Relative reticulo-rumen pH indicators for subacute ruminal acidosis detection in dairy cows
Résumé
Subacute ruminal acidosis (SARA) is usually characterized by abnormal and intermittent drops in rumen pH. Nevertheless, high
individual animal variability in rumen pH and the difference in measurement methods for pH data acquisition decrease the
sensitivity and accuracy of pH indicators for detecting SARA in ruminants. The aim of this study was to refine rumen pH indicators
in long-term SARA based on individual dairy cow reticulo-rumen pH kinetics. Animal performances and rumen parameters were
studied weekly in order to validate SARA syndrome and rumen pH was continuously measured using reticulo-rumen sensors. In
total, 11 primiparous dairy cows were consecutively fed two different diets for 12 successive weeks: a control diet as low-starch
diet (LSD; 13% starch for 4 weeks in period 1), an acidotic diet as high-starch diet (HSD; 32% starch for 4 weeks in period 2), and
again the LSD diet (3 weeks in period 3). There was a 1-week dietary transition between LSD and HSD. Commonly used absolute
SARA pH indicators such as daily average, area under the curve (AUC) and time spent below pH <5.8 and pH <6 were processed
from absolute (raw) daily kinetics. Then signal processing was applied to raw pH values in order to calculate relative pH indicators
by filtering and normalizing data to remove inter-individual variability, sensor drift and sensor noise. Normalized AUC, times spent
below NpH <− 0.3 and NpH <− 0.5, NpH range and NpH standard deviation were calculated. Those relative pH indicators were
compared with commonly used pH indicators to assess their ability to detect SARA. This syndrome induced by HSD was confirmed
by consistent expected changes in milk quality, dry matter intake and acetate : propionate ratio in the rumen, whereas the ruminal
concentration of lipopolysaccharide was increased. Commonly used pH SARA indicators were not able to discriminate SARA
syndrome due to high animal variability and sensor drift and noise, whereas relative pH indicators developed in this study
appeared more relevant for SARA detection as assessed by receiver operating characteristic tests. This work shows that absolute
pH kinetics should be corrected for drift, noise and animal variability to produce relative pH indicators that are more robust for
SARA detection. These relative pH indicators could be more relevant for identifying affected animals in a herd and also for
comparing SARA risk among studies.