The recent development of tagged RT-PCR and rTth RT-PCR has greatly improved strand-specific detection of hepatitis C virus (HCV) RNA but these assays are still prone to some false detection of the incorrect strand of RNA. In this study we aimed to address additional factors which contribute towards false detection of HCV RNA. Firstly the benefits of both tagged primers and the thermostable reverse transcriptase rTth during cDNA synthesis were combined and it was found that strand specificity was greatly improved without compromising sensitivity. The reliability of the assay was then optimised by addressing the following issues: control synthetic transcripts should be free of contaminating plasmid DNA, residual RT activity should be minimised in the presence of PCR primers and cDNA should be free of unincorporated tagged RT primer prior to PCR amplification. The alterations made to the assay eliminated completely false detection of the incorrect strand of RNA in the control assay whilst the correct strand was consistently detected at a cDNA dilution of 10(-3)-10(-4). Negative strand was not detected in RNA isolated from serum but was detected, at a ten-fold lower level than positive strand, in RNA isolated from liver tissue.