means of evaluating 189
the overall validity of a diagnostic tool (Šimundić, 2009). Receiver operating 190
characteristic curves represent the paired sensitivity and specificity for each cutoff, and 191
Page 10 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
11
when plotted, the resulting curve allows for estimation of the discriminative power of a 192
test. Areas under the ROC between 0.9 and 1.0 represent “excellent” diagnostic accuracy 193
according to ROC diagnostic categories (Šimundić, 2009). In the current study, area 194
under the ROC for the BHBCheck meter was 0.98, 0.98, and 0.98 for Exp 1, Exp 2, and 195
Overall. The area under the ROC curve for the Precision Xtra meter was 0.98, 0.99, and 196
0.99 for Exp 1, Exp 2, and Overall. This evaluation indicates that both the Precision Xtra 197
and BHBCheck meters portray adequate diagnostic abilities. 198
Overall, performance of the Precision Xtra meter within the current study was 199
consistent with previous studies (Iwersen et al., 2009; Bach et al., 2016). The Precision 200
Xtra meter has been previously validated and has set the expectation for sensitivity and 201
specificity for a cowside BHB meter. While it is advantageous to have several diagnostic 202
options, diagnostic tools need to be evaluated to ensure test accuracy. Within the current 203
study, the BHBCheck meter performed similarly to the Precision Xtra meter in this study 204
and to other meters previously examined (Bach et al., 2016; Iwersen et al., 2009; 205
Voyvoda and Erdogan, 2010). Both meters were consistent with expectations of a 206
diagnostic tool and displayed adequate diagnostic accuracy. We conclude that the 207
BHBCheck meter is an accurate and valuable option for use in HYK detection protocols. 208
209
ACKNOWLEDGEMENTS 210
This research was partially funded by gifts from the Wisconsin Alumni Research 211
Foundation. BHBCheck meters and strips were provided by PortaCheck. The authors 212
would like to thank the University of Wisconsin-Madison Dairy Cattle Center staff as 213
Page 11 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
12
well as the owner and herdsman of the privately-owned dairy for their cooperation during 214
this research project. 215
216
Page 12 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
13
REFERENCES 217
Bach, K. D., W. Heuwieser, and J. A. A. McArt. 2016. Technical note: Comparison of 4 218 electronic handheld meters for diagnosing hyperketonemia in dairy cows. J. Dairy 219 Sci. 99:9136–9142. doi:10.3168/jds.2016-11077. 220
Bland, J. M., and D. G. Altman. 1986. Statistical methods for assessing agreement 221 between two methods of clinical measurement. Lancet. 1:307-310. 222
Chandler, T. L., R. S. Pralle, G. R. Oetzel, R. H. Fourdraine, and H. M. White. 2015. 223 Development of a ketosis prevalence tool in Holstein dairy cows based on milk 224 component data and cow test-day information. J. Dairy Sci. Vol 98, Suppl 2. 225
Drackley, J. K. 1999. ADSA Foundation Scholar Award. Biology of dairy cows during 226 the transition period: the final frontier? J. Dairy Sci. 82:2259–2273. 227 doi:10.3168/jds.S0022-0302(99)75474-3. 228
Duffield, T. 2000. Subclinical ketosis in lactating dairy cattle. Vet. Clin. North Am. Food 229 Anim. Pract. 16:231–253. 230
Gordon, J. L., S. J. Leblanc, and T. F. Duffield. 2013. Ketosis treatment in lactating dairy 231 cattle. Vet. Clin. North Am. Food Anim. Pract. 29:433–445. 232 doi:10.1016/j.cvfa.2013.03.001. 233
Gordon, J. L., T. F. Duffield, T. H. Herdt, D. F. Kelton, L. Neuder, and S. J. LeBlanc. 234 2017. Effects of a combination butaphosphan and cyanocobalamin product and 235 insulin on ketosis resolution and milk production. J. Dairy Sci. 100:2954–2966. 236 doi:10.3168/jds.2016-11925. 237
Grummer, R. R. 1993. Etiology of Lipid-Related Metabolic Disorders in Periparturient 238 Dairy Cows. J. Dairy Sci. 76:1–15. 239
Herdt, T. H. 2000. Ruminant adaptation to negative energy balance. Influences on the 240 etiology of ketosis and fatty liver. Vet. Clin. North Am. Food Anim. Pract. 16:215–241 230. doi: 10.1016/S0749-0720(15)30102-X. 242
Iwersen, M., U. Falkenberg, R. Voigtsberger, D. Forderung, and W. Heuwieser. 2009. 243 Evaluation of an electronic cowside test to detect subclinical ketosis in dairy cows. J. 244 Dairy Sci. 92:2618–2624. doi:10.3168/jds.2008-1795. 245
McArt, J. A. A., D. V. Nydam, and G. R. Oetzel. 2012a. A field trial on the effect of 246 propylene glycol on displaced abomasum, removal from herd, and reproduction in 247 fresh cows diagnosed with subclinical ketosis. J. Dairy Sci. 95:2505–2512. 248 doi:10.3168/jds.2011-4908. 249
McArt, J. A. A., D. V. Nydam, and G. R. Oetzel. 2012b. Epidemiology of subclinical 250 ketosis in early lactation dairy cattle. J. Dairy Sci. 95:5056–5066. 251 doi:10.3168/jds.2012-5443. 252
Page 13 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
14
McArt, J. A. A., D. V. Nydam, and M. W. Overton. 2015. Hyperketonemia in early 253 lactation dairy cattle: A deterministic estimate of component and total cost per case. J. 254 Dairy Sci. 98:2043–2054. doi:10.3168/jds.2014-8740. 255
Oetzel, G. R. 2004. Monitoring and testing dairy herds for metabolic disease. Vet. Clin. 256 North Am. Food Anim. Pract. 20:651–674. doi:10.1016/j.cvfa.2004.06.006. 257
Santschi, D. E., R. Lacroix, J. Durocher, M. Duplessis, R. K. Moore, and D. M. Lefebvre. 258 2016. Prevalence of elevated milk β-hydroxybutyrate concentrations in Holstein 259 cows measured by Fourier-transform infrared analysis in Dairy Herd Improvement 260 milk samples and association with milk yield and components. J. Dairy Sci. 99:9263–261 9270. doi:10.3168/jds.2016-11128. 262
Šimundić, A. 2009. Measures of diagnostic accuracy: basic definitions. eJIFCC. 19:203-263 211. 264
Suthar, V. S., J. Canelas-Raposo, A. Deniz, and W. Heuwieser. 2013. Prevalence of 265 subclinical ketosis and relationships with postpartum diseases in European dairy 266 cows. J. Dairy Sci. 96:2925–2938. doi:10.3168/jds.2012-6035. 267
Voyvoda, H., and H. Erdogan. 2010. Use of a hand-held meter for detecting subclinical 268 ketosis in dairy cows. Res.Vet. Sci. 89:344–351. doi:10.1016/j.rvsc.2010.04.007. 269
270
Page 14 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
15
Table 1. Demographics of samples from experiment (Exp) 1 (collected between 4 and 18 271
DIM), Exp 2 (collected between 1 and 45 DIM), and Overall used to evaluate the 272
cowside blood β-hydroxybutyrate meters1 compared with a colorimetric laboratory 273
assay2. 274
Exp 1 Exp 2 Overall
Unique cows 39 40 79 Samples 58 368 426 Cow variables
parity ± SE 2.7 ± 0.18 3.3 ± 0.08 3.3 ± 0.08
DIM ± SE 10.2 ± 0.56 15.1 ± 0.66 14.3 ± 0.58
BHB3, mmol/L
mean ± SE 1.3 ± 0.12 1.0 ± 0.03 1.0 ± 0.03
median 1.2 0.7 0.8 min 0.3 0.3 0.3
max 3.9 5.4 5.4 275
1BHBCheck meter (PortaCheck, Moorestown, NJ) and Precision Xtra meter (Abbott 276
Laboratories, Abbott Park, Illinois). 277
2LiquiColor colorimetric assay (EKF Diagnostics - Stanbio, Boerne, TX). 278
3As quantified by the colorimetric laboratory assay. 279
280
281
282
Page 15 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
16
Table 2. Sensitivity and specificity of cowside blood β-hydroxybutyrate meters1 to 283
correctly identify samples as greater than or equal to 1.2 mmol/L, based on a colorimetric 284
laboratory assay2 within experiment (Exp) 1 (collected between 4 and 18 DIM), Exp 2 285
(collected between 1 and 45 DIM), and Overall. 286
Exp 1 Exp 2 Overall Samples 58 368 426 Samples ≥ 1.2 mmol/L 29
77
106
BHBCheck Sens, % 89 [71,98] 3
91 [84,96]
91 [84,96]
Spec, % 84 [66,95]
93 [89,95]
93 [89,95] Precision Xtra
Sens, % 100 [87,100] 97 [91,99] 98 [93,99] Spec, % 84 [66,95] 92 [88,95] 92 [88,94]
287
1BHBCheck meter (PortaCheck, Moorestown, NJ) and Precision Xtra meter (Abbott 288
Laboratories, Abbott Park, Illinois). 289
2LiquiColor colorimetric assay (EKF Diagnostics - Stanbio, Boerne, TX). 290
3 [95% CI] 291
292
Page 16 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
17
Sailer Figure 1. Linear regression of blood β-hydroxybutyrate (BHB) concentrations 293
quantified by BHBCheck (PortaCheck, Moorestown, NJ; left panels) and Precision Xtra 294
meter (Abbott Laboratories, Abbott Park, Illinois; right panels) compared with a 295
colorimetric laboratory assay (LiquiColor colorimetric assay; EKF Diagnostics - Stanbio, 296
Boerne, TX) within experiment (Exp) 1 (top panels; 58 samples, collected between 4 and 297
18 DIM), Exp 2 (368 samples, collected between 1 and 45 DIM), and Overall (bottom 298
panels). 299
300
Sailer Figure 2. Differences of blood BHB concentrations determined in 426 samples 301
(collected between 1 and 45 DIM across two studies) by a laboratory assay (LiquiColor 302
colorimetric assay; EKF Diagnostics - Stanbio, Boerne, TX) and BHBCheck 303
(PortaCheck, Moorestown, NJ; top panel) or Precision Xtra meter (Abbott Laboratories, 304
Abbott Park, Illinois; middle panel) and between the two meters (bottom panel). 305
306
307
Page 17 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
18
308 Sailer Figure 1. 309 310 311
312 313 314
BHBCheck BHB, mmol/L Precision Xtra BHB, mmol/L
r = 0.98
r = 0.96
r = 0.96
r = 0.98
r = 0.97
r = 0.97
Page 18 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science
For Peer Review
19
Sailer Figure 2. 315
316
317
318
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Ass
ay B
HB
-B
HB
Che
ck
BH
B, m
mol
/L
Average BHB for Assay and BHBCheck, mmol/L
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Ass
ay B
HB
-P
reci
sion
Xtr
a B
HB
, mm
ol/L
Average BHB for Assay and Precision Xtra, mmol/L
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Pre
cisi
on X
tra
BH
B -
BH
BC
heck
BH
B, m
mol
/L
Average BHB for PrecisionXtra and BHBCheck, mmol/L
+2 SD = 0.34
-2 SD = -0.44
mean = -0.05
+2 SD = 0.33
-2 SD = -0.51
mean = -0.09
+2 SD = 0.38 mean = 0.04
-2 SD = -0.29
Page 19 of 19
ScholarOne support: (434) 964 4100
Journal of Dairy Science