Canonical Correlation of Milk Composition Parameters and Blood Biomarkers in High-Producing Dairy Cows During Different Lactation Stages

This study explores milk composition and blood markers in cows across lactation stages. Holstein cows were divided into four groups: beginning of lactation (BL; <i>n</i> = 21), peak of lactation (PL; <i>n</i> = 21), middle of lactation (ML; <i>n</i> = 21), and end of lactation (EL; <i>n</i> = 20). Blood (1 × 15 mL) and milk samples (1 × 100 mL) were collected for biomarker analysis. Blood chemistry profiles were determined using a clinical chemistry analyser, and milk lactose, fat, and protein levels (%) were determined using an infrared absorbance analyser. Minerals (Ca, P, and Mg) in milk were determined by atomic absorption spectrometry after mineralizing the samples. Glucose was higher in the EL group than in the BL group (<i>p</i> < 0.01), whereas D-beta-hydroxybutyrate (D-BHB) was higher in the BL group than in the PL and ML groups (<i>p</i> < 0.001). Cholesterol was higher in the PL, ML, and EL groups than in the BL group (<i>p</i> < 0.001). Gamma-glutamyl transferase was increased in the PL group compared to the BL group. Phosphorus levels were lower in the PL than in the BL group, whereas protein levels were higher in the EL than in the PL group. Spearman and partial correlation analysis showed several significant associations between the observed variables. Using canonical correlation analysis were identified three significant correlations (<i>r</i>_<i>c</i>1 = 0.853; <i>r</i>_<i>c</i>2 = 0.823; <i>r</i>_<i>c</i>3 = 0.739). The main canonical correlation identified blood TG and milk urea as the strongest variables. According to the canonical loading, the biomarkers TG, Mg, urea, cholesterol, and alkaline phosphatase (<i>U1</i>) are the primary variables associated with milk parameters (<i>V1</i>), specifically with milk urea, milk Mg and P, protein, and lactose.