Rapidly progressive dementia and CJD: more detail

How is it defined, and what is the pace of assessment?

Decline over weeks to months, occasionally days, in contrast to the years most dementias take. The standard reference chapter states that assessment should happen without delay and usually as an inpatient, precisely because some patients have treatable conditions.

What is the differential?

The chapter organises it with the mnemonic VITAMINS: vascular, infectious, toxic-metabolic, autoimmune, metastases and neoplasm, iatrogenic, neurodegenerative, systemic. Worked examples from each category:

Neurodegenerative: sporadic, genetic and iatrogenic CJD, Alzheimer disease, dementia with Lewy bodies, frontotemporal dementia, corticobasal degeneration, progressive supranuclear palsy. Infectious: herpes simplex encephalitis, HIV-associated dementia, progressive multifocal leukoencephalopathy, subacute sclerosing panencephalitis, neurosyphilis, Lyme disease, Whipple disease. Toxic and metabolic: deficiencies of B12, thiamine and niacin, uraemia, Wilson disease, and bismuth, lithium or mercury toxicity. Autoimmune: Hashimoto encephalopathy, paraneoplastic limbic encephalitis, lupus cerebritis, sarcoidosis. Neoplastic: CNS metastases, primary CNS lymphoma, intravascular lymphoma. Vascular: stroke, thrombotic thrombocytopenic purpura, hyperviscosity syndromes, CNS vasculitis. Psychiatric causes are included, with the caveat that they require a neurological workup first.

Prion disease sits at the top of the list when there is prominent motor or cerebellar dysfunction alongside cognitive decline, and recent registry work has reinforced the difficulty of the parkinsonian presentations specifically, where clinical signs alone separate prion disease from Lewy body disease poorly and MRI and EEG carry much of the discrimination.

What do referral series show?

The UCSF series of about 825 referrals over five years, most carrying a presumed diagnosis of CJD, resolved as 54 percent prion disease (37 percent probable or definite sporadic, 15 percent genetic, 2 percent acquired), 18 percent non-prion conditions, many of them treatable, and 28 percent undetermined, mostly because records were insufficient although most met criteria for possible CJD.

Within the non-prion group the breakdown was 26 percent neurodegenerative, 15 percent autoimmune, 11 percent infectious, 11 percent psychiatric and 9 percent miscellaneous. The source also reports a 28 percent undetermined figure within that group, which overlaps with the cohort-level figure, so the sub-breakdown should be quoted with care.

Two caveats apply before transferring these proportions to a local population. This is a referral series at a specialist prion centre, so prion disease is enriched relative to a general neurology service. And the series predates current seeding assays and the 2017 revision of the surveillance criteria, so the undetermined fraction would be smaller today.

What is current test performance?

Better than the figures in the older series, and still not sufficient alone.

In that series, CSF 14-3-3 had a sensitivity of 48 percent and a specificity of 66 percent, and EEG sensitivity was below 45 percent on first testing. Those numbers are worth knowing because both tests are still ordered and still over-interpreted in both directions.

For CSF RT-QuIC, the US national reference centre states a diagnostic sensitivity of 90.3 percent and specificity of 98.5 percent, citing Rhoads and colleagues in Neurology 2020. Its operational requirements are important for a rapid workup: patient aged 12 or over, 2 ml of CSF with the first 2 ml of flow discarded, a non-bloody tap, immediate freezing at minus 20 degrees or colder, shipping frozen Monday to Thursday, and a 5 business day turnaround from receipt, with no STAT option. Bloody or discoloured samples cannot be run.

Assay choice changes the answer in genetic disease. The 2026 Canadian comparison on 116 diagnostic CSF samples found second-generation IQ-CSF RT-QuIC at 72.4 percent sensitivity across a panel enriched for inherited cases, against 87.1 percent for a hamster end-point assay and 96.6 percent for a modified assay using a deer mouse substrate. All 61 sporadic CJD samples were positive on both end-point assays, so the shortfall was concentrated in the inherited subtypes, with none of 13 fatal familial insomnia cases detected by IQ-CSF. A negative CSF seeding result in a patient with a family history should therefore prompt a question about which substrate was used.

Blood markers are at the validation stage. A 2021 study measured plasma neurofilament light and total tau in 83 sporadic CJD cases against healthy controls and six comparison groups. Both markers discriminated CJD from non-CJD dementias with an AUC of 0.93. Against rapidly progressive Alzheimer disease, NfL gave an AUC of 0.95 and total tau 0.92, and the difference was not statistically significant, so NfL should not be described as the more accurate of the two in that comparison. Total tau was raised only in CJD, while NfL was also elevated in several non-prion dementias, which limits its specificity. Plasma total tau was weakly associated with disease duration, with correlation coefficients around minus 0.32 and model concordance about 0.6; NfL showed no prognostic value. The authors stated that their proposed cutoffs need validation in independent cohorts before clinical use, and the study had no external validation cohort.

How prion disease is diagnosed covers the diagnostic panel, and the suspected case pathway the referral route.

What should the workup include?

From the reference chapter: screen every dementia patient for B12 deficiency; consider HIV testing in every rapidly progressive dementia evaluation; test for syphilis with serum RPR or VDRL and CSF VDRL; CT of the body with and without contrast to look for sarcoidosis, malignancy and paraneoplastic disease; tumour markers and anti-thyroid peroxidase antibodies where autoimmune causes are suspected; and serum and CSF to a research laboratory for specialist antibody panels when autoimmune encephalopathy is strongly suspected.

Three procedural points from the same source. Biopsy is often what makes the diagnosis. Where intravascular lymphoma is suspected, go to biopsy rather than angiogram, and avoid steroids before a lymphoma biopsy. Exclude tuberculosis before starting immunosuppression for sarcoidosis. Prion precautions should be used when prion disease is possible, which in practice means for brain, spinal cord and posterior eye tissue rather than for routine care.

What should happen alongside the workup?

Symptom management and planning, concurrently rather than afterwards. The 2026 multidisciplinary recommendations advise palliative involvement early and future care planning as soon as possible, ideally at diagnosis, because capacity is often lost during the diagnostic process itself. Where no family is available, an advocate such as an Independent Mental Capacity Advocate may be appointed.

Families should also be told that there is no evidence of person-to-person transmission through regular social and clinical contact, and that routine care needs no infection control beyond standard practice.

Where next?

Rapidly progressive dementia and CJD is the plain version for families. How prion disease progresses: more detail covers the course after diagnosis.

This page summarises published literature and is not clinical advice. Investigation and management of an individual patient belong with the treating team and the national prion centre or reference laboratory.