Garske T, Ghani AC 2010, Uncertainty in the tail of the variant Creutzfeldt-Jakob disease epidemic in the UK (PLoS One 5:e15626, PMID 21203419, doi 10.1371/journal.pone.0015626)

A mathematical modelling paper from Imperial College London that asks how long variant CJD might keep appearing in the UK and whether it could grow again. The authors simulate two routes of spread, food-borne exposure and red blood cell transfusion, fit the model to UK case counts and an appendix tissue survey, and conclude that a long, low tail is plausible, with a possible second peak driven largely by transfusion, but no self-sustaining epidemic. The record is open access and its bibliographic details were checked in three registries. This entity is aggregate epidemiology only.

Key facts

  • The paper is by Tini Garske and Azra C. Ghani, Department of Infectious Disease Epidemiology, MRC Centre of Outbreak Analysis and Modelling, Imperial College London, and appeared in PLoS One volume 5, issue 12, article e15626, on 23 December 2010 under a CC BY licence.
  • The model is a stochastic, individual-based branching process stratified by age and by prion protein codon 129 genotype (MM, MV, VV), with an infection, latent, infectious and death sequence, and with both clinical and never-symptomatic infections allowed.
  • It was fitted to 171 UK cases (167 definite or probable deaths plus 4 possible) to the end of 2009, of which 3 were linked to blood transfusion, to the genotype distribution of cases, and to an appendix survey with 3 positives in 11,246 samples; the tonsil survey results were left out of the fit.
  • The most likely peak incidence of the projected tail is about 10 to 11 cases per year (95% credible interval 1 to 65), occurring between roughly 2020 and 2035, with the second wave running from about 2010 to about 2080.
  • Median projected cumulative future cases for 2010 to 2179 are about 390 (95% credible interval 84 to 3000), most of them via transfusion and predominantly in the MM genotype; the authors judge such transfusion cases unlikely to be found by donor and recipient case linking.
  • In every scenario consistent with the data the basic reproduction number stays below 1 (median 0.0056), so the model finds no self-sustaining epidemic even if leucodepletion and the donor ban were ineffective.
  • Many inputs could not be estimated from the data (transfusion transmissibility, subclinical fraction, genotype susceptibility), so the wide intervals reflect assumed ranges and uniform priors; surgical and dental routes were not modelled.

Timeline

  1. 1998UK red cell leucodepletion introduced; the model assumes a baseline 40 percent reduction in transmissibility from this date.
  2. 2004UK ban on blood donation by people who have received a transfusion; the model assumes a baseline 90 percent effectiveness.
  3. 2010Manuscript received by PLoS One (12 August 2010).
  4. 2010Paper published online on 23 December 2010.

Last checked 1 November 2026

Sources for this page