Vetnews | Mei 2025 22 « BACK TO CONTENTS Article 2.4. Statistical Analysis Data were analyzed using Reference Value Advisor v.2.1 (Ecole Nationale Veterinaire de Toulouse, Toulouse, France), and reference intervals were calculated according to the guidelines of the American Society for Veterinary Clinical Pathology (Madison, WI, USA) [17,18]. Tukey’s method was used to assess for outliers and suspect data. These values were retained unless there was a convincing error or a result pattern inconsistent with health. The distribution of the data was evaluated by the Anderson–Darling method, with a p-value < 0.05 indicating non-Gaussian distribution. The symmetry of the data was evaluated using the runs test, with a p-value < 0.05 indicating asymmetry. The parametric method was selected for Gaussian distributions. The nonparametric method was selected when data had a non-Gaussian and asymmetrical distribution, both before and after Box-Cox transformation. The bootstrap method was used to establish confidence intervals. In any instance where outliers were removed, the data were re-analyzed using the methods above. 3. Results The study population consisted of 28 owned donkeys in New York (23%), 6 in Massachusetts (5%), 18 in California (15%), 44 in Texas (37%), and 24 (20%) free-roaming donkeys upon capture from the wild in Death Valley National Park, California. Most donkeys sampled were standard-sized (n = 102), followed by a miniature (n = 17), and only a single mammoth, with 55 being intact males, 16 castrated males, and 49 females. A total of 120 EDTA whole blood samples (and blood smears) were collected for CBC analysis, but 9 samples were rejected due to clotting in the blood tube, and 2 samples were rejected because it was discovered that they were collected from foals less than six months of age. A total of 120 serum samples were collected for biochemistry analysis, but 1 was lost during shipping, and 2 samples were rejected because it was discovered that they were collected from foals less than six months of age. The majority of suspect and outlier data was retained as the population was well defined with general health established. The outliers removed are detailed in Table 2. CBC and biochemistry reference interval data are presented in Tables 3 and 4, respectively. Measurand Method Albumin Bromocresol green dye-binding AST NADH oxidation Electrolytes Indirect ion-selective potentiometry Bicarbonate Phosphoenolpyruvate carboxylase based Bilirubin, total Jendrassik-Grof based- Diazo method (diazonium ion) Bilirubin, direct Jendrassik-Grof based- (diazotized sulfanilic acid) Calcium 5-nitro-5′-methyl-BAPTA (NM-BAPTA) Cholesterol Cholesterol esterase Creatinine kinase Creatine phosphate cleavage Creatinine Modified Jaffe GGT L-gamma-glutamyl-3-carboxy-4- nitroanilde substrate Glucose Hexokinase GLDH Alpha-oxoglutarate substrate Magnesium Xylidyl blue, diazonium salt Phosphate Ammonium molybdate SDH D-fructose substrate Triglycerides Lipoprotein lipase Total protein Biuret Urea Nitrogen Urease- kinetic Table 1. Basic methodology for biochemical testing performed on the Roche Cobas. Measurand n Reason(s) Manual WBC Differential 17 Uneven distribution of WBCs on smears causing erroneous differential percentages/counts Platelet 21 Marked platelet clumping causing falsely decreased counts MPV 21 Marked platelet clumping causing falsely increased MPV TP (on CBC) 1 Moderate to marked hemolysis causing interference with a refractometer Electrolytes 1 Multiple outliers from a single sample, suggest subclinical disease Calcium 1 Low outlier associated with low albumin, a presumed artefact of decreased protein binding Glucose 12 Very low values (<50 mg/ dL) incompatible with health, presumed artefact of delayed serum separation from cells (10 of 12 outliers were received as whole clotted blood). AST, SDH, GLDH 1 Multiple high hepatocellular enzyme outliers from a single sample, suggest subclinical liver disease. GGT 6 Multiple high outliers are associated with a single herd. Samples received as whole clotted blood, suspect pre-analytical artefact. GGT 3 Multiple high outliers are associated with a single herd. Follow-up communication with the submitter reveals a history of liver flukes in herds. Cholesterol and Triglycerides 3 Samples with both lipid values marked as suspect or outlier data (high), suggesting subclinical disease Triglycerides 1 Single sample with marked hypertriglyceridemia, incompatible with health CK 5 Samples with CK > 800, suggestive of traumatic handling or restraint Table 2. Outliers removed from analysis.
RkJQdWJsaXNoZXIy OTc5MDU=