Pain detection of cows with mastitis in the barn and at milking: a piece of cake?

International audience

anglais. Objectives: Pain is one of the most detrimental factors affecting dairy cows’ welfare and induces production losses. Mastitis is a good model for pain research in cattle, as it is a commonly occurring inflammatory painful disease. When induced by LPS infusion in the udder, mastitis has a limited duration, and non-steroidal anti-inflammatory drugs (NSAIDs) can easily modulate it. To date, there is a lack of knowledge on the behavioral reaction of mastitic cows in cubicles and at milking. This study aimed at refining behavioral indicators of pain in dairy cows with mastitis, according to time and whether cows were observed in the cubicle barn or at milking. Materials & Methods: Twenty-seven cows received an intra-mammary infusion of 25 μg E. coli LPS in one healthy quarter. Thirteen cows received 3 mg/kg ketoprofen IM (Ketofen® 10%) in addition to LPS (LPS+NSAID cows), and 14 cows received placebo (20mL NaCl 0.9% IM) instead of NSAID (LPS cows). Evolution of the local immune response was assessed using somatic cell counts (SCC) and cytokines/chemokines quantification. Cows’ response to the challenge was monitored at regular intervals from 24 hours before to 48 hours post-infusion (hpi) through direct clinical observations (cardiac, ruminal, and respiratory frequencies), indicators of inflammation (in milk: cytokines/chemokines, haptoglobin, serum amyloid A (SAA)), and stress (cortisol in blood and milk). Direct behavioral observations were recorded in the barn (i.e., postures, activity, social behavior, ear positions, etc.) and at milking (i.e., steps, lifts, kicks). Results: In LPS cows, infusion induced significant changes in endocrinal, inflammatory, and clinical parameters. Blood cortisol peaked at (mean [95% CI]) 69.18 [48.98-97.72] ng/ml at 3 hpi and remained above 65 ng/ml at 7 hpi. Milk cortisol peaked at 1.69 [1-3.01] ng/ml at 7 hpi. SCC peaked at 1.8 × 10⁷ cells/ml at 7 hpi and remained above 8 × 10⁶ cells/ml until 48 hpi. IL-6, IL-1β, and CXCL8 peaked at 7 hpi (respectively: 1.9 [1.7-3.0] × 10³ ng/ml; 1.6 [0.7-2.3] × 10² ng/ml; 2.7 [1.7-7.0] × 10³ ng/ml). SAA in milk significantly increased at 7 hpi, 24 hpi, and 31 hpi, peaking at 7.4 [4.5-12.8] × 10⁴ ng/ml at 48 hpi. More LPS cows stopped feeding/ruminating and pressed their tail between their legs at 3, 5, and 24 hpi than before (P < 0.05 in all cases). They also tended to be more apathetic, dropping their head and ears at 5 hpi (P = 0.07). At milking, significantly more LPS cows lifted their hooves at forestripping at 7 hpi than before (P = 0.04). LPS+NSAID and LPS cows showed similar patterns of responses for milk cortisol, SCC, respiratory rate, mAA, haptoglobin, IL-6, IL-1β, and CXCL8 (P > 0.05). Compared to LPS cows, LPS+NSAID cows had significantly lower plasmatic cortisol levels at 3 hpi (61.66 [24.57-144.54] ng/ml). At 7 hpi, LPS+NSAID cows' rectal temperature decreased by 0.66 [(-0.85)-(-0.48)] °C, and their rumen motility rate increased by 1.04 [0.34-1.75] contractions/min. At 31 hpi, their heart rate increased by 8.44 [1.00-15.91] beats/min, and rumen motility rate increased by 0.89 [0.19-1.60] contractions/min. Compared to LPS cows, a larger proportion of LPS+NSAID cows were feeding or ruminating (P = 0.02), fewer had ears down (P = 0.04) at 5 hpi, and they were more likely to lie down at 24 hpi (P = 0.02). Conclusion: This study confirms the benefits of using behavioral indicators to detect pain associated with mastitis in dairy cows, particularly during milking. These behavioral indicators (feeding/ruminating, tail position, reactivity at forestripping) could be useful for early detection of mastitis and for decision-making regarding the initiation of pain-relief treatment, contributing to animal welfare improvement.