Development of a selection index for resistance to subclinical ketosis in Holstein Friesian dairy cows

At the onset of lactation, high-yielding dairy cows could often experience a period of negative energy balance. This is reflected in a loss of body condition, due to body fat mobilization, and an increase of circulating ketone bodies, particularly β-hydroxybutyrate (BHB). This condition, known as hyperketonaemia, can result in (sub)clinical ketosis with negative implications on cow productivity and functionality, including health and fertility. The objective of the present study was to develop a genetic evaluation of resistance to subclinical ketosis for Holstein dairy cattle using data routinely available from the national milk recording system and linear classification. Milk BHB and fat-to-protein ratio (FPR) was available on more than 2.2 million test-days records belonging to Holstein cows in the first 90 days-in-milk from first up to the third lactation. These records were subsequently matched to the closest linear classification date when body condition score (BCS) was measured by an expert evaluator. The pedigree of cows has traced back up to 6 generations. (Co)variance components were estimated using trivariate linear mixed models; in particular, for BHB and FPR the fixed effects of herd-test-day, the two-way interaction between the week of lactation and parity, and the three-way interaction between classes of age at calving, parity and year of calving were considered. The linear model for BCS included the fixed effects of herd-year-round of classification, year of calving and the two-way interaction between age at calving and stage of lactation. The additive genetic effect and, only for BHB and FPR, the permanent environment were the two random terms. Due to computational constraints, (co)variance components were estimated on ten different subsets including 400 herds each, and subsequently averaged. Milk BHB and FPR and BCS averaged 0.056, 1.152 and 2.99, respectively. Heritability estimates were 0.093, 0.090 and 0.157 for BHB, PFR and BCS, while repeatability estimates were 0.179 (BHB) and 0.209 (FPR). The genetic (phenotypic in parenthesis) correlations were 0.159 (0.279; BHB vs. FRP), −0.161 (−0.038; BHB vs. BCS) and −0.140 (−0.049; FPR vs. BCS). The present study suggests that an exploitable additive genetic variation exists for milk BHB, and it could be used to set up breeding strategies aiming at improving resistance to subclinical ketosis through genetic selection.