Population screening for variant Creutzfeldt-Jakob disease using a novel blood test: diagnostic accuracy and feasibility study.

Abstract

Our study indicates a prototype blood-based variant Creutzfeldt-Jakob disease (vCJD) assay has sufficient sensitivity and specificity to justify a large study comparing vCJD prevalence in the United Kingdom with a bovine spongiform encephalopathy-unexposed population. In a clinical diagnostic capacity, the assay's likelihood ratios dramatically change an individual's pretest disease odds to posttest probabilities and can confirm vCJD infection. To determine the diagnostic accuracy of a prototype blood test for vCJD and hence its suitability for clinical use and for screening prion-exposed populations. Retrospective, cross-sectional diagnostic study of blood samples from national blood collection and prion disease centers in the United States and United Kingdom. Anonymized samples were representative of the US blood donor population (n = 5000), healthy UK donors (n = 200), patients with nonprion neurodegenerative diseases (n = 352), patients in whom a prion disease diagnosis was likely (n = 105), and patients with confirmed vCJD (n = 10). Presence of vCJD infection determined by a prototype test (now in clinical diagnostic use) that captures, enriches, and detects disease-associated prion protein from whole blood using stainless steel powder. The assay's specificity among the presumed negative American donor samples was 100% (95% CI, 99.93%-100%) and was confirmed in a healthy UK cohort (100% specificity; 95% CI, 98.2%-100%). Of potentially cross-reactive blood samples from patients with nonprion neurodegenerative diseases, no samples tested positive (100% specificity; 95% CI, 98.9%-100%). Among National Prion Clinic referrals in whom a prion disease diagnosis was likely, 2 patients with sporadic CJD tested positive (98.1% specificity; 95% CI, 93.3%-99.8%). Finally, we reconfirmed but could not refine our previous sensitivity estimate in a small blind panel of samples from unaffected individuals and patients with vCJD (70% sensitivity; 95% CI, 34.8%-93.3%). In conjunction with the assay's established high sensitivity (71.4%; 95% CI, 47.8%-88.7%), the extremely high specificity supports using the assay to screen for vCJD infection in prion-exposed populations. Additionally, the lack of cross-reactivity and false positives in a range of nonprion neurodegenerative diseases supports the use of the assay in patient diagnosis.