In their article and accompanying press release, the Buffalo researchers played a statistical trick by how they dealt with 52 “borderline” patients whom did not receive a complete Doppler assessment. They did CCSVI prevalence statistics by excluding the borderlines from the analysis (the obvious and proper procedure) but then also did the statistics putting the borderlines in the negative CCSVI category. This latter calculation has absolutely no justification and took the prevalence of CCSVI in those with MS from 63% to 56%.