The aim of this study was to investigate the reliability of an immunoturbidometric assay for measuring the acute phase protein serum amyloid A (saa) in horses in clinical practice. The assay was compared to a previously validated assay, and overlap performance was assessed by measuring the concentration of saa in clinically healthy horses and horses with inflammatory and non-inflammatory diseases. In pools of serum with low and high saa concentrations the assay's intra-assay coefficients of variation were 11·7 per cent and 4·6 per cent, and its interassay coefficients of variation were 9·1 per cent and 5·6 per cent, respectively. Slight inaccuracies were observed, but they were negligible in comparison with the range of the saa response. The assay systematically underestimated the concentrations of saa in comparison with the results of the validated assay. The assay detected the expected difference in saa concentrations between the healthy and diseased horses.