Variability in diagnosing Creutzfeldt-Jakob disease using standard and proposed diagnostic criteria

Creutzfeldt-Jakob disease (CJD) is a rapidly progressive dementia with a median survival of 2-14 months. The diagnosis can only be made accurately by biopsy/autopsy. However, this is not always feasible or desirable. Thus, diagnostic criteria have been proposed by UCSF, European MRI-CJD Consortium, and WHO. We will compare these criteria.

Retrospective study of 31 patients (average age of 69.2 years) between 2003 to 2010 by ICD9 codes 046.1, 046.11, and 046.19.

All patients presented with rapidly progressive dementia (mean duration of 4.25 months). Pyramidal and extrapyramidal findings, myoclonus, cerebellar changes, akinetic mutism, and visual disturbances were observed in 6.5-48.4%. Five had periodic pattern on EEG. CSF biomarker 14-3-3 was positive in 11. Tau was positive in 6. Neuron specific enolase was positive in 9. By consensus (kappa = 0.62), MRI was "typical" of CJD in 23 with cortical ribboning (n = 16), basal ganglia hyperintensity (n = 15), or combination (n = 8). By WHO criteria, which does not include neuroimaging, CJD was diagnosed in 10, but 14 if any CSF biomarker was used (p = NS). The UCSF criteria, which does not include CSF biomarkers, diagnosed 18 cases, and the European MRI-CJD Consortium, which includes neuroimaging and CSF biomarkers but with less neurological signs, diagnosed 23 cases (p < 0.05 and p < 0.001, respectively). CJD-mimics included urosepsis, neurosarcoidosis, idiopathic left temporal lobe epilepsy, alcohol intoxication, central nervous system vasculitis, viral encephalitis, and non-Hodgkin's lymphoma.

This study illustrates the variability in diagnosing CJD and emphasizes the diagnostic utility of neuroimaging. It also highlights false-positives that occur with neuroimaging.