In a previous epidemiological investigation among schoolchildren of Northern Italy, a conservative 1% prevalence of type 1 von Willebrand disease (VWD) was found. Diagnosis was based on a positive family history and low von Willebrand factor (VWF) ristocetin cofactor activity. To investigate whether the type 1 VWD phenotype as detected by our original methodology cosegregates with one or more specific alleles of the VWF gene, we have performed genotype analysis in affected subjects and their family members. Eleven of the 14 subjects previously identified as having VWD, all with mild personal bleeding symptoms, agreed to participate in the genetic study. Remarkably, the laboratory measurements of the previous investigation were completely confirmed in 10 of the 11 subjects. Clear cosegregation of the VWD type 1 and a specific VWF allele was observed in one family and was likely in the family of two other pro-bands. In three additional propositi and their families a possible association of the phenotype with a VWF allele was found. No association was observed in the remaining five subjects and their families. During 13-year follow-up few additional bleeding episodes were recorded among investigated subjects, most often occurring in the one family manifesting clear cosegregation. The results of this study illustrate that a personal and family bleeding history and persistently low VWF ristocetin cofactor activity, fitting the usual criteria for type 1 VWD, may not cosegregate with genetic markers at the VWF gene locus. Thus the prevalence of VWD defined as a disorder involving the VWF locus might be overestimated in population study. However, phenotypic diagnosis still remains fundamental to identify patients at risk of bleeding. Further research should clarify whether in families with more severe clinical and laboratory phenotype a clear association with markers of VWF is found.