Improvement of the Specificity of Enzyme Immunoassays for the Detection of Rotavirus and Adenovirus in Fecal Specimens

Abstract
A variable rate of false-positive results may be observed with commercial enzyme immunoassays (EIAs) for the detection of rotavirus and adenovirus antigen in stool specimens, depending on the quality of the reagents and the presence of potentially interfering substances in stool samples. The present study was performed in an attempt to improve the specificity of current commercial rotavirus and adenovirus EIAs without significant loss of sensitivity by optimizing the cut-off value. A collective of 174 stool samples obtained from children suffering from acute gastroenteritis was tested. Electron microscopy (EM) and PAGE were used as reference methods for rotavirus detection. For the evaluation of the adenovirus kits, virus isolation in cell culture and the polymerase chain reaction served as reference standards. The highest sensitivity for rotavirus and adenovirus detection was achieved by the Ridascreen® Rotavirus and Ridascreen® Adenovirus. However, the Ridascreen® Rotavirus and Ridascreen® Adenovirus produced the highest number of false-positive results (n = 9) for each rotavirus and adenovirus detection. Cross-reactivities to coronaviruses and reoviruses were observed with the rotavirus antigen EIAs. For Rotazyme II, Ridascreen Rotavirus and Ridascreen Adenovirus, the specificity could be markedly increased without loss of sensitivity by doubling the cut-off value. For the alternative immunoassays, which were overall more specific, it was not possible to significantly decrease the rate of false-positive results without impairment of sensitivity by raising the cut-off value. In conclusion, at least for some rotavirus and adenovirus antigen EIAs, the cut-off value set by the manufacturer may not permit an optimal differentiation between true-positive and -negative samples. By raising the cut-off value from 50 to 100%, the specificity of two rotavirus antigen and one adenovirus antigen EIA can be improved markedly without significant loss of sensitivity.

This publication has 1 reference indexed in Scilit: