Amylase activity in serum and urine: comparison of results by the amyloclastic, dyed-starch, and nephelometric techniques.

We assayed serum and urine specimens for amylase activity by the nephelometric (I),dyed-starch (Amylochrome) (II), and mayloclastic (III) techniques. For serum, the correlation coefficients of the regression lines were: I vs. II, 0.978 (n = 106); I vs. III, 0.736 (n = 110); and II vs. III, 0.739 (n = 108). For urine, they were I vs. II, 0.938 (n = 49); and I vs. III, 0.752 (n = 46). Because calculation of the Kolmogorov-Smirnov statistic showed the distributions to be nongaussian, Spearman rank correlation coefficients were determined and showed that I and II correlated well but neither method correlated with III. The clinical data show that I and II gave above-normal activities in every case of pancreatitis, but III gave normal values in two of eight cases. In all cases, I and II were more sensitive, giving higher amylase activities (as compared with the upper limit of normal) than did III. The nephelometric procedure is most suitable for routine and emergency testing; the dyed-starch assay is equally sensitive and reliable, but less convenient. The amyloclastic procedure appears to be less reliable.