BACKGROUND/AIMS: Several strains of the hepatitis C virus exist; distinct genotypes and subtypes can be identified by sequence comparison of the viral genomes. Recent evidence that the genotype/subtype of hepatitis C virus may influence the clinical course of chronic hepatitis C and the response to interferon-alpha therapy for this disease suggests that methods to identify the genotype may become clinically useful. In the present study we evaluated a recently introduced reverse hybridization assay. METHODS: HCV-RNA was isolated from serum samples from 61 consecutive patients attending our out-patient clinic and subsequently sequenced in the 5'-noncoding and the nonstructural-5 region by the dideoxynucleotide chain termination method. HCV-genotyping was performed by phylogenetic analysis of nonstructural-5 sequences. The amplification product for the reverse hybridization assay was obtained by "nested" polymerase chain reaction using biotinylated primers corresponding to the 5'-noncoding region. The assay is based on hybridization of the resulting polymerase chain reaction product with oligonucleotide probes immobilized as parallel lines on membrane strips. RESULTS: According to the phylogenetic analysis of the nonstructural-5 region the prevalence of hepatitis C virus subtypes was as follows: 1a 18%, 1b 51%, 2a 3%, 2b 3%, 2c 7% and 3a 18%. The reverse hybridization assay correctly identified each hepatitis C virus genotype (1, 2, and 3). However, differentiation of hepatitis C virus subtypes was insufficient. 1/11 HCV-1a isolates was incorrectly classified by the reverse hybridization assay as HCV-1b and vice versa 3/31 HCV-1b isolates as HCV-1a. Classification of hepatitis C virus subtypes 2a, 2b and 3a was correct, but 4/4 HCV-2c isolates were misinterpreted by the assay as HCV-2a. CONCLUSIONS: The reverse hybridization assay can differentiate between hepatitis C virus genotypes 1, 2, and 3, but is not completely reliable for hepatitis C virus subtyping.