ABSTRACT
OBJECTIVE: A retrospective study was conducted to establish the prerace venous acid-base and blood gas values of Standardbred horses at rest using big data analytics. SAMPLES: Venous blood samples (73,382) were collected during seven racing seasons from 3 regional tracks in the Commonwealth of Pennsylvania. Horses were detained 2 hours prior to race time. PROCEDURES: A mixed-effects linear regression model was used for estimating the marginal model adjusted mean (marginal mean) for all major outcomes. The interaction between age and gender, track, and the interaction between month, treatment (furosemide), and year were the major confounders included in the model. Random effects were set on individual animal nested within trainer. Partial pressure of venous carbon dioxide (PVCO2), partial pressure of oxygen (PVO2), and pH were measured, and base excess (BE), total carbon dioxide (TCO2), and bicarbonate (HCO3-) were calculated. RESULTS: Significant (P < .001) geographical differences in track locations were seen. Seasonal reductions in acid-base values started in January with significant (P < .001) decreases from adjacent months seen in June, July, and August followed by a gradual return. There were significant increases (P < .001) in BE and TCO2 and decreases in PVO2 with age. Significant differences (P < .001) in acid-base values were seen when comparing genders. A population of trainers were significantly different (P < .001) from the marginal mean and considered outliers. CLINICAL RELEVANCE: In a population of horses, big data analytics was used to confirm the effects of geography, season, prerace furosemide, gender, age, and trainer influence on blood gases and the acid-base profile.