Genomic and phylogenetic analysis of hepatitis C virus isolates: a survey of 535 strains circulating in southern France.

The present study examines the distribution of Hepatitis C virus (HCV) genotypes in Marseille, France in 2001-2002 and evaluates the efficiency of two in house direct sequence PCR protocols based on 5'NC analysis or NS5B analysis. By 5'NC sequencing, the distribution of 535 HCV strains derived from patients attending gastroenterology and AIDS referral centers, or dialysis units was as follows: 33% were infected by genotype 1a; 26% by 1b; 7% by 2; 22% by 3a; 10.7% by 4. In univariate analysis, HCV distribution was associated with age and source of infection, whereas in multivariate analysis only injecting drug use was an independent determinant for genotype distribution. Among the 535 specimens submitted to 5'NC direct sequencing, 18% could not be classified accurately into subtypes. A subset of 187 samples was amplified efficiently and sequenced by targeting the NS5B region of the viral genome. The two methods yielded concordant results in 70% of cases. Specimens unsubtypeable or misclassified most frequently by 5'NC analysis were type 1b and subtypes 2a/2c and 4a/4c. The data show that 5'NC direct sequence analysis is a sensitive method to identify genotypes in all cases, but that it can lead to subtyping misclassification (in particular, subtype 1b and 1a) or doubtful results (in particular subtypes 2a/2c and 4a/4c). Conversely, NS5B direct sequence assay, based on phylogenetic analysis, allowed better discrimination between subtypes. These two approaches are complementary and should be made available in clinical laboratories to ensure a reliable survey of HCV strains. Copyright 2003 Wiley-Liss, Inc.