A mixed-effects time-to-event analysis of the relationship between first-lactation lameness and subsequent lameness in dairy cows in the UK

Data for 611 second-lactation and 251 third-lactation cows were examined using mixed-effects time-to-event models to determine the shape of the hazard, quantify relative risk and estimate herd- and sire-level variation in time to lameness. The semi-parametric Cox and fully parametric Weibull models were suggested from univariable Kaplan-Meier plots. Time to all-lameness, claw-horn lameness and skin lameness were modelled. Explanatory variables were season of current-lactation calving, age at first calving and first-lactation lameness history (whether all-lameness or claw-horn and skin lameness). In mixed-effects models of lactation-2 lameness, previously lame cows had a significantly increased hazard (hazard ratio (HR) = 2.0 for all types of lameness and HR = 3.2 for claw-horn lameness) compared to those not previously lame. These relationships were less marked in the third lactation. There was little evidence for an effect of age at first calving, whilst possible differences between calving seasons were observed. The hazard function suggested that the rate of lameness was roughly flat across each lactation. Herd-level variation was more evident for infectious foot diseases. The contribution of the sire increased with parity and might be important for sole ulcer and white-line disease. (C) 2002 Elsevier Science B.V. All rights reserved.