We are pleased to read that Germer et al. confirm our results 1 showing that the Cobas AmpliPrep/Cobas TaqMan assay (CAP/CTM; Roche Molecular Systems, Pleasanton, CA) underestimates hepatitis C virus (HCV) RNA levels in a substantial number of patients infected with HCV genotype 4. We recently reported two cases where the simultaneous presence of a G-to-A substitution at position 145 (G145A) and an A-to-T substitution at position 165 (A165T) of the 5Ј noncoding sequence of HCV genome abolished HCV RNA detection in the CAP/CTM assay, but not in two other assays, including another real-time polymerase chain reaction test (m2000 SP /m2000 RT Real-Time Assay, Abbott Molecular, Des Plaines, IL) and the third-generation branched DNA (bDNA)-based signal amplification assay (Versant HCV RNA 3.0; Siemens Medical Solutions Diagnostics, Tarrytown, NJ). 2 In order to confirm the role of nucleotide substitutions at positions 145 and 165 in the failure of CAP/CTM to accurately quantify HCV RNA, we generated in vitro RNA transcripts from a plasmid into which the 5Ј noncoding sequence of a wild-type genotype 4 strain originating from an HCV-infected patient was inserted. Substitutions at positions 145 and 165 were introduced by means of site-directed mutagenesis. The transcripts harboring one or both substitutions, as well as the wild-type transcript, were quantified without extraction with the three assays described above. The results are shown in Table 1. Each single substitution was responsible for an approximately 2-log underestimation of HCV RNA levels in the Cobas TaqMan assay relative to the two other assays. In the presence of both G145A and A165T substitutions, HCV RNA levels were high in m2000 RT and bDNA, but undetectable with the Cobas TaqMan assay.
The frequency of substitutions at positions 145 and 165 in Gen-Bank reported by Germer et al. and ourselves 2 is not negligible. We agree that the presence of the double substitution at positions 145 and 165 is rare, at least in the sequences reported in GenBank. Given the fact that the incidence and prevalence of genotype 4 infection is increasing worldwide, especially in the injection drug user population, and that HCV RNA quantification is critical to make treatment decisions, accuracy is essential. The current version of the CAP/CTM assay, which underquantifies approximately 15% of HCV genotype 2 and 30% of HCV genotype 4 samples, 1 is not appropriate for use in patients infected with these genotypes. An improved version of this assay is currently in development. It is expected to bring accurate quantification of all HCV genotypes. Careful evaluation of this new assay will be needed in expert centers.