Département de Sciences Cliniques, Faculté de Médecine Vétérinaire, Université de Montréal, 3200 rue Sicotte, St-Hyacinthe, Québec, J2S 2M2, Canada.
Département de Sciences Cliniques, Faculté de Médecine Vétérinaire, Université de Montréal, 3200 rue Sicotte, St-Hyacinthe, Québec, J2S 2M2, Canada.
J Dairy Sci. 2020 Apr;103(4):3381-3391. doi: 10.3168/jds.2019-17129. Epub 2020 Feb 11.
Digital dermatitis (DD) is a worldwide infectious disease of cattle that causes lameness, discomfort, and economic losses. The reference standard test to diagnose DD is visual observation in a trimming chute, which cannot be practically performed daily on dairy farms. Moreover, some lesion misclassification may occur using this standard diagnostic method. The possibility of misclassification makes the use of a trimming chute debatable as a perfect reference standard test. The objective of this study was, therefore, to assess the diagnostic accuracy of a commercial borescope and trimming chute exam. The accuracy (sensitivity and specificity) of the tests and DD prevalence were investigated using Bayesian latent class analyses. Our hypothesis was that a commercial borescope can be routinely used to diagnose DD in a milking parlor without previous feet cleaning. A cross-sectional study was performed in a freestall facility. The lesions were scored (M0, M1, M2, M3, M4, M4.1) in the milking parlor with a borescope, followed by an examination in the trimming chute 48 to 72 h after the borescope exam. A total of 870 hind feet were scored during 2 sessions of trimming chute exams and borescope exams in the milking parlor. The data were analyzed in 2 ways. First, data were dichotomized into DD lesions (M1, M2, M3, M4, M4.1) and absence of DD lesions (M0). Second, data were dichotomized into active DD lesions (M1, M2, M4.1) and inactive lesions (M0, M3, M4). A Bayesian latent class model allowing for conditional dependence between tests was used to estimate tests' accuracy, likelihood ratio, and DD prevalence. When the data were dichotomized into DD lesions (M1-M4.1) versus absence of DD (M0) lesions, the sensitivity and specificity of the borescope was 55% [95% credible interval (CrI) 40-71%] and 81% (95% CrI 75-88%). The sensitivity of trimming chute exams was 79% (95% CrI 68-88%), and specificity was 80% (95% CrI 71-89%). When the data were dichotomized into active lesions (M1, M2, M4.1) versus inactive lesions or absence of lesions (M3, M4, M0), the sensitivity and specificity of the borescope were, respectively, 32% (95% CrI 13-58%) and 91% (95% CrI 88-95%). The sensitivity and specificity of trimming chute exams were 91% (95% CrI 81-97%) and 81% (95% CrI 75-89%), respectively. In conclusion, it is possible to use the borescope in the milking parlor without cleaning the feet to monitor prevalence of DD lesions. However, an isolated borescope examination, especially for diagnosing active DD lesions, has low sensitivity for use as a surveillance method. For such use, the sensitivity could be improved by repeating the borescope exam on a regular basis.
数字性皮炎（DD）是一种全球性的牛传染病，会导致跛行、不适和经济损失。诊断 DD 的参考标准测试是在修剪滑道中进行目视观察，但在奶牛场实际上无法每天进行。此外，使用这种标准诊断方法可能会出现一些病变分类错误。这种分类错误的可能性使得修剪滑道作为完美的参考标准测试存在争议。因此，本研究的目的是评估商业内窥镜和修剪滑道检查的诊断准确性。使用贝叶斯潜在类别分析评估了测试的准确性（敏感性和特异性）和 DD 的流行率。我们的假设是，可以在不事先清洁脚部的情况下，使用商业内窥镜在挤奶厅常规诊断 DD。这是一项在一个自由站立设施中进行的横断面研究。使用内窥镜在挤奶厅对病变进行评分（M0、M1、M2、M3、M4、M4.1），然后在内窥镜检查后 48 至 72 小时在修剪滑道中进行检查。在两次修剪滑道检查和挤奶厅内窥镜检查中总共对 870 个后脚进行了评分。数据以两种方式进行分析。首先，数据被分为 DD 病变（M1、M2、M3、M4、M4.1）和无 DD 病变（M0）。其次，数据被分为活跃 DD 病变（M1、M2、M4.1）和不活跃病变（M0、M3、M4）。使用允许测试之间条件依赖的贝叶斯潜在类别模型来估计测试的准确性、似然比和 DD 的流行率。当数据被分为 DD 病变（M1-M4.1）与无 DD（M0）病变时，内窥镜的敏感性和特异性分别为 55%（95%可信区间[CrI]为 40-71%）和 81%（95% CrI 为 75-88%）。修剪滑道检查的敏感性为 79%（95% CrI 为 68-88%），特异性为 80%（95% CrI 为 71-89%）。当数据被分为活跃病变（M1、M2、M4.1）与不活跃病变或无病变（M3、M4、M0）时，内窥镜的敏感性和特异性分别为 32%（95% CrI 为 13-58%）和 91%（95% CrI 为 88-95%）。修剪滑道检查的敏感性和特异性分别为 91%（95% CrI 为 81-97%）和 81%（95% CrI 为 75-89%）。总之，在不清洁脚部的情况下，可以在内窥镜检查中使用挤奶厅来监测 DD 病变的流行率。然而，作为一种监测方法，单独使用内窥镜检查，特别是用于诊断活跃的 DD 病变，其敏感性较低。为了提高这种用途的敏感性，可以定期重复进行内窥镜检查。
