Because of the enormous variability of hepatitis C virus (HCV), the development of reliable genotyping assays is a formidable challenge. The optimal genotyping region appears to be the 5' untranslated region (UR) because of high conservation within, but considerable variability between, genotypes. In this study, 21 probes dispersed over seven variable 5' UR areas were applied to a line probe assay (LiPA) and used to analyze 506 HCV-infected sera from different geographical regions representing a multitude of subtypes. At least 31 different reactivity patterns emerged, with 404 (80%) of 506 distributed over 11 prototype patterns, in general corresponding to subtypes 1a, 1b, 2a/2c, 2b, 3a, 5a, and 6a and several type 4 subtypes. Subtyping specificity ranged from 97% in Hong Kong to 90% in Europe but was only 11% in West Africa, while typing specificity was always 100% when samples from Vietnam were excluded. In a second evaluation, the subtype prediction by LiPA of 448 GenBank 5' UR HCV sequences was scored. Of the 58 theoretically predicted patterns, 321 sequences (72%) were covered by the 11 prototype patterns. We concluded that (i) the selected probes detected the corresponding signature motifs in the seven variable regions with 100% reliability; (ii) these motifs allowed correct type interpretation of samples collected worldwide, with the exclusion of Vietnam, Thailand, or Vietnamese patients residing in European hospitals; and (iii) subtyping specificities vary according to geographical region, with 11 prototype subtyping patterns identifying the majority of samples from Europe and the Americas. These results indicate that the LiPA is a reliable assay applicable to routine typing and subtyping of HCV specimens.