Discordant results from hepatitis C virus genotyping by procedures based on amplification of different genomic regions

We compared the results of genotyping hepatitis C virus (HCV) either by PCR amplification of the core region or by hybridization of PCR-amplified products of the 5' untranslated region (5'UTR assay). Serum samples from 144 Italian anti-HCV-positive patients (106 drug abusers and 38 patients with chronic viral liver disease but no history of drug abuse) were studied. The original core region assay described by Okamoto et al. (H. Y. Okamoto, Y. Sugiyama, S. Okada, K. Kurai, Y. Akahane, Y. Sugai, T. Tanaka, K. Sato, F. Tsuda, Y. Miyakawa, and M. Mayumi, J. Gen. Virol. 73:673-679, 1992) allowed genotyping of 75 of 144 samples. A modified version of Widell et al. (A. Widell, S. Shev, S. Mansson, Y.-Y. Zhang, U. Foberg, G. Norkrans, A. Fryden, O. Weiland, J. Kurkus, and E. Nordenfelt, J. Med. Virol. 44:272-279, 1994) allowed genotyping of 11 of 79 samples (50 of 79 samples remained unclassified by the method of Okamoto et al. In contrast, all 144 samples were genotyped by the 5'UTR assay. Forty-six of 75 (61 percent) of the samples genotyped by the method of Okamoto et al. and 10 of 11 (91 percent) of the samples genotyped by the method of Widell et al. had results consistent with those obtained by the 5'UTR assay. According to the results of direct sequencing, the method of Okamoto et al. erroneously classified seven samples as having mixed infections. In conclusion, HCV genotyping seems more reliable when it is performed by the 5'UTR assay than by either of two core region assays. The major advantage provided by the 5'UTR assay is a much lower proportion of negative or indeterminate results in younger patients with histories of drug abuse or infection by genotypes other than HCV type 1.