Assessment of the Degree of Dehydration in Dogs Based on Biochemical Parameters Using Ordinal Logistic Regression

Abstract

            This experimental study was carried out in the faculty of veterinary medicine, at Suez Canal University, Ismailia, Egypt. A total of 40 dogs were categorized based on the severity of dehydration into three categories (mild, moderate, and severe) and a fourth group for dogs without dehydration. Many biochemical parameters were utilized to evaluate dehydration, including blood electrolytes (Na, K, CL, Ca, Mg, and Ph), liver enzymes (ALT and AST), kidney function parameters (urea, creatinine, and uric acid), and lactate. The most prevalent clinical manifestations of gastroenteritis in dogs were vomiting, followed by profuse watery yellowish to bloody diarrhea, anorexia, and mild, moderate, and severe degrees of dehydration manifested by STT retardation. Four OLR models ranging from univariable to multivariable logistic regression were developed. Lactate, AST, creatinine, urea, and uric acid were recorded as positive predictors for the severity of dehydration; however, only lactate, AST, and uric acid were recorded as positive significant (p<0.05) predictors for the degree of dehydration. Na, K, Cl, Ca, Mg, Ph, and ALT were all negative predictors of dehydration level. Na and K were significant (P0.05) negative predictors of the degree of dehydration, whereas the remaining variables were not substantially related to the degree of dehydration. It was observed that biochemical markers are good indicators of dehydration; including these factors in the OLR model will help in differentiating between different degrees of dehydration.