Analyte comparisons between 2 clinical chemistry analyzers

The purpose of this study was to assess agreement between a wet reagent and a dry reagent analyzer. Thirteen analytes (albumin, globulin, alkaline pho sphatase, alanine aminotransferase, amylase, urea nitrogen, calcium, choles terol, creatinine, glucose, potassium, total bilirubin, and total protein) for both canine and feline serum were evaluated. Concordance correlations, linear regression, and plots of difference against mean were used to analyz e the data. Concordance correlations were excellent for 8 of 13 analytes (r greater than or equal to 0.90); the correlations for albumin, potassium, a nd calcium were clinically unreliable. The linear regression analysis revea led that several analytes had slopes significantly different from unity, wh ich was likely related to methodological differences. Compared to the wet r eagent analyzer, the dry reagent analyzer showed excellent agreement for al kaline phosphatase, alanine aminotransferase, amylase (feline), urea nitrog en, cholesterol, creatinine, glucose, total bilirubin (canine), and total p rotein. However, it showed only slight to substantial agreement for amylase (canine), calcium, albumin, potassium, and total bilirubin (feline).