Abstract #T226

# T226
Relative reticulo-rumen pH indicators for subacute ruminal acidosis detection in dairy cows.
C. Villot*1, B. Meunier1, J. Bodin2, C. Martin1, M. Silberberg1, 1Institut national de la recherche agronomique, UMR1213 Herbivores, Saint-Genès-Champanelle, France, 2BR3 Consultants, Lyon, France.

Subacute ruminal acidosis (SARA) in dairy cows is a nutritional disease difficult to diagnose on field due to a lack of specific clinical signs. Abnormal and intermittent drops in rumen pH remain relevant observations to characterize SARA. The aim of this study was to assess the ability of different pH indicators calculated from rumen pH kinetics to detect SARA. Eleven primiparous dairy cows were submitted to a long-term feed challenge to experimentally induce SARA. Animals were consecutively fed 2 different diets: a low starch control diet (LSD; 13% starch, 4 weeks), a high-starch acidotic diet (HSD; 32% starch, 4 weeks), and again the LSD (3 weeks). Rumen pH was monitored continuously using a reticulo-rumen sensor (eCow bolus). Commonly used pH indicators of SARA were daily average, area under the curve (AUC), time spent under pH < 5.8 and pH < 6.0. Rumen pH was also processed to remove sensor drift (8 weeks filter) and noise (48h filter) and normalized on 0 to suppress inter-individual variability. Then relative pH indicators such as normalized AUC, times spent under pH < −0.3 and pH < −0.5 (drop of 0.3 or 0.5 pH unit regarding normalized pH on 0), pH range and pH standard deviation were calculated and compared with commonly used pH indicators. SARA syndrome induced by HSD was confirmed by consistent expected changes in animal performances and rumen conditions. Due to high variability in animals (initial rumen pH ranged from 6.1 to 6.7) and sensor drifts and noises, absolute pH indicators appeared less relevant than relative pH indicators for SARA detection during HSD.). The observed AUC of receiver operating characteristic (ROC) curves for absolute pH indicators (from 0.48 to 0.71) were lower than those of relative pH indicators (0.78 to 0.91). Moreover AUC of ROC curve of relative pH indicators were statistically higher than time spent under pH < 6.0 (DeLong’s test for 2 correlated ROC curves, P < 0.05). This work shows that relative pH indicators can take into account drift and noise of pH sensors and animal variability. Relative pH indicators appear relevant and robust for SARA detection in a herd and to compare SARA situations among studies.

Key Words: subacute ruminal acidosis (SARA), reticulo-rumen pH sensor, relative pH indicators