The transition period, 3 weeks prior to parturition to 3 weeks after parturition,1 is arguably the time of greatest metabolic challenge for adult dairy cows. Dairy cows frequently undergo periods of negative energy balance during the latter stages of gestation and the early stages of lactation because their dry matter consumption decreases in the days just prior to parturition, and immediately after parturition, they frequently cannot consume enough feed to meet the energy and protein requirements for initiation of lactation.2,3 Negative energy balance results in the mobilization of fatty acids, some of which are metabolized into ketones (acetone, acetoacetate, and BHBA) by the liver. Dairy cows normally have small concentrations of ketones in their blood during the early stages of lactation. A maladaptive response to negative energy balance can cause abnormally high concentrations of ketones in the circulation,3 and that response can begin before parturition in some cows. Several methods for ketone detection are commercially available and include tests that use milk, urine, and whole blood samples; however, the gold standard for detection of abnormally increased concentrations of ketones, or ketosis, remains measurement of serum BHBA concentration by standard laboratory methods.
Cows with serum BHBA concentrations > 1.0 to 1.4 mmol/L during the early postpartum period have lower milk production and are at an increased risk of developing clinical ketosis, displaced abomasum, and metritis, compared with similar cows with serum BHBA concentrations < 1.0 mmol/L.4–6 Results of 1 study7 indicate that the risk of clinical disease is greater for cows in which subclinical ketosis (BHBA concentration, 1.2 to 2.9 mmol/L) is diagnosed during the first week after parturition, compared with cows in which subclinical ketosis is diagnosed during the second week after parturition. The time at which ketosis is diagnosed also affects how well affected cows will respond to treatment. Cows in which ketosis (blood BHBA concentration, ≥ 1.2 mmol/L) is diagnosed between 3 and 9 days after parturition have a lower cure rate following treatment with propylene glycol, compared with cows in which ketosis is diagnosed > 9 days after parturition.a An abnormal increase in the concentration of circulating ketones in dairy cattle during the first week after parturition may be an indication of metabolic problems during the late gestation period, and detection of those metabolic problems prior to parturition may be beneficial for the management of affected cows during the transition period.
Cows with serum NEFA concentrations between 0.29 and 0.50 mEq/L during the week prior to parturition are at an increased risk of developing displaced abomasum, metritis, and ketosis and having lower milk production, compared with cows with serum NEFA concentrations < 0.29 mEq/L.6,8,9 Although evaluation of prepartum serum NEFA concentration is useful, it is inconvenient because blood samples should be collected only at certain times after feeding and must be kept chilled until processing and analysis, which can only be performed in a laboratory.
The association of prepartum serum BHBA concentration with various postpartum diseases and production variables has been investigated in multiple studies.8,10,11 In 1 study,10 cows with a serum BHBA concentration ≥ 0.6 mmol/L during the last 10 days of gestation had a lower milk yield on the first test day after parturition than did cows with a prepartum serum BHBA concentration < 0.6 mmol/L. In other studies, cows with a BHBA concentration ≥ 0.7 mmol/L 1 week prior to parturition were at increased risk of being culled within 60 days after parturition, compared with cows with prepartum BHBA concentrations < 0.7 mmol/L,11 and cows with a BHBA concentration ≥ 0.8 mmol/L 10 days prior to parturition were at an increased risk of developing a displaced abomasum within 60 days after parturition, compared with cows with a prepartum BHBA concentration < 0.8 mmol/L.8
A commercially available handheld device is available for cow-side measurement of BHBA concentration in whole blood samples. Results of a study12 performed to validate the device for measurement of postpartum BHBA concentration indicate that BHBA measurements obtained with the handheld device were highly correlated (r = 0.95) with BHBA measurements obtained by standard laboratory methods, and the sensitivity and specificity of BHBA concentration as determined by the handheld device for identification of cows with subclinical ketosis were 88% and 96%, respectively, when the cutoff for diagnosis was a BHBA concentration ≥ 1.2 mmol/L and 96% and 97%, respectively, when the cutoff for diagnosis was a BHBA concentration ≥ 1.4 mmol/L, compared with that determined by standard laboratory methods. The sensitivity and specificity (ie, test performance) of postpartum BHBA concentrations as determined by the handheld device for identification of cows with ketosis are positively associated with the blood BHBA concentration used as the cutoff for diagnosis of the disease.12 The decrease in test performance as the cutoff concentration of BHBA for diagnosis of ketosis decreases has raised concern about use of the device for identification of prepartum cows at risk of developing postpartum ketosis because prepartum BHBA concentration is generally lower than postpartum BHBA concentration.
The primary objective of the study reported here was to evaluate the use of a handheld device for measurement of BHBA concentration for identification of prepartum dairy cattle at risk of developing hyperketonemia during the first week after parturition. Other objectives were to validate the accuracy of BHBA concentration as determined by the handheld device, compared with that determined by standard laboratory methods in prepartum dairy cattle; to evaluate the association between prepartum BHBA concentration as determined by the handheld device and the incidence of hyperketonemia within 1 week after parturition; and to assess the agreement between prepartum BHBA and NEFA concentrations for the identification of cows at risk of developing hyperketonemia within the first week after parturition.
Materials and Methods
Animals—The study protocol was approved by the Animal Care Committee of the University of Guelph, and herd owners consented to having blood samples obtained from the cows enrolled in the study. A convenience sample of 6 commercial, free stall–housed dairy herds that were part of a larger ongoing field trial to investigate ketosis treatments was selected for the study reported here. The herds were chosen on the basis of their close proximity to the University of Guelph and were clients of a veterinary clinic that was interested in participating in the study. The herds ranged in size from 100 to 400 lactating cows, and all cows were fed a total mixed ration. Each herd was visited once weekly from May through August 2012. At each visit, cows expected to calve in the next 3 to 9 days were enrolled in the study; thus, the size of the study population was dictated by the number of cows that were eligible for enrollment during the 4-month sampling period.
Sample collection and processing—Blood samples (approx 10 mL) were collected by coccygeal venipuncture into evacuated serum-separator tubesb from all cows at study enrollment and at weekly intervals thereafter until 30 days after parturition. Immediately after collection of each blood sample, the blood tube was inverted to pool a small amount of whole blood into the cap for use for determination of BHBA concentration by a handheld device.c Once the blood sample was applied to the ketone strip used for the handheld device, the cap was placed back on the tube, and the remainder of the sample was placed on ice and transported back to the research facility.
Blood samples were centrifuged at the research facility. The serum was harvested from each sample, and the extent of hemolysis was qualitatively scored on the basis of the color of the sample on a scale of 0 (no color) to 3 (dark red). The serum samples were then frozen and stored at −20°C until analysis. Because hemolysis can falsely increase BHBA concentration,13 it was decided a priori to exclude all samples with a hemolysis score ≥ 2 from the analysis. None of the serum samples had a hemolysis score > 1, and all samples were included in the analysis. The prepartum serum samples were submitted to the University of Guelph Animal Health Laboratory, and serum BHBA and NEFA concentrations were determined by use of standard laboratory methods with an automated biochemistry analyzer.d
Statistical analysis—All analyses were performed with commercially available statistical software.e The correlation between BHBA concentration as determined by the handheld device (device BHBA concentration) and that as determined by standard laboratory methods (laboratory BHBA concentration) was evaluated by use of the Lin CCC14 and a Bland-Altman plot15 as described. Three cutoffs (≥ 0.6, ≥ 0.7, and ≥ 0.8 mmol/L) for prepartum laboratory BHBA concentration were selected on the basis of previously identified prepartum BHBA concentrations associated with postpartum disease and impaired performance.8,10,11 Each cutoff was used to model the outcome in a logistic regression equation that included prepartum device BHBA concentration, month of parturition, and herd; however, when the level of significance was set at α = 0.05, only device BHBA concentration was retained in the model. An ROC curve was then generated for each cutoff to determine the corresponding device BHBA concentration cutoff that optimized both sensitivity and specificity for the device measurement when the given cutoff was used as the gold standard for the diagnosis of hyperketonemia.16
Logistic regression was used to assess the association between hyperketonemia (BHBA concentration, ≥ 1.2 mmol/L) during the first week after parturition and prepartum BHBA concentration, both of which were measured with the handheld device. Fixed effects assessed in the model were postpartum BHBA concentration, month of parturition, prepartum NEFA concentration, and herd. Only variables with values of P ≤ 0.05 were retained in the final model.
The κ statistic was used to evaluate the level of agreement beyond chance between prepartum serum NEFA concentrations and prepartum BHBA concentrations for identification of cows at risk of developing postpartum hyperketonemia. The level of agreement was determined between each of 2 cutoffs for prepartum serum NEFA concentration (≥ 0.4 and ≥ 0.5 mEq/L) commonly associated with an increased risk of postpartum disease10 and recommended for use in prepartum monitoring programs17 and each of 3 cutoffs for prepartum BHBA concentration (≥ 0.6, ≥ 0.7, and ≥ 0.8 mmol/L) as determined by both standard laboratory methods and the handheld device. The level of agreement between prepartum NEFA and BHBA concentrations was classified as slight (κ, 0 to 0.2), fair (κ, > 0.2 to 0.4), moderate (κ, > 0.4 to 0.6), substantial (κ, > 0.6 to 0.8), or almost perfect (κ, > 0.8 to 1.0).18
Results
Cows—The number of cows sampled from each of the 6 herds ranged from 18 to 49. Blood samples were obtained from 211 cows. The prepartum BHBA concentration for 1 cow was below the lower limit of detection for the handheld device; therefore, the data for that cow were excluded from all analyses. For the 210 cows included in the analyses, the mean ± SD number of days between collection of the prepartum blood sample and parturition was 5.54 ± 2.23 days.
Correlation between device BHBA concentration and laboratory BHBA concentration—The mean ± SD prepartum laboratory BHBA concentration was 0.53 ± 0.25 mmol/L, and the mean ± SD prepartum device BHBA was 0.54 ± 0.27 mmol/L. Of the 210 cows evaluated, 55 (26%) and 88 (42%) had prepartum BHBA concentrations ≥ 0.6 mmol/L as determined by laboratory methods and the handheld device, respectively. Only 8 (4%) cows had a prepartum BHBA concentration ≥ 1 mmol/L, regardless of the analytic method used for measurement. The proportion of cows with a prepartum BHBA concentration ≥ 0.8 mmol/L varied among the 6 herds and ranged between 2% and 32% (Figure 1).
Proportion of cows with prepartum BHBA concentrations, as determined by use of a handheld device, < 0.60 mmol/L (white bars), between 0.60 and < 0.70 mmol/L (light gray bars), between 0.70 and < 0.80 mmol/L (dark gray bars), and ≥ 0.80 mmol/L (black bars) for each of 6 free stall–housed dairy herds in Ontario, Canada, and for the total study population. For each herd from May through August 2012, a blood sample was obtained from each cow between 3 and 9 days before its expected calving date. Blood samples were obtained from 210 cows (46 cows from herd A, 39 cows from herd B, 28 cows from herd C, 30 cows from herd D, 49 cows from herd E, and 18 cows from herd F).
Citation: Journal of the American Veterinary Medical Association 246, 10; 10.2460/javma.246.10.1112
Proportion of cows with prepartum BHBA concentrations, as determined by use of a handheld device, < 0.60 mmol/L (white bars), between 0.60 and < 0.70 mmol/L (light gray bars), between 0.70 and < 0.80 mmol/L (dark gray bars), and ≥ 0.80 mmol/L (black bars) for each of 6 free stall–housed dairy herds in Ontario, Canada, and for the total study population. For each herd from May through August 2012, a blood sample was obtained from each cow between 3 and 9 days before its expected calving date. Blood samples were obtained from 210 cows (46 cows from herd A, 39 cows from herd B, 28 cows from herd C, 30 cows from herd D, 49 cows from herd E, and 18 cows from herd F).
Citation: Journal of the American Veterinary Medical Association 246, 10; 10.2460/javma.246.10.1112
Proportion of cows with prepartum BHBA concentrations, as determined by use of a handheld device, < 0.60 mmol/L (white bars), between 0.60 and < 0.70 mmol/L (light gray bars), between 0.70 and < 0.80 mmol/L (dark gray bars), and ≥ 0.80 mmol/L (black bars) for each of 6 free stall–housed dairy herds in Ontario, Canada, and for the total study population. For each herd from May through August 2012, a blood sample was obtained from each cow between 3 and 9 days before its expected calving date. Blood samples were obtained from 210 cows (46 cows from herd A, 39 cows from herd B, 28 cows from herd C, 30 cows from herd D, 49 cows from herd E, and 18 cows from herd F).
Citation: Journal of the American Veterinary Medical Association 246, 10; 10.2460/javma.246.10.1112
The prepartum device BHBA concentration was moderately correlated with the prepartum laboratory BHBA concentration (mean ± SD Lin CCC, 0.77 ± 0.03 [95% CI, 0.72 to 0.83]). The difference between the mean prepartum device BHBA concentration and the mean prepartum laboratory BHBA concentration was small (0.004 mmol/L) and nonsignificant (P = 0.72). Evaluation of the Bland-Altman plot did not reveal any evidence of bias because the difference between the prepartum device and laboratory BHBA concentrations did not change significantly as the mean for the prepartum device and laboratory BHBA concentration changed (Figure 2). However, there were 2 outliers that may have influenced the overall correlation between the 2 concentrations. The cow with the largest discrepancy between the 2 tests had a prepartum device BHBA concentration of 0.4 mmol/L and a laboratory BHBA concentration of 2.1 mmol/L.
When laboratory BHBA concentration cutoffs of ≥ 0.6, ≥ 0.7, and ≥ 0.8 mmol/L were used as the standard against which the device BHBA concentrations were compared, the corresponding areas under the ROC curve for the optimum device BHBA concentration cutoff (ie, the cutoff that maximized the sensitivity and specificity of the device BHBA concentration) were very high (ie, > 0.90; Table 1). The optimum device BHBA concentration cutoff was equivalent to the laboratory BHBA concentration cutoffs of ≥ 0.6 and ≥ 0.7 mmol/L; however, for the laboratory BHBA concentration cutoff of ≥ 0.8 mmol/L, a device BHBA concentration cutoff of ≥ 0.7 mmol/L had better sensitivity and specificity than did a device BHBA concentration cutoff of ≥ 0.8 mmol/L. A device BHBA concentration cutoff of ≥ 0.8 mmol/L had a specificity of 94% for detection of a laboratory BHBA concentration cutoff of ≥ 0.8 mmol/L.
Area under the ROC curve, sensitivity, specificity, and associated 95% CIs for the optimum cutoff of prepartum BHBA concentration as determined by use of a handheld device (device BHBA concentration) when compared with various cutoffs of prepartum BHBA concentration as determined by use of standard laboratory methods (laboratory BHBA concentration; gold standard) to identify dairy cows at risk of developing hyperketonemia (BHBA concentration ≥ 1.2 mmol/L) during the first week after parturition.
| Laboratory BHBA concentration cutoff (mmol/L) | Optimum device BHBA concentration cutoff (mmol/L) | Area under the ROC curve (95% CI) | Sensitivity (95% CI) | Specificity (95% CI) |
|---|---|---|---|---|
| 0.6 | 0.6 | 0.902 (0.86–0.95) | 0.91 (0.80–0.97) | 0.76 (0.68–0.82) |
| 0.7 | 0.7 | 0.931 (0.88–0.99) | 0.85 (0.65–0.96) | 0.87 (0.81–0.92) |
| 0.8 | 0.7 | 0.929 (0.84–1.00) | 0.93 (0.66–0.99) | 0.83 (0.77–0.89) |
Blood samples were obtained from 210 dairy cows from 6 herds in Ontario, Canada, between 3 and 9 days prior to their expected calving date. The χ2 for each BHBA concentration cutoff comparison was significant (P < 0.001) as determined on the basis of a Wald test.
Association between hyperketonemia during the first week after parturition and prepartum BHBA concentration—The final multivariable model for hyperketonemia (BHBA concentration, ≥ 1.2 mmol/L) during the first week after parturition included prepartum BHBA concentration, prepartum serum NEFA concentration, and herd. When herd and prepartum serum NEFA concentration were controlled, cows with a prepartum device BHBA concentration ≥ 0.6 mmol/L were 2.2 times (95% CI, 1.0 to 4.8; P = 0.04) as likely to develop hyperketonemia within the first week after parturition, compared with cows with a prepartum device BHBA concentration < 0.6 mmol/L. Interestingly, hyperketonemia within the first week after parturition was not significantly associated with prepartum device BHBA concentration cutoffs of ≥ 0.7 and 0.8 mmol/L.
Bland-Altman plot of prepartum BHBA concentration as determined by a handheld device and standard laboratory methods for the cows of Figure 1. Each circle represents the results for 1 cow. The mean and 95% limits of agreement for the difference between the BHBA concentration as determined by the handheld device and by standard laboratory methods are represented by the middle and outer horizontal lines, respectively. Each circle outlined with a square indicates an outlier measurement. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 246, 10; 10.2460/javma.246.10.1112
Bland-Altman plot of prepartum BHBA concentration as determined by a handheld device and standard laboratory methods for the cows of Figure 1. Each circle represents the results for 1 cow. The mean and 95% limits of agreement for the difference between the BHBA concentration as determined by the handheld device and by standard laboratory methods are represented by the middle and outer horizontal lines, respectively. Each circle outlined with a square indicates an outlier measurement. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 246, 10; 10.2460/javma.246.10.1112
Bland-Altman plot of prepartum BHBA concentration as determined by a handheld device and standard laboratory methods for the cows of Figure 1. Each circle represents the results for 1 cow. The mean and 95% limits of agreement for the difference between the BHBA concentration as determined by the handheld device and by standard laboratory methods are represented by the middle and outer horizontal lines, respectively. Each circle outlined with a square indicates an outlier measurement. See Figure 1 for remainder of key.
Citation: Journal of the American Veterinary Medical Association 246, 10; 10.2460/javma.246.10.1112
Agreement beyond chance between prepartum serum NEFA concentration and prepartum BHBA concentration—The κ statistics calculated for each combination of 2 cutoffs for prepartum NEFA concentration and 3 cutoffs for prepartum BHBA concentration as determined by standard laboratory methods and the handheld device were summarized (Table 2). For identification of cows that developed hyperketonemia within 1 week after parturition, there was substantial agreement between prepartum NEFA concentration ≥ 0.5 mEq/L and both prepartum laboratory (κ, 0.62) and device (κ, 0.64) BHBA concentration ≥ 0.8 mmol/L.
Level of agreement beyond chance (κ statistic) between various cutoffs for prepartum serum NEFA concentration and prepartum BHBA concentration as determined by either standard laboratory methods or a handheld device for identification of cows at risk of developing postpartum hyperketonemia.
| κ | |||
|---|---|---|---|
| Serum NEFA concentration cutoff (mEq/L) | BHBA concentration cutoff (mmol/L) | Laboratory | Handheld device |
| 0.4 | 0.6 | 0.40 | 0.27 |
| 0.7 | 0.26 | 0.36 | |
| 0.8 | 0.39 | 0.49 | |
| 0.5 | 0.6 | 0.39 | 0.18 |
| 0.7 | 0.56 | 0.40 | |
| 0.8 | 0.62 | 0.64 |
A κ between 0 and 0.2 indicates slight agreement, between > 0.2 and 0.4 indicates fair agreement, between > 0.4 and 0.6 indicates moderate agreement, between > 0.6 and 0.8 indicates substantial agreement, and between > 0.8 and 1.0 indicates almost perfect agreement.
See Table 1 for remainder of key.
Discussion
Results of the present study indicated moderate agreement (Lin CCC, 0.77) between the BHBA concentration as determined by standard laboratory methods (ie, laboratory BHBA concentration) and the BHBA concentration as determined by use of a handheld device (ie, device BHBA concentration). In another study,12 BHBA concentration as determined by the same handheld device as used in the present study was highly correlated (Pearson correlation coefficient, 0.95) with that determined by standard laboratory methods. However, in that study,12 a larger population of cows was evaluated than in the present study, and all blood samples were collected from cows during the immediate postpartum period when blood BHBA concentration is expected to be higher than that during the prepartum period. The Pearson correlation coefficient does not account for accuracy19 and may slightly overestimate the correlation, compared with the Lin CCC. Also, the BHBA concentration data set for the present study included 2 outliers that might have affected the Lin CCC because they were well outside of the 95% limits of agreement on the Bland-Altman plot14 and occurred at fairly high BHBA concentrations. The largest difference between the device BHBA concentration and laboratory BHBA concentration for an individual cow was approximately 2.0 mmol/L, which was unusual for the study population and may have represented a reporting or sample-handling error.
The optimum cutoff for the prepartum device BHBA concentration for identification of cows at risk of developing hyperketonemia within 1 week after parturition was equivalent to that for prepartum laboratory BHBA concentration, except when the cutoff for the prepartum laboratory BHBA concentration was ≥ 0.8 mmol/L, at which the optimum cutoff for the prepartum device BHBA concentration was ≥ 0.7 mmol/L. When a prepartum laboratory BHBA concentration ≥ 0.8 mmol/L was used as the gold standard for identification of cows that developed postpartum hyperketonemia, a device BHBA concentration ≥ 0.7 mmol/L had a sensitivity of 93%, whereas a device BHBA concentration ≥ 0.8 mmol/L had a sensitivity of 79%; however, the specificity for a device BHBA concentration ≥ 0.7 mmol/L (83%) was lower than the specificity for a device BHBA concentration ≥ 0.8 mmol/L (93%). Generally, the cutoff used for identification of cows at risk of developing hyperketonemia within 1 week after parturition for the prepartum device BHBA concentration can be interchanged with that for the prepartum laboratory BHBA concentration. Although results of the present study suggested that a prepartum device BHBA concentration cutoff of ≥ 0.7 mmol/L should be used when a prepartum laboratory BHBA concentration ≥ 0.8 mmol/L is considered the gold standard for identification of cows at risk of developing hyperketonemia during the first week after parturition, the cutoff chosen for the prepartum device BHBA concentration should be made on the basis of whether maximal sensitivity or specificity is desired. In the present study, only 23 of 210 (11%) cows had prepartum BHBA concentrations ≥ 0.8 mmol/L, so small discrepancies between the device and laboratory BHBA concentrations might be overrepresented. The 2 outlier prepartum BHBA concentrations were both > 0.8 mmol/L. Additional prepartum blood samples with BHBA concentrations ≥ 0.8 mmol/L would need to be evaluated to assess the extent of agreement between the device and laboratory BHBA concentrations at high concentrations with greater precision.
For each laboratory BHBA concentration cutoff assessed (≥ 0.6, ≥ 0.7, and ≥ 0.8 mmol/L) for identification of cows that developed hyperketonemia within 1 week after parturition, the optimum cutoff for device BHBA concentration for the same purpose had a sensitivity that ranged between 85% and 93% and specificity that ranged between 76% and 87%. As the laboratory BHBA concentration increased, the specificity for the optimum cutoff for the device BHBA concentration likewise increased, whereas the sensitivity remained fairly stable. Results of Bland-Altman analysis indicated no bias between the device and laboratory BHBA concentrations; thus, the application of a correction factor to the prepartum device BHBA concentration is not necessary for interpretation. However, some consideration of the desired sensitivity and specificity for the device BHBA concentration in conjunction with the purpose of testing may be required.
A novel finding of the present study was the fact that cows with a prepartum device BHBA concentration ≥ 0.6 mmol/L were 2.2 times as likely to develop hyperketonemia within 1 week after parturition, compared with cows with a prepartum device BHBA concentration < 0.6 mmol/L. The lack of a significant association between the development of postpartum hyperketonemia and prepartum device BHBA concentration cutoffs ≥ 0.7 or ≥ 0.8 mmol/L was likely caused by a lack of power because of the small number of samples with BHBA concentrations that exceeded those cutoffs. A study in which the proportion of cows with a prepartum device BHBA concentration ≥ 0.7 mmol/L is greater than that of the present study is necessary to determine whether prepartum device BHBA concentration cutoffs ≥ 0.7 and ≥ 0.8 mmol/L are associated with the development of postpartum hyperketonemia.
The substantial agreement between prepartum serum NEFA concentration and prepartum BHBA concentration for identification of cows that develop hyperketonemia within 1 week after parturition was surprising because NEFA and BHBA result from separate biological processes. However, the prepartum NEFA and BHBA concentration cutoffs compared are associated with postpartum disease and impaired performance.10 Regardless, the results of the present study suggested that determination of prepartum BHBA concentration by use of the handheld device is a convenient alternative to determination of prepartum serum NEFA concentration by means of standard laboratory methods for identification of dairy cows at risk of developing postpartum hyperketonemia.
In the present study, prepartum device BHBA concentration was moderately correlated with prepartum laboratory BHBA concentration, and the sensitivity and specificity of prepartum device BHBA concentration cutoffs ≥ 0.6 mmol/L for identification of cows at risk of developing hyperketonemia during the first week after parturition were good when the gold standard was similar cutoffs for laboratory BHBA concentration. Thus, use of the handheld device evaluated in this study is a valid method for the measurement of BHBA concentration in dairy cows, and we recommend that a prepartum BHBA concentration ≥ 0.7 mmol/L as determined by the handheld device be used as a cutoff for identification of cows at risk of developing postpartum hyperketonemia. The use of this valid and cost-effective tool for identification of cows at risk of developing postpartum hyperketonemia should aid in the development and monitoring of health management programs and improve the timeliness of interventions such as the initiation of propylene glycol administration at the time of parturition.
ABBREVIATIONS
| BHBA | β-Hydroxybutyrate |
| CCC | Concordance correlation coefficient |
| CI | Confidence interval |
| NEFA | Nonesterified fatty acid |
| ROC | Receiver operating characteristic |
Gordon JL. Risk factors for and treatment of ketosis in lactating dairy cattle. DVSc thesis, University of Guelph, Guelph, ON, Canada, 2013.
Vaccutainer, BD Diagnostics, Mississauga, ON, Canada.
Precision Xtra, Abbott Laboratories Ltd, Saint-Laurent, QC, Canada.
Cobas 6000 c501 biochemistry analyzer, Roche Diagnostics, Laval, QC, Canada.
SAS, version 9.2, SAS Institute Inc, Cary, NC.
References
1. Grummer R. Impact of changes in organic nutrient metabolism on feeding the transition dairy cow. J Anim Sci 1995; 73: 2820–2833.
2. Hayirli A, Grummer RR & Nordheim EV, et al. Animal and dietary factors affecting feed intake during the prefresh transition period in Holsteins. J Dairy Sci 2002; 85: 3430–3443.
3. Herdt TH. Ruminant adaptation to negative energy balance: influences on the etiology of ketosis and fatty liver. Vet Clin North Am Food Anim Pract 2000; 16: 215–230.
4. Duffield TF, Lissemore KD & McBride BW, et al. Impact of hyperketonemia in early lactation dairy cows on health and production. J Dairy Sci 2009; 92: 571–580.
5. LeBlanc SJ, Leslie KE, Duffield TF. Metabolic predictors of displaced abomasum in dairy cattle. J Dairy Sci 2005; 88: 159–170.
6. Ospina PA, Nydam DV & Stokol T, et al. Evaluation of nonesterified fatty acids and β-hydroxybutyrate in transition dairy cattle in the northeastern United States: critical thresholds for prediction of clinical diseases. J Dairy Sci 2010; 93: 546–554.
7. McArt JAA, Nydam DV, Oetzel GR. Epidemiology of subclinical ketosis in early lactation dairy cattle. J Dairy Sci 2012; 95: 5056–5066.
8. Chapinal N, Carson ME & Duffield TF, et al. The association of serum metabolites with clinical disease during the transition period. J Dairy Sci 2011; 94: 4897–4903.
9. Ospina PA, Nydam DV & Stokol T, et al. Association between the proportion of sampled transition cows with increased nonesterified fatty acids and beta-hydroxybutyrate and disease incidence, pregnancy rate, and milk production at the herd level. J Dairy Sci 2010; 93: 3595–3601.
10. Chapinal N, LeBlanc SJ & Carson ME, et al. Herd-level association of serum metabolites in the transition period with disease, milk production, and early lactation reproductive performance. J Dairy Sci 2012; 95: 5676–5682.
11. Roberts T, Chapinal N & Leblanc SJ, et al. Metabolic parameters in transition cows as indicators for early-lactation culling risk. J Dairy Sci 2012; 95: 3057–3063.
12. Iwersen M, Falkenberg U & Voigtsberger R, et al. Evaluation of an electronic cowside test to detect subclinical ketosis in dairy cows. J Dairy Sci 2009; 92: 2618–2624.
13. Duffield TF, Sandals D & Leslie KE, et al. Efficacy of monensin for the prevention of subclinical ketosis in lactating dairy cows. J Dairy Sci 1998; 81: 2866–2873.
14. Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics 1989; 45: 255–268.
15. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986; 1: 307–310.
16. Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin Chem 1993; 39: 561–577.
17. Oetzel GR. Monitoring and testing dairy herds for metabolic disease. Vet Clin North Am Food Anim Pract 2004; 20: 651–674.
18. McGinn T, Wyer PC & Newman TB, et al. Tips for learners of evidence-based medicine: 3. Measures of observer variability (kappa statistic). CMAJ 2004; 171: 1369–1373.
19. King TS, Chinchilli VM. Robust estimators of the concordance correlation coefficient. J Biopharm Stat 2001; 11: 83–105.
