Objective: To assess the effect of usage of three different versions of Creutzfeldt-Jakob disease (CJD) diagnostic criteria on estimates of CJD incidence. Methods: A total of 428 patients referred for suspected sporadic CJD between 1991 and 1997 were classified according to different criteria to be compared after analysis of medical records. Specificity, sensitivity, and positive and negative predictive values were calculated for each set of criteria in the subgroup of patients with a postmortem examination. Positive and negative predictive values of the clinical diagnosis were applied to cases without postmortem examination. Subsequently, the true number of cases of CJD among the referred cases was estimated. Results: By comparison with the French and European study criteria, the Masters’ criteria showed higher sensitivity but lower specificity and positive predictive value. Comparison with an estimate of the true total number of CJD cases showed that Masters’ criteria overestimated the incidence by 7%, whereas the French and the European study criteria led to an underestimate of 12%. Detection of the 14-3-3 protein in CSF, considered as an additional diagnostic criterion for clinically probable CJD, resulted in a slight increase in the estimated incidence when the French or European study criteria were applied. Conclusions: Different diagnostic criteria could lead to an under- or overestimation of the true incidence of CJD. Therefore, comparisons of CJD incidence in different countries should rely on diagnostic classifications using identical criteria. Taking into account 14-3-3 protein detection as a criterion for probable CJD will result in a small increase in the estimated CJD incidence.