• Fertility is a key driver of productive lifespan in Swiss dairy cows. • Fat-protein yield is associated with longevity, highlighting its economic value. • Production, fertility, and health traits explain 20–40% of lifespan variation. • Good udder and conformation scores are linked to higher survival rates. • Longevity depends on a complex interplay of fertility, productivity, and health. Based on herd book data from over 2.44 million Swiss dairy cows with approximately 7.33 million lactations (01 January 1999 – 31 August 2019), culling risk factors were analyzed using Weibull proportional hazard models. Time-dependent fixed effects included fat-protein production level compared to herd mates, number of inseminations, calving interval, calving ease, lactation cell count, herd size, herd size changes, calving season, and production zone. Time-independent fixed effects comprised age at first calving, farm changes between first and last lactation, and the origin of replacement heifers, while herd-year–season was included as random effect. All traits significantly affected culling risk. Together, the investigated traits improved model fit substantially, with Maddala’s R² values ranging between 0.20 and 0.40 for productive lifespan. However, most individual trait effects were small in magnitude despite statistical significance. The number of inseminations accounted for the largest proportion of PL variation across all breeds, followed, in most breeds, by the cow’s fat-protein yield relative to her herd mates. The importance of other traits varied by breed. Linear composite traits explained only a small proportion of PL variation (1.2 to 3.7%), with a good udder score improving longevity. Cows with a high final linear score had longer PL, though its explanatory power was low (0.7 to 2.5%), and few cows were scored at the extremes . The findings highlight the critical role of fertility management for longevity and the economic importance of productivity for cow survival, while emphasizing that longevity is influenced by many factors with generally small individual contributions.
Bieber et al. (Fri,) studied this question.