Whole blood samples were incubated overnight in buffer containing stainless steel powder, as previously described.³  The powder was heat treated and sequentially incubated with biotinylated anti-PrP antibody, NeutrAvidin-HRP, and chemiluminescent substrate. Samples were scored reactive if the mean signal from 3 replicates exceeded an on-plate negative control cutoff. Repeat-reactive samples were considered positive for vCJD; nonreactive and single-reactive samples were considered negative for vCJD.

Anonymous blood samples from unaffected blood donors were obtained from the American Red Cross (US normals) and the National Health Service Blood and Transplant service of England and Wales (UK normals). Positive control samples consisted of 0.1% wt/vol vCJD-infected brain homogenate in whole unaffected human blood (vCJD spikes) and were used to preserve scarce stocks of endogenous vCJD patient samples (vCJD patients). Clopper and Pearson’s “exact” method was used for confidence intervals of proportions. Likelihood ratios and their confidence intervals were calculated using Haldane’s correction and a derived approximation for risk ratios.

HRP, horseradish peroxidase.