Models to predict the risk of subacute ruminal acidosis in dairy cows based on dietary and cow factors: A meta-analysis

Abstract

The present research aimed at developing practical and feasible models to optimize feeding adequacy to maintain desired rumen pH conditions and prevent subacute ruminal acidosis (SARA) in dairy cows. We conducted 2 meta-analyses, one using data from recent published literatures (study 1) to investigate the prediction of SARA based on nutrient components and dietary physical and chemical characteristics, and another using internal data of our 5 different published experiments (study 2) to obtain adjustments based on cow status. The results of study 1 revealed that physically effective neutral detergent fiber inclusive of particles >8 mm (peNDF >8) and dietary starch [% of dry matter (DM)] were sufficient for predicting daily mean ruminal pH {y = 5.960 - (0.00781 × starch) + (0.03743 × peNDF >8) - [0.00061 × (peNDF >8 × peNDF >8)]}. The model for time of pH suppression (<5.8 for ruminal pH or <6.0 for reticular pH, min/d) can be predicted with additionally including DMI (kg/d): 124.7 + (1.7007 × DMI) + (20.9270 × starch) + (0.2959 × peNDF >8) - [0.0437 × (DMI × starch × peNDF >8)]. As a rule of thumb, when taken separately, we propose 15 to 18% peNDF >8 as a safe range for diet formulation to prevent SARA, when starch or NFC levels are within 20 to 25% and 35 to 40% ranges, respectively. At dietary starch content below 20% of DM, grain type was insignificant in affecting ruminal pH. However, increasing dietary starch contents by using corn as the sole grain source could lead to more severe drops of pH compared with using grain mix based on barley and wheat, as underlined by an interaction between starch content and grain type. Data from study 2 emphasized an increased risk of SARA for cows in the first and second lactation with lower mean pH (0.2 units) and double amounts of time at pH <5.8 compared with the cows with ≥3 parities. Given that a lower ruminal pH is expected in these high-risk cows, it is advisable to keep the lower end of recommended starch (20%) and higher peNDF >8 (18%) contents in the diet of these cows. Overall, the present study underlines the possibility of predicting SARA based on dietary factors including peNDF >8 and starch contents, as well as DMI of the cows, which can be practically implemented for optimal diet formulation for dairy cows. With more data available, future studies should attempt to improve the predictions by including additional key dietary and cow factors in the models.