Biochemical assessment of canine body cavity effusions using three bench-top analysers.

Abstract

OBJECTIVES: To assess the performance of three bench-top chemistry instruments for the analysis of canine effusions. Acceptable results were compared with those obtained by a reference chemistry analyser. METHODS: Total protein, albumin, creatinine and bilirubin concentrations were measured in 74 effusions using the VetScanVS2, VetTest8008 and SpotchemEZ analysers. Cholesterol and triglyceride concentrations were also measured by the VetTest and Spotchem. Results were analysed using Westgard's error analysis, Spearman's correlation, Bland-Altman plots and Deming regression. Results were considered acceptable when observed total error (TE(obs) ) was less than allowable total error (TE(A) ). RESULTS: VetScan error analysis revealed acceptable results for total protein (TE(obs) =1.11, TE(A) =4.7) and creatinine (TE(obs) =42.2, TE(A) =78.1). Correlation was fair for protein (r(s) =0.66) and creatinine (r(s) =0.76), but poor and not significant for bilirubin (r(s) =0.01, P=0.08), precluding error analysis. VetTest error analysis was acceptable for creatinine only (TE(obs) =5.55, TE(A) =25.5). Correlation was good (r(s) =0.81). The difference plot revealed a bias (95% confidence interval) of -1.5 (-37 to 40) and four outliers. The Spotchem did not generate a precise arithmetic value in most (56.9 to 73.6%) samples, precluding further analysis. CLINICAL SIGNIFICANCE: Acceptable results were obtained for total protein (VetScan) and creatinine [Vetscan, Vettest (with good correlation)]. The Spotchem is of limited value in canine effusion analysis.