The current classification of human sporadic prion diseases recognizes six major phenotypic subtypes with distinctive clinicopathological features, which largely correlate at the molecular level with the genotype at the polymorphic codon 129 (methionine, M, or valine, V) in the prion protein gene and with the size of the proteaseresistant core of the abnormal prion protein, PrP Sc (i.e. type 1 migrating at 21 kDa and type 2 at 19 kDa). We previously demonstrated that PrPSc typing by Western blotting is a reliable means of strain typing and disease classification. Limitations of this approach, however, particularly in the interlaboratory setting, are the association of PrPSc types 1 or 2 with more than one clinicopathological phenotype, which precludes definitive case classification if not supported by further analysis, and the difficulty of fully recognizing cases with mixed phenotypic features. In this study, we tested the inter-rater reliability of disease classification based only on histopathological criteria. Slides from 21 cases covering the whole phenotypic spectrum of human sporadic prion diseases, and also including two cases of variant Creutzfeldt-Jakob disease (CJD), were distributed blindly to 13 assessors for classification according to given instructions. The results showed good-to-excellent agreement between assessors in the classification of cases. In particular, there was full agreement (100 %) for the two most common sporadic CJD subtypes and variant CJD, and very high concordance in general for all pure phenotypes and the most common subtype with mixed phenotypic features. The present data fully support the basis for the current classification of sporadic human prion diseases and indicate that, besides molecular PrPSc typing, histopathological analysis permits reliable disease classification with high interlaboratory accuracy.