Influence of common health disorders on the length of productive life and stayability in German Holstein cows.

The aim of this study was to infer phenotypic and genetic effects of health disorders on longevity traits, considering Holstein dairy cow records from large-scale co-operator herds. In this regard, we focused on 13 different disease traits and on 2 longevity definitions: length of productive life (LPL) and stayability (STAY). The LPL was defined as the interval in days from first calving to culling. For LPL, we considered 90,215 cows with known culling dates. For binary STAY, we defined 3 survival stages in the first 3 lactations: from calving to DIM 59, from DIM 60 to DIM 299, and from DIM 300 to the next calving date. Due to the earlier trait recording possibilities, 129,386 cows were considered for the STAY analysis. Accordingly, the presence or absence of diseases in lactation stages were defined as binary traits. A further data set for the 90,215 cows with a culling date included the subjective culling reasons defined by farmers. Comparison of culling reasons, as defined by farmers, with diagnoses from the disease data set indicated some disagreements. For example, only 18.71% of the cows with the farmer culling reason "metabolic diseases" were diagnosed with a metabolic disorder. Better agreements were identified for mastitis (84.09%). Phenotypically, in most cases, occurrence of diseases at different lactation stages had negative influence on LPL and STAY. In this regard, we identified strong detrimental effects of clinical mastitis and of metabolic disorders from early lactation stages on longevity traits. For example, the presence of clinical mastitis in the first stage of first lactation was associated with LPL decrease of 95.35 d. Using generalized linear mixed models for binary health disorders, heritabilities ranged from <0.01 (±0.079 standard error) for ruminal acidosis early in first, second, and third lactation to 0.24 (±0.039) for interdigital hyperplasia from the last stage in third lactation. Heritabilities from single-trait and bivariate animal models ranged from 0.03 (±0.003) to 0.10 (±0.007) for LPL, and from 0.01 (±0.002) to 0.06 (±0.007) for STAY. Genetic correlations between longevity traits and health disorders were mostly negative (i.e., favorable in a breeding sense). For improvements to longevity genetic evaluations for young bulls with a limited number of daughter culling dates, we suggest consideration of health traits from a well-organized co-operator herd monitoring system as early longevity predictors, especially for censored data. Genetic correlations between mastitis from different lactation stages with LPL and STAY ranged from -0.28 (±0.07) to -0.69 (±0.05), and from -0.26 (±0.08) to -0.77 (±0.08), respectively. Interestingly, only diagnoses for dermatitis digitalis showed opposite results phenotypically and genetically. Strong genetic associations between ruminal acidosis and STAY were observed (genetic correlations: -0.48 ± 0.18 to -0.98 ± 0.31), supporting the inferred phenotypic associations. Genetic correlations between longevity traits LPL and STAY were quite large, between 0.77 (±0.11) and 0.94 (±0.02) for the different lactation stages, suggesting utility of early STAY information when attempting genetic improvements for longevity.