| dc.description.abstract | Ketosis in dairy cattle is defined as an increase in the circulating ketone bodies. It has negative effects on milk production and is a risk factor for many diseases related to the transition period. In the Netherlands, a relatively new method of detecting acetone in milk recently became included in the regular milk production registration (MPR). The objective of this study was to determine if prepartum walking and eating activity of dairy cattle could be used as an early indicator of receiving a ketosis attention on the MPR result form in the first 60 days postpartum. The prepartum walking activity (steps/d) and eating activity (h/d) of respectively 302 and 310 dairy cows, monitored by Nedap SmartTag sensors during the last four weeks prepartum, was analysed. The dairy cows were divided in a ketosis group and a no ketosis group. The means, regression coefficients and residual standard deviations of three prepartum periods (first period d -28 to -15; second period d -14 to -3; third period d -3 to 0) were used as independent variables in logistic regression analysis (backward model reduction). The mean walking activity of the total prepartum period (steps/d ± SD; d -28 to 0) differed significantly between the ketosis group and the no ketosis group (𝑥̅k = 2665 ± 902; 𝑥̅no = 3000 ± 819). In logistic regression, the mean (steps/d x1000) of the first period (p = 0.001) and the residual standard deviation of the third period (p = 0.037) were found to be significant model parameters. The odds ratios and the corresponding confidence intervals of the two model parameters were respectively 0.473 [0.299, 0.749] and 1.473 [1.025, 2.119]. In contrast to walking activity, the mean eating activity of the total prepartum period (h/d ± SD; d -28 to 0) did not differ significantly (p = 0.087) between the ketosis and the no ketosis group (𝑥̅k = 5.89 ± 1.29; 𝑥̅no = 6.22 ± 1.04). The remaining model parameters in logistic regression were the residual standard variation of the first period (p = 0.055) and the mean of the second period (p = 0.093). The odds ratios were respectively 0.292 [0.083, 1.027] and 0.812 [0.637, 1.035]. With respect to the 95% confidence interval, neither of the remaining parameters of prepartum eating activity had a statistically significant association with postpartum ketosis attentions, although a clear trend was established. The sensitivity and specificity of both prepartum walking and prepartum eating activity are not yet high enough to be useful as early indicators of a ketosis attention postpartum. However, the trends that were established in this study between both prepartum walking and eating activity and postpartum ketosis attentions, might contribute to further research into prevention and control of transition period related diseases such as ketosis. | |
