Can automated measures of lying time help assess lameness and leg lesions on tie-stall dairy farms?

Abstract

a b s t r a c tThe time that dairy cows spend lying down is an important measure of their comfort and lameness and injuries to hocks and knees are associated with alterations in lying time. We examined whether automated measures of lying time could identify cows and farms with problems of lameness or leg lesions. Data were collected from 40 lactating Holstein dairy cows from each of 100 tie-stall farms. The occurrence of lameness, hock and knee injuries was recorded and lying times were recorded automatically using accelerometers. There was large variation between individual cows, and between farms in all measures of lying time. At the cow level, there was no relationship (P > 0.10) between being lame and daily duration of lying time. A lower daily duration of lying time was found among cows with hock injuries (mean ± SE: non-injured = 12.79 ± 0.06 h, injured = 12.21 ± 0.06 h; P < 0.001) and cows with knee injuries (mean ± SE: non-injured = 12.54 ± 0.05 h, injured = 12.25 ± 0.06 h; P = 0.04) than those without lesions. The median daily duration of lying time on a farm was negatively correlated with the prevalence of lameness (r p = −0.27, P = 0.006), of hock injuries (r p = −0.35, P = 0.003) and of knee injuries (r p = −0.28, P = 0.004). A canonical discriminant function with canonical coefficients of 0.63 for mean daily duration of lying down, and of 0.54 for mean bout frequency could correctly identify 72% of the farms that were above the median for percent of cows with hock or knee injuries or being lame (linear discriminant function: constant = −95.10, daily duration = 11.39, bout frequency = 4.83) and 68% of the farms below the median (linear discriminant function: constant = −106.95, daily duration = 11.92, bout frequency = 5.30) (Wilks Lambda test P = 0.002). A criterion of a median lying time between 12 h and 13 h alone could identify over 60% of farms above or below the median for lameness, hock and knee lesion prevalence. Automated measures of lying time may be a useful animal-based measure to indicate farms with a high percentage of lame cows or cows with leg lesions.

Cow- and farm-level risk factors for lameness on dairy farms with automated milking systems

“…The mean prevalence of lameness in our study was 15%. This is similar to the 14% reported for AMS farms in Denmark (Klaas et al, 2003), but lower than lameness prevalence on 141 Canadian freestall farms with a milking parlor (21%; Solano et al, 2015) and 100 Canadian tiestall farms (24%; Charlton et al, 2015). The latter 2 studies followed the same data collection protocol as described in the current paper.…”

Invited review: Lying time and the welfare of dairy cows

“…The mean daily duration of time spent lying down is around 11 h/d but varies among cows from less than 6 h/d to more than 16 h/d (Figure 1). Ito et al (2009) Canada 43 2,033 11.0 9.5-12.9 Thomsen et al (2012) Denmark 39 1,233 10.7 Wechsler et al (2000) Switzerland 18 10/farm 11.3 Gomez and Cook (2010) US 16~13/farm 11.9 Yunta et al (2012) Spain 10 40/farm 12.0 Thorup et al(2015) Denmark 4 200/farm 11.4 Solano et al (2016) Canada 141 40/farm 10.6 8.2-13.2 Freestall with AMS 2 Westin et al (2016) Canada 36 40/farm 11.3 9.7-13.5 Deming et al (2013) Canada 13 30/farm 10.8 Helmreich et al 2014Switzerland 4 138 10.6 9.4-11.5 King et al (2016) Canada 41 30/farm 11.5 10.1-11.7 Tiestall Charlton et al (2016) Canada 100 40/farm 12.5 10.0-15.0 Dry lots Steensels et al (2012) Israel 3 210 8.9 Tresoldi et al (2017) US 10 10-15/farm 9.0 7.4-10.6 Deep-bedded pack Endres and Barberg (2007) US 12 147 9.3 6.1-11.4 Pasture Sepúlveda-Varas et al (2014) Chile 6 274 8.5 Beggs et al (2018) Australia 10 30/farm 9.5 1 Only includes studies that examined multiple farms.…”

Invited review: Cattle lameness detection with accelerometers

“…However, in some studies, lying variables have been reported to not differ to a statistically significant extent. For example, daily lying time was only marginally different between lame and nonlame cows in 3 studies (Chapinal et al, 2011;Charlton et al, 2016;Blackie and Maclaurin, 2019). Longer lying bouts by lame cows have generally been reported (King et al, 2017;Schindhelm et al, 2017;Weigele et al, 2018), whereas one study with Jerseys reported shorter bouts (Blackie and Maclaurin, 2019).…”

“…If 2 or more models of the 9 indicated lameness, an alert was triggered (de Mol et al, 2013). Two studies reported using differing forms of discriminant analysis, partial least squares discriminant analysis (Garcia et al, 2014), and canonical discriminant analysis (Charlton et al, 2016). Most of the reviewed studies that reported classifying cows as lame or not reported using a binary output.…”