Development of dose-response models of Creutzfeldt-Jakob disease infection in nonhuman primates for assessing the risk of transfusion-transmitted variant Creutzfeldt-Jakob disease

Estimates for the risk of transmitting variant Creutzfeldt-Jakob disease (vCJD) via blood transfusion have relied largely on data from rodent experiments, but the relationship between dose (amount of infected blood) and response (vCJD infection) has never been well quantified. The goal of this study was to develop a dose-response model based on nonhuman primate data to better estimate the likelihood of transfusion-transmitted vCJD (TTvCJD) in humans. Our model used dose-response data from nonhuman primates inoculated intracerebrally (i.c.) with brain tissues of patients with sporadic and familial CJD. We analyzed the data statistically by using a beta-Poisson dose-response model. We further adjusted model parameters to account for the differences in infectivity between blood and brain tissue and in transmission efficiency between intravenous (i.v.) and i.c. routes to estimate dose-dependent TTvCJD infection. The model estimates a mean infection rate of 76% among recipients who receive one unit of whole blood collected from an infected donor near the end of the incubation period. The nonhuman primate model provides estimates that are more consistent with those derived from a risk analysis of transfused nonleukoreduced red blood cells in the United Kingdom than prior estimates based on rodent models. TTvCJD was recently identified as one of three emerging infectious diseases posing the greatest immediate threat to the safety of the blood supply. Cases of TTvCJD were reported in recipients of nonleukoreduced red blood cells and coagulation factor VIII manufactured from blood of United Kingdom donors. As the quantity of abnormal prions (the causative agent of TTvCJD) varies significantly in different blood components and products, it is necessary to quantify the dose-response relationship for a wide range of doses for the vCJD agent in transfused blood and plasma derivatives. In this paper, we suggest the first mechanistic dose-response model for TTvCJD infection based on data from experiments with nonhuman primates. This new model may improve estimates of the possible risk to humans.