Orthopedic disease, presenting as lameness, is a common cause of pain in dairy cattle, often implying a prolonged course of disease, with significant economic losses and impaired animal welfare. To mitigate these negative effects, early identification of lameness and monitoring of pain levels during recovery are crucial. This study aimed to evaluate whether the Cow Pain Scale (CPS) can be used to detect pain behaviors in stationary dairy cows with mild to moderate lameness or not, and to investigate the association between certain CPS items and orthopedic disease. Data were collected on a research dairy farm with a loose-housing system and included 36 clinical lameness cases from 34 individual cows. Each lameness case consisted of 2 trials: when the cow was found to be lame (initial trial), and after treatment and improvement (follow-up trial). Each trial included an on-site pain assessment with the CPS with simultaneous video recording, followed by Sprecher lameness scoring and clinical examination. Blinded pain assessments using the CPS were performed from the video recordings by 3 trained observers at 2 different occasions, 3 wk apart. Using linear mixed models, a significant positive correlation between CPS total scores and Sprecher lameness scores was identified for both video and on-site pain assessments. Predicted CPS total scores increased with increasing lameness scores
however, there was a large overlap in CI, indicating a complex relationship between pain score and pain intensity. Multiple correspondence analysis identified different CPS item combinations, including facial expressions, which were associated with orthopedic pain. Video scoring showed moderate inter- and good intrarater agreement in CPS total scores, but there was considerable variation in inter- and intrarater agreement for different scale items. The scale items "head position" and "back position" showed an overall strong agreement and "attention toward the surroundings" a moderate agreement. In contrast, intrarater agreement for the items "facial expression" and "ear position" was generally weak, which may indicate that these items are harder to assess. The CPS proved to be a reliable tool for video pain assessment in dairy cows with orthopedic pain. Furthermore, we showed that behavioral patterns varied among lame cows, which is why single items within the CPS should be interpreted with caution.