In order to shorten the time required for the detection of enteroviruses in stool specimens, an 18-h immunoperoxidase test combining low-speed centrifugation and the use of a group specific anti-VP1 monoclonal antibody (5-D8/1, Dako) was developed. This rapid culture assay (RCA) was compared blindly to a conventional culture assay (CCA) on a panel of 180 children's stool specimens received for routine diagnosis of enterovirus infection. The same cell lines (human embryonic fibroblasts and KB continuous cell line) were used in both tests. Discrepancies in results were analysed by a PCR technique with primers located in a conserved part of the 5' non-coding region of the enterovirus genome. Fourteen specimens were positive and 158 were negative with both tests. Four samples were positive with the RCA yet negative with the CCA and 3 others showed the opposite pattern; an additional sample positive by RCA was uninterpretable by CCA due to bacterial contamination. Subsequent PCR testing of these 8 samples showed no discrepancies; all were positive. Using CCA as the reference, the sensitivity and specificity of RCA were 77.8 and 98% respectively. Kinetic studies using enterovirus isolates demonstrated that RCA was much more sensitive than CCA during the first three days of culture. These results further suggested that RCA sensitivity could be improved by a factor of at least 10 times by prolonging the incubation period by 24 hr. With this change, the RCA assay described below is suggested as a rapid alternative to CCA for the routine diagnosis of enterovirus infection in stool specimens. When an identification at the serotype level is required, samples found positive using RCA could then be subjected to CCA.