Estimations of prepartum feed intake and its effects on transition metabolism and subsequent milk production

The objectives of this study were to identify factors associated with prepartum DMI, evaluate the performance of linear models to estimate prepartum DMI using different classes of predictors, and investigate the consequences of different levels of prepartum DMI on transition metabolism and lactation performance. Individual feed intake of nulliparous (n = 100) and parous cows (n = 173) was measured by automatic feeding bins from d -35 to 98 relative to calving. Rumination and physical activities were monitored by wearable sensors. Blood metabolites were measured on d -21, -10, -3, 0, 3, 7, 10, 14, and 21. Body weight and BCS were assessed throughout the study. The average prepartum DMI as percentage of BW (DMIpBW) was calculated for each cow and used as dependent variable of linear models. Parity, prepartum BCS and BW, 305-d milk production in the previous lactation, milk yield at dry-off (MYDO), and the length of the dry period were associated with DMIpBW. These factors explained 41% of the variation in DMIpBW across all cows and 49% in parous cows.. Estimations of DMIpBW were improved when data on prepartum rumination and blood metabolites were added into the predictive models. In the latter, the adjusted R² increased to values between 47% and 61%, and selected models performed consistently in a 5-fold cross-validation analysis. To evaluate the implications of DMIpBW to transition metabolism and performance, cows were ranked within parity and classified into terciles as low (LFI), moderate (MFI), or high feed intake (HFI). The mean DMI was 1.44%, 1.70%, and 1.91% ± 0.01% of BW, respectively. No differences in BW were observed in nulliparous cows, but all 3 groups of parous cows differed (LFI = 892, MFI = 849, HFI = 798 ± 8 kg). The proportion of cows with BCS >3.5 at enrollment differed among all groups, and averaged 67.4%, 55.1%, and 36.5% ± 6%, respectively. For parous cows, 305-d milk production and MYDO differed among all groups and averaged 9,808, 10,457, and 11,182 ± 233 kg, and 18.1, 23.1, and 26.2 ± 1 kg/d, respectively. After calving, DMI (LFI = 20.9, MFI = 21.9, and HFI = 22.1 ± 0.2 kg/d) and milk yield (LFI = 36.7, MFI = 38.2, and HFI = 38.3 ± 0.4 kg/d) was lower in LFI cows compared with the other 2 groups. Postpartum energy balance differed among all groups and averaged -2.79, -1.63, and -0.66 ± 0.3 Mcal/d for LFI, MFI, and HFI, respectively. During the transition period, LFI cows had higher serum concentrations of nonesterified fatty acids, BHB, Cl (prepartum only), and aspartate aminotransferase (AST; postpartum only), and lower serum concentrations of cholesterol, P, glutamate dehydrogenase, gamma-glutamyl transferase (prepartum only), AST (prepartum only), urea (parous only), and superoxide dismutase activity (parous only). In conclusion, a low level of prepartum DMI was associated with fatter and heavier cows, lower milk production in previous lactation, important adjustments in energy metabolism during the transition period, and moderate losses in DMI and milk yield in the subsequent lactation. Moreover, the inclusion of prepartum rumination activity and target blood metabolites into predictive models improved the estimations of prepartum DMI.