Abstract
Accurate diagnosis of rotavirus is important in both clinical and research situations. A total of 100 stool specimens from children with diarrhea were tested for rotavirus by electron microscopy. These specimens were then coded and tested for rotavirus by four procedures: a monoclonal antibody-based enzyme immunoassay (EIA) (Pathfinder; Kallestad Laboratories, Inc., Austin, Tex.), two polyclonal antibody-based EIAs (Rotazyme II; Abbott Laboratories, North Chicago, Ill.; and an EIA performed with reagents from the National Institutes of Health, Bethesda, Md. [NIH reagent EIA]), and a latex agglutination (LA) assay (Rotalex; Medical Technology Corp., Somerset, N.J.). The sensitivity of the monoclonal antibody EIA (95%) was superior to those of the polyclonal antibody EIAs (73% for Rotazyme II and 57% for the NIH reagent EIA) and the LA assay (61%). The specificity of the LA assay (98%) was slightly better than those of the other systems (88 to 96%). The positive and negative predictive values of the monoclonal antibody EIA (93 and 96%, respectively) were better than those of Rotazyme II (82 and 80%, respectively), the LA assay (96 and 76%, respectively), and the NIH reagent EIA (93 and 74%, respectively). The visual readings of the monoclonal antibody EIA correlated better with the spectrophotometric optical density readings than did the visual readings of the polyclonal antibody EIAs; however, the agreement of both with electron microscopy results was poor when 1+ or plus-minus readings were observed. The monoclonal antibody EIA is more sensitive and predictive than other rotavirus detection systems and second only to the LA assay in specificity in detecting rotavirus in stool specimens.