Association between lying duration and hoof lesions in lactating dairy cows

Introduction
Lameness is an important disease of dairy cattle associated with pain and change in normal behavior (Chapinal et al., 2010). Lying behavior is a vital indicator of cow welfare (Norring et al., 2014), and may play a significant role in the etiology of lameness (Weigele et al., 2017). Hoof lesions (HL) are the major cause of lameness in dairy cattle however; there is limited information about the behavior of dairy cows that develop HL. Previous work that evaluated lying behavior in dairy cows focused on differences between healthy and already lame cows. The objective of this study was to evaluate the association between lying behavior and the development of HL in lactating dairy cows. 

Materials and Methods
This study was conducted in a commercial dairy herd milking 3x with a total of 10,000 Jersey cows in MN. Cows were housed in a cross-ventilated free-stall barn with stalls bedded with recycled sand. Cows were selected for this study from a concurrent study that used convenience sample of cows as they exited the rotary parlor. From the concurrent study 435 cows without any visible HL on all four legs were enrolled at 20 ± 3 DIM (d20) and a pedometer (AfiTagII) was secured to the right hind leg of each cow. At 120 ± 3 DIM (d120), cows were reexamined for HL and pedometers were removed. Cows that left the study (culled, moved or died) before lesion status could be determined at d120 were excluded from further analyses. Lying duration data of first day after enrolment were considered outliers and were not used for model fitting. Separate models were built using each outcome of lying behavior of interest and included the fixed effects of lesion status at d120 (healthy, lesion), DIM, interaction between lesion status at d120 and DIM, parity (change in BCS, and season of calving. Data was analyzed using restricted maximum linear growth mixed measures model with an autoregressive covariance structure. In each model, cow was included as random effect. Parity was forced into all final models 

Results
Due to data back failures only data from 344 cows with a minimum of 24,813 days of data were available for analysis. At d120, 58.4% of cows were diagnosed with a HL. Of the total lesions diagnosed at d120, the majority were sole hemorrhage (93%). Fifteen cows were diagnosed with non-infectious lesions sole ulcer and white line disease, whereas eight cows had infectious lesion digital dermatitis at d120. At enrollment, the cows that developed HL already had reduced lying duration -0.5 units (95% Confidence Interval (CI): -0.78 to -0.28) compared with healthy cows. Also, there was evidence for an interaction between DIM with lesion status 0.003 (95% CI: 0.002 to 0.005). Irrespective of lesion status at d120, average lying duration increased progressively for both cows that remained healthy and herdmates that developed HL. 

Conclusions
Results from this study showed that lying duration was already decreased in cows that went on to develop a HL a 100 days prior to HL diagnosis. This suggest that the causative mechanism for HL development might already have occurred prior to enrolment in our study.