Comparison of polyacrylamide gel electrophoresis, an enzyme-linked-immunosorbent assay, and an agglutination test for the direct identification of bovine rotavirus from feces and coelectrophoresis of viral RNAs.

The dsRNA concentrated polyacrylamide gel electrophoresis (CPAGE) detected rotavirus directly from 19% of 77 stool specimens from diarrheic calves. A commercial enzyme-linked immunosorbent assay (ELISA) detected 25%, latex agglutination test, 23%, and polyacrylamide gel electrophoresis (PAGE), 19%. Establishing CPAGE as the "standard," the commercial ELISA and the latex agglutination test both had higher sensitivity (84%) than PAGE (79%). However, PAGE produced the highest specificity (100%), followed by agglutination (88%) and ELISA (84%). The commercial ELISA had a slightly higher sensitivity than agglutination, PAGE, and CPAGE, but the ELISA specificity was generally lower. The latex agglutination test had a lower sensitivity than ELISA, but specificity was higher. Agglutination had similar negative predictive values (94%), compared with agglutination and PAGe, but had the lowest positive predictive value (a measure of accuracy) (70%). Agreement with CPAGE was highest for PAGE (94.8%), followed by agglutination (87%) and ELISA (84.4%). The calculated percentages of total disagreement with all other tests indicated that ELISA differed from the other rotavirus detection assays in 10.4% of the cases, agglutination in 7.8%, PAGE in 2.6%, and CPAGE in 1.3%. The 2 PAGE assays allowed the detection of atypical rotaviruses from feces based on the characteristic "super-short" migration pattern of the 11 genomic segments of rotaviruses and of other members of the Reoviridae.