In veterinary medicine, measurement of canine C-reactive protein (cCRP) is used widely to detect inflammatory diseases. We evaluated the precision of Randox and Fuji assays for cCRP, as well as accuracy, correlation, and agreement compared to a reference ELISA. Blood samples from 71 client-owned dogs (20 healthy, 51 diseased) were analyzed with the 3 assays. Inter-assay CVs were ~3.5% with both the Randox and Fuji assays. The mean biases were -1.90% for the Randox and -5.93% for the Fuji test; the targeted biases were ~8.5% for both assays. The CV, bias, and observed total error were acceptable for the 2 assays compared to ASVCP recommendations based on biological variation studies. The Spearman correlation coefficient for cCRP concentration compared with the reference ELISA was 0.83 for the Randox test and 0.92 for the Fuji test. Both assays measured cCRP precisely at intermediate and increased concentrations. Correlation with the reference ELISA was good, and both assays could be used to evaluate cCRP concentrations in veterinary practice. However, the assays did not reach analytical agreement; hence the results obtained by these assays are not interchangeable, and serial monitoring of cCRP requires the use of the same assay.
Keywords: C-reactive protein; dogs; inflammation; point-of-care testing.