Etiologies of Rapidly Progressive Dementia: A Cross-Sectional Multicentric Tertiary Hospital Study
Abstract
Background: Rapidly progressive dementia (RPD) is linked to several disorders. Reference centers for Creutzfeldt–Jakob disease have reported prion diseases as the most frequent cause of RPD. Alternatively, tertiary centers found high frequencies of nonprion diseases causing RPD, including a significant number of potentially treatable disorders. This study is aimed at describing the etiologies of RPD among inpatients from four tertiary centers in Florianópolis, Brazil.
Methods: We reviewed medical records of patients hospitalized from 2001 to 2020 under codes of the 10th revision of the International Classification of Diseases potentially related to RPD. Patients who fulfilled the criteria for RPD were selected. RPD was defined when (1) the criteria of the National Institute on Aging–Alzheimer’s Association for dementia were met and (2) the disease duration was 730 days or less. We revised all etiological diagnoses according to established clinical criteria.
Results: One hundred four patients with RPD were identified. The most frequent diagnostic groups were dementias of vascular origin (N = 27, 26.0%), nonprion degenerative diseases (N = 24, 23.0%), and mixed dementias (N = 9, 8.7%). All 11 patients with disease duration between 1 and 2 years had either dementia of vascular origin or nonprion degenerative dementia. Ten patients (9.6%)—diagnosed with autoimmune disorders, infectious disorders, neoplasia, or posterior reversible encephalopathy syndrome—showed improvement.
Conclusions: Dementias of vascular origin were the most frequent cause of RPD in our sample, which differs from what has been previously reported by the literature. Patients with dementia duration between 1 and 2 years seem to have disorders usually associated with slow progression, while all potentially reversible cases had a duration of 1 year or less. Thorough investigation of RPD is paramount for the identification of potentially reversible etiologies.
1. Introduction
Dementia is characterized by cognitive or behavioral symptoms that impair one’s functionality, represent a decline from a previous level of performance, and are not caused by delirium or a major psychiatric illness [1]. Rapidly progressive dementia (RPD) is a syndrome in which the development of dementia takes an accelerated course [2, 3], generally developing over weeks to months [3]. There is not a universally accepted chronologic cutoff point to differentiate RPD from other forms of dementia [3–5]. However, it is usually accepted that, in RPD, the time lapse between the onset of symptoms and functional impairment is inferior to 1–2 years [2–5].
RPD is linked to a heterogeneous group of disorders [2–5]. Classically associated with prion diseases, of which Creutzfeldt–Jakob disease (CJD) is the most common, RPD may be caused by nonprion degenerative diseases, autoimmune disorders, infectious diseases, toxic–metabolic conditions, vascular disorders, and neoplasms [2–5]. Studies on RPD from centers of reference for the diagnosis of CJD have found prion diseases to be the most prevalent cause of RPD [6–8]. Nonetheless, several patients with suspected CJD were found to have alternative diagnoses [6–8], including a significant number of patients diagnosed with potentially treatable disorders [7]. Studies from reference centers for CJD tend to include a higher proportion of irreversible RPD cases since most patients are referenced on the hypothesis of a prion disease [4]. Therefore, research studies from such centers may overestimate the frequency of prion diseases among patients with RPD [4, 5].
Several tertiary center studies have tried to address this issue by recruiting patients according to clinical criteria to define both RPD syndrome and etiological diagnoses [9–14]. These studies found high frequencies of nonprion diseases causing RPD, including many potentially reversible disorders [10–14]. According to tertiary center research, nonprion degenerative diseases, autoimmune disorders, toxic–metabolic disorders, vascular diseases, neoplasia, and infectious diseases are responsible for many cases of RPD [10–14].
A large part of RPD patients seems to suffer from potentially reversible conditions. Therefore, it is crucial to better understand the diagnostic profile of RPD and its etiologies. Better characterization of the epidemiology of RPD-causing disorders may increase clinicians’ awareness of the importance of thorough clinical investigation of RPD, prompting a better standard of care for patients with RPD.
This study is aimed at determining the relative frequency of different etiologies of RPD among inpatients from four tertiary centers in Florianópolis, in southern Brazil.
2. Methods
2.1. Patient Selection
This was a cross-sectional multicentric study. Four tertiary hospitals from the metropolitan area of Florianópolis, Santa Catarina, Brazil, were included: Hospital Universitário Polydoro Ernani de São Thiago–Universidade Federal de Santa Catarina (HU-UFSC), Hospital Governador Celso Ramos (HGCR), Hospital Regional Homero de Miranda Gomes (HRHMG), and Instituto de Psiquiatria (IPQ). These are major public reference centers to which patients with dementia are referenced from all over the state of Santa Catarina, which has a population of over 6 million people [15].
The four centers’ computerized systems of medical hospitalizations were screened for medical records of 18-year-old patients or older who were hospitalized from January 2001 to December 2020 under specific codes of the 10th revision of the International Classification of Diseases (ICD-10) [16] which could be related to a RPD syndrome (complete list of ICD-10 codes available in the Supporting Information (available here)). The authors analyzed such medical records. When the initial search in the computerized system of a given center revealed more than 1500 medical records of interest, a random sample of 1500 from the center was selected for analysis. The list of medical records was double-checked for duplicates.
The authors thoroughly reviewed the selected medical records, and those patients who fulfilled the criteria for RPD were selected. RPD was defined as any clinical syndrome in which (1) the criteria of the National Institute on Aging–Alzheimer’s Association (NIA-AA) [1] for dementia of any cause were met and (2) disease duration (defined as the time from symptom onset to characterization of functional impairment) was of 730 days or less. Patients whose cognitive or behavioral symptoms were attributed to delirium or a primary psychiatric diagnosis during hospitalization were excluded.
2.2. Medical Record Analysis and Diagnostic Revision
The authors retrieved relevant clinical information from each patient, including age at the time of hospitalization, age of symptom onset, gender, symptoms reported by patient or caregiver, comorbidities, drugs in use, physical examination, cognitive testing results, brain imaging scans, laboratory testing including cerebrospinal fluid (CSF) analysis, electroencephalogram (EEG), etiologic diagnosis registered on the medical record, and outcome.
All etiologic diagnoses were then revised by the authors according to established clinical criteria for the diagnosis of several etiologies of dementia. For the diagnosis of CJD, Alzheimer’s disease (AD), dementia with Lewy bodies (DLB), Parkinson disease dementia (PDD), behavioral variant of frontotemporal dementia (bvFTD), primary progressive aphasia (PPA), progressive supranuclear palsy (PSP), corticobasal degeneration (CBD), multiple system atrophy (MSA), vascular dementia (VD), and autoimmune encephalopathy, the following criteria were, respectively, employed: European Magnetic Resonance Imaging—Creutzfeldt–Jakob Disease Consortium Criteria [17], NIA-AA criteria for AD [1], fourth consensus criteria of the DLB Consortium [18], Movement Disorder Society criteria for PDD [19], international consortium criteria for bvFTD [20], 2011 criteria for the classification of PPA and its variants [21], Movement Disorder Society criteria for PSP [22], 2013 criteria for the diagnosis of CBD [23], criteria of the second consensus for the diagnosis of MSA [24], criteria of the National Institute of Neurological Disorders and Stroke–Association Internationale pour la Recherche et l’Enseignement en Neurosciences (NINDS-AIREN) for the diagnosis of VD [25], and Graus’ criteria for the diagnosis of autoimmune encephalitis [26]. When a certain patient did not meet any of the aforementioned criteria, the authors judged each case individually.
The relative prevalence of different etiological diagnoses both at hospitalization (initial diagnosis) and after diagnostic revision (revised diagnosis) was analyzed and reported. The initial diagnosis was that made by the attending medical teams in each of the four centers during hospitalization and was obtained from the medical records. The initial diagnoses were reported because, even though it was not possible to guarantee that they had been made strictly according to established clinical criteria, they were nonetheless made by clinical teams directly involved with the patients’ medical assistance, while the revised diagnoses were made retrospectively based on reported clinical and complementary tests data. For the inferential analyses, only the revised diagnoses were employed, since they adhered to commonly used clinical criteria.
2.3. Statistical Analysis
Continuous variables were expressed as “mean (standard deviation (SD)),” while categorical variables were expressed as “absolute number (percentage).” For hypothesis testing, Student’s t-test was employed for continuous variables. Age of onset and disease duration were compared across different etiologic groups (after diagnostic revision). Patients were further divided into two groups: (1) RPD patients with disease duration equal to or inferior to 365 days and (2) RPD patients with disease duration superior to 365 days and equal to or inferior to 730 days. The distribution of revised etiological groups of RPD was compared between both groups. We investigated possible associations between etiological group and RPD clinical characteristics and complementary test findings. For categorical variables, we employed chi-square test or Fisher’s exact test, as appropriate. SPSS Version 28.0 was used for statistical analysis. A p value of 0.050 was used for statistical significance.
2.4. Ethical Standards
This study was developed according to the Declaration of Helsinki and approved by the Federal University of Santa Catarina Ethics Committee, which waived the need for informed consent, under Number CAAE 40477420.9.0000.0121. PAR Inc. and Editora Hogrefe CETEPP have granted permission to use the Brazilian Portuguese version of the Mini-Mental State Examination (MMSE) [27] in this study.
2.5. Data Sharing and Accessibility
The data supporting this study’s findings are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
3. Results
3.1. Patient Selection and Sample Characteristics
The search through the computerized systems revealed 14,096 medical records of patients hospitalized under an ICD-10 code of interest (1421 from HU-UFSC, 6595 from HGCR, 5787 from HRHMG, and 293 from IPQ). Since HGCR and HRHMG had more than 1500 medical records found, a random sample of 1500 medical records was selected from each. There was no difference of age at the time of hospitalization between randomized and nonrandomized patients from HRHMG—61.5 (16.76) years × 61.9 (16.47) years (p = 0.420)—and neither any difference of gender (54.4% male × 56.1% male, p = 0.264). As for HGCR, there was a slight difference in age between randomized and nonrandomized patients—60.0 (16.98) × 58.9 (17.84) years (p = 0.037)—but no difference in gender (53.4% male × 52.3% male, p = 0.480).
Therefore, 4714 medical records were searched for RPD patients. Twenty-nine records could not be localized (0.6% of all selected medical records). The analysis of the remaining 4685 medical records identified 104 individuals who met our working criteria for RPD (48 from HU-UFSC, 26 from HGCR, 14 from HRHMG, and 16 from IPQ) and were selected. Sixty-five (62.5%) patients were male. The mean age of onset was 63.7 (15.56) years, and the mean disease duration was 188.1 (196.84) days. Ninety-three (89.4%) patients had a disease duration equal to or inferior to 365 days, and 11 (10.6%) had a disease duration superior to 365 days and equal to or inferior to 730 days. Only 29 (27.9%) medical records contained information regarding patient education, with a mean of 6.5 (4.44) years of formal schooling. Clinical and demographical characteristics are shown in Table 1.
| HU-UFSC | IPQ | HRHMG | HGCR | Total | |
|---|---|---|---|---|---|
| Gender† | |||||
| Male | 33 (68.8) | 11 (68.8) | 7 (50.0) | 14 (53.8) | 65 (62.5) |
| Female | 15 (31.3) | 5 (31.3) | 7 (50.0) | 12 (46.2) | 39 (37.5) |
| Age of onset (years)‡ | 65.8 (15.60) | 66.1 (12.64) | 64.6 (15.73) | 57.6 (16.26) | 63.7 (15.56) |
| Disease duration (days)‡ | 194.6 (205.24) | 280.0 (218.62) | 238.7 (223.26) | 92.2 (96.74) | 188.1 (196.84) |
| Disease duration† | |||||
| ≤ 365 days | 42 (87.5) | 13 (81.3) | 12 (85.7) | 26 (100.0) | 93 (89.4) |
| > 365 days | 6 (12.5) | 3 (18.8) | 2 (14.3) | 0 (0.0) | 11 (10.6) |
- Abbreviations: HGCR, Hospital Governador Celso Ramos; HRHMG, Hospital Regional Homero de Miranda Gomes; HU-UFSC, Hospital Universitário Professor Polydoro Ernani de São Thiago–Universidade Federal de Santa Catarina; IPQ, Instituto de Psiquiatria.
- †Data shown as “absolute number (percentage).”
- ‡Data shown as “mean (standard deviation).”
3.2. Clinical Manifestations and Complementary Tests
The most frequently reported cognitive–behavioral symptoms among RPD patients (referred by patients or caregivers) were memory impairment (51.0%), confusion (47.1%), agitation (45.2%), disorientation (27.9%), aggressiveness (25.0%), psychotic symptoms (25.0%), and disinhibition (20.2%). Commonly patient- or caregiver-reported noncognitive–behavioral symptoms included gait impairment (31.7%), new-onset epileptic seizures (18.3%), sphincter control impairment (17.3%), muscular weakness (12.5%), involuntary movements (11.5%), loss of consciousness (10.6%), dysphagia (10.6%), and headache (10.6%).
The most frequent findings on bedside cognitive and mental status examination as reported by attending medical teams were disorientation (67.3%), confusion (22.1%), memory impairment (18.3%), global aphasia (17.3%), drowsiness (16.3%), and attention impairment (14.4%). Alternatively, the most frequent findings on general and neurologic examination were muscular weakness (44.2%), hyperreflexia (26.0%), pyramidal signs (22.1%), appendicular ataxia (16.3%), dysarthria (15.4%), tremor (14.4%), and muscle stiffness (13.5%).
Thirty-nine (37.5%) patients were submitted to either standardized cognitive testing or formal neuropsychological testing. Reasons for the absence of standardized cognitive testing were usually not reported in the medical records. Thirty-eight (36.5%) patients were submitted to the MMSE [27] (mean score: 15.2; SD: 4.90); three (2.9%) to the Montreal Cognitive Assessment [28] (mean score: 12.0; SD: 10.44); three (2.9%) to tasks from the Consortium to Establish a Registry for Alzheimer’s Disease (CERAD) [29], including the list of words memorization (mean score: 10.7; SD: 3.79), evoking (mean score: 2.3; SD: 2.08), and recognition (mean score: 4.3; SD: 1.53) and semantic verbal fluency (mean score: 9.0; SD: 6.24) tasks; and two (1.9%) to the clock drawing test scored according to Sunderland [30] (mean score: 5.0; SD: 1.41). Formal neuropsychological testing was performed with eight (7.7%) patients.
Eighty-one (77.9%) patients had a computer tomography (CT) brain scan. Frequent findings included diffuse brain atrophy (46.9%), gliosis (40.7%), cerebral microangiopathy (39.5%), ventriculomegaly (17.3%), and intracranial atherosclerosis (12.3%). Brain magnetic resonance imaging (MRI) was performed in 62 (59.6%) patients, showing cerebral microangiopathy (58.1%), diffuse brain atrophy (50.0%), gliosis (19.4%), temporal lobe atrophy (17.7%), hippocampal atrophy (14.5%), acute ischemic lesions (11.3%), ischemic lacunae (11.3%), and ventriculomegaly (11.3%) as the most frequent findings.
Twenty-six (25.0%) patients had an EEG performed. The most frequent EEG findings were disorganized base activity (61.5%), epileptiform discharges (19.2%), and triphasic waves (15.4%). CSF examination was available for 44 (42.3%) patients. The mean protein concentration was 68.6 mg/dL (SD: 73.89). Mean cell count was 15.2 cells/mm3 (SD: 42.97). High protein concentration (characterized by values higher than 45.0 mg/dL) was found in 47.6%, while CSF pleocytosis (characterized by values higher than 4.0 cells/mm3) was present in 27.9%. CSF protein 14-3-3 determination was available for three patients (2.9%). One (1.0%) patient had CSF amyloid beta, total tau, and phosphorylated tau determinations.
3.3. Etiological Diagnoses
The etiological diagnoses at the time of medical hospitalization are shown in Table 2. At the time of medical hospitalization, the most frequently identified diagnostic groups of RPD were dementias of vascular origin (N = 23, 22.1%), mixed dementias (N = 19, 18.3%), and nonprion degenerative diseases (N = 13, 12.5%). The most common etiological diagnoses were VD (N = 21, 20.2%), CJD (N = 8, 7.7%), and bvFTD (N = 4, 3.9%) and the combination of AD and VD (N = 4, 3.9%). Nine (8.7%) patients had no etiological diagnosis or diagnostic group recorded.
| Diagnosis | At hospitalization (%) | After diagnostic revision (%) |
|---|---|---|
| Prion diseases | 8 (7.7) | 8 (7.7) |
| Creutzfeldt–Jakob disease | 8 (7.7) | 8 (7.7) |
| Nonprion degenerative diseases | 13 (12.5) | 24 (23.0) |
| AD | 4 (3.8) | 7 (6.7) |
| DLB | 1 (1.0) | 2 (1.9) |
| bvFTD | 4 (3.8) | 5 (4.8) |
| Possible AD or possible DLB | — | 4 (3.8) |
| Other | 4 (3.8) ∗ | 6 (5.8)‡ |
| Dementia of vascular origin | 23 (22.1) | 27 (26.0) |
| VD | 21 (20.2) | 23 (22.1) |
| Possible VD | — | 2 (1.9) |
| Hypoxic–ischemic encephalopathy | 1 (1.0) | 1 (1.0) |
| PRES | 1 (1.0) | 1 (1.0) |
| Toxic–metabolic disorders | 4 (3.9) | 6 (5.8) |
| Wernicke–Korsakoff syndrome | 3 (2.9) | 2 (1.9) |
| Alcoholic dementia | 1 (1.0) | 2 (1.9) |
| Marchiafava–Bignami disease | — | 1 (1.0) |
| Osmotic demyelination syndrome | — | 1 (1.0) |
| Normal pressure hydrocephalus | 3 (2.9) | 2 (1.9) |
| Autoimmune disorders | 8 (7.7) | 8 (7.7) |
| Defined anti-NMDAR encephalitis | 1 (1.0) | 1 (1.0) |
| Anti-LGI1 limbic encephalitis | 2 (1.9) | 1 (1.0) |
| Probable seronegative AE | — | 1 (1.0) |
| Possible AE | 3 (2.9) | 4 (3.8) |
| Possible neurosarcoidosis | 1 (1.0) | 1 (1.0) |
| Paraneoplastic encephalitis | 1 (1.0) | — |
| Neoplasia | 7 (6.7) | 7 (6.7) |
| Glioblastoma | 3 (2.9) | 3 (2.9) |
| Other | 4 (3.8) ∗∗ | 4 (3.8) ∗∗ |
| Dementia associated with TBI | 3 (2.9) | 5 (4.8) |
| Posttraumatic dementia | 3 (2.9) | 4 (3.8) |
| Chronic subdural hematoma | — | 1 (1.0) |
| Infectious disorders | 7 (6.7) | 8 (7.7) |
| Herpetic encephalitis | 2 (1.9) | 2 (1.9) |
| Other | 5 (4.8) ∗∗∗ | 6 (5.8)§ |
| Mixed dementia | 19 (18.3) | 9 (8.7) |
| VD and possible AD | 4 (3.9) | 5 (4.8) |
| Other | 15 (14.4)† | 4 (3.8)¶ |
| Indeterminate | 9 (8.7) | — |
| Total | 104 (100.0) | 104 (100.0) |
- Note: Bold was used for total frequencies of each etiological group.
- Abbreviations: AD, Alzheimer’s disease; AE, autoimmune encephalitis; bvFTD, frontotemporal dementia, behavioral variant; DLB, dementia with Lewy bodies; LGI1, leucine-rich-glioma-inactivated 1 protein; NMDAR, N-methyl-D-aspartate receptor; PRES, posterior reversible encephalopathy syndrome; TBI, traumatic brain injury; VD, vascular dementia.
- ∗One case of Parkinson’s disease dementia (PDD), one of primary progressive aphasia (PPA), one of progressive supranuclear palsy (PSP), and one of Huntington’s disease.
- ∗∗One case of anaplastic astrocytoma, one of meningioma, one of craniopharyngioma, and one of cerebral metastasis.
- ∗∗∗One case of neurotuberculosis and neurotoxoplasmosis; one of acquired immunodeficiency syndrome (AIDS) dementia complex, progressive multifocal leukoencephalopathy, and neurocryptococcosis; and one of AIDS dementia complex and neurotoxoplasmosis.
- †One case of vascular dementia (VD) and dementia with Lewy bodies (DLB); one of VD and posttraumatic dementia (PTD); one of VD and alcoholic dementia; one of VD and neurosyphilis; one of VD and Hashimoto’s encephalopathy; one of VD and neoplasia; one of Alzheimer’s disease (AD) and normal pressure hydrocephalus (NPH); one of AD and unspecified toxic–metabolic disorder; one of NPH, Wernicke encephalopathy, subacute combined degeneration, and alcoholic dementia; one of VD, alcoholic dementia, Vitamin B12 deficiency, and Vitamin B1 deficiency; and one of Wernicke encephalopathy, subacute combined degeneration, alcoholic dementia, PTD, and neurotuberculosis.
- ‡One case of possible frontotemporal dementia, behavioral variant (bvFTD); one of PSP; one of corticobasal degeneration (CBD); one of Huntington’s disease; one of possible AD, possible DLB, or possible bvFTD; and one of possible AD, possible LBD, or possible CBD.
- §One case of neurotuberculosis, one of neurocryptococcosis, one of brain abscess, one of AIDS dementia complex, one of neurotuberculosis and neurotoxoplasmosis, and one of AIDS dementia complex and neurotuberculosis.
- ¶One case of VD and DLB, one of VD and PTD, one of VD and possible AD or possible DLB, and one of AD and NPH.
After diagnostic revision, 31 (29.8%) patients did not fulfill any of the sets of criteria listed in the Methods section. Seven patients were diagnosed with dementia due to neoplasia because their cognitive impairment could be safely attributed to one or more intracranial neoplastic lesions identified on brain imaging. Four patients diagnosed with posttraumatic dementia had a clear temporal relationship between symptom onset and a traumatic brain injury, and all had abnormal brain imaging scans. Two were diagnosed with Wernicke–Korsakoff syndrome, both with a history of alcohol abuse; one had an acute onset encephalopathy with typical bithalamic and hypothalamic hyperintensities on brain MRI, while the other had both an acute onset encephalopathy with ataxia but unremarkable brain imaging. Two patients diagnosed with alcoholic dementia had a long history of alcohol abuse and had neither eye movement abnormalities nor ataxia, and both had unremarkable complementary tests. Two patients diagnosed with normal pressure hydrocephalus (NPH) presented with the triad of gait impairment, urinary incontinence, and dementia and showed ventriculomegaly on brain imaging; one of them showed partial improvement after tap test, and the other had reduced sulci at the vertex on MRI. Two patients presenting with fever, an acute onset of behavioral and cognitive changes, and CSF pleocytosis had positive polymerase chain reaction (PCR) testing for Herpes simplex. Two patients diagnosed with neurotoxoplasmosis had typical nodular lesions with ring enhancement on brain scans; one of them who also had HIV and showed cognitive worsening despite neurotoxoplasmosis adequate treatment was diagnosed also with AIDS dementia complex, while the other had a concomitant disseminated tuberculosis with CSF elevated protein, pleocytosis, and low glucose. Two patients diagnosed with neurotuberculosis showed CSF with elevated protein, pleocytosis, and low glucose; one of them was HIV-positive and did not improve despite adequate treatment for tuberculosis and was also diagnosed with AIDS dementia complex. One developed cognitive and behavioral symptoms temporally related to a cardiopulmonary arrest and was diagnosed with hypoxic–ischemic encephalopathy. Another had a subacute course with brain imaging showing transient bilateral vasogenic edema in the watershed regions of the parietal lobes and was diagnosed with posterior reversible encephalopathy syndrome (PRES). One patient diagnosed with Marchiafava–Bignami disease had an acute onset encephalopathy with typical callosal lesions on brain MRI. One patient diagnosed with osmotic demyelination syndrome had a history of alcohol, cocaine, and crack abuse; had sodium and potassium disturbances; and had signs of pontine and extrapontine demyelination on brain MRI. One patient was diagnosed with possible neurosarcoidosis; over the course of a few months, he developed an encephalopathy with several nodular enhancing lesions on brain MRI suggestive of granulomatous disease, combined with elevated CSF protein and pleocytosis, while infectious diseases were excluded. One patient with subacute presentation showed a chronic subdural hemorrhage on brain CT. One patient developed dementia over a few months and was found to have a brain abscess through brain imaging; cultures of the abscess material were positive for Nocardia. One patient had CSF elevated protein and pleocytosis and was positive for Cryptococcus neoformans.
After the diagnostic revision, 43 (41.3%) of the RPD patients had their diagnoses modified. Seventeen (16.3%) patients initially diagnosed with mixed dementia had their diagnoses changed for a single cause of dementia. The diagnoses of five (4.8%) patients were modified from a single cause to mixed dementia. A diagnosis could be attributed to seven (6.7%) patients who previously did not have a specific diagnosis registered on their medical records. Three (2.9%) patients initially diagnosed as having probable AD, probable CJD, and probable PDD did not meet the criteria for those conditions and had their diagnoses modified to possible AD or possible LBD, possible AD or possible LBD or possible CBD, and possible AD or possible LBD, respectively. Two (1.9%) patients initially diagnosed with anti-LGI1 limbic encephalitis and paraneoplastic encephalitis had their diagnoses modified to seronegative autoimmune encephalitis and possible autoimmune encephalitis, respectively. One patient had a diagnosis change from Wernicke encephalopathy to osmotic demyelination syndrome and another from PPA to bvFTD. Eight (7.7%) patients underwent modifications of both etiological diagnosis and diagnostic group: one from CJD to possible AD or possible DLB, one from alcoholic dementia to posttraumatic dementia, one from bvFTD to the combination of LBD and VD, two from NPH to VD, one from possible autoimmune encephalitis to CJD, one from VD to bvFTD, and one from VD to NPH.
The most frequent revised diagnostic groups were dementias of vascular origin (N = 27, 26.0%), nonprion degenerative diseases (N = 24, 23.0%), and mixed dementias (N = 9, 8.7%). The most frequent revised etiological diagnoses were probable VD (N = 23, 22.1%), probable CJD (N = 8, 7.7%), and probable AD (N = 7, 6.7%). Table 2 compares the relative frequency of different diagnostic groups before and after diagnostic revision.
Since many patients did not have MRI or CSF results, we separately analyzed the etiological profile of patients who had both MRI and CSF results. In this subgroup (N = 33), the most frequent cause of RPD was autoimmune disorders (N = 7, 21.2%), followed by prion diseases (N = 5, 15.2%), dementias of vascular origin (N = 5, 15.2%), infectious disorders (N = 5, 15.2%), nonprion degenerative diseases (N = 3, 9.1%), toxic–metabolic disorders (N = 3, 9.1%), mixed dementia (N = 3, 9.1%), NPH (N = 1, 3.0%), and dementia associated with traumatic brain injury (N = 1, 3.0%). There were no patients with neoplasia in this subgroup. Patients in this subgroup were mostly male (N = 20, 60.6%); the mean age at onset was 58.6 (15.15) years and the mean disease duration was 104.7 (100.51) days. No patient in this subgroup had a disease duration superior to 365 days.
RPD patients with revised diagnoses of autoimmune disorders and infectious disorders had an earlier age of onset, while those with nonprion degenerative diseases and dementias of vascular origin were older at disease onset (Table 3). Disease duration was shorter among patients with prion diseases and autoimmune disorders and longer among patients with nonprion degenerative diseases and mixed dementia (Table 3). Gender was not significantly associated with etiological group (data not shown).
| Diagnosis | Age of onset (years)† | p‡ | Disease duration (days)† | p‡ |
|---|---|---|---|---|
| Prion diseases | 65.4 (12.54) | 0.749 | 91.4 (48.85) | < 0.001 ∗∗ |
| Other | 63.5 (15.83) | 196.1 (202.43) | ||
| Nonprion degenerative diseases | 68.9 (12.78) | 0.021 ∗ | 333.5 (216.20) | < 0.001 ∗∗ |
| Other | 61.4 (16.19) | 123.4 (148.40) | ||
| Dementia of vascular origin | 68.8 (10.67) | 0.006 ∗∗ | 195.4 (219.26) | 0.787 |
| Other | 61.1 (17.01) | 184.3 (186.03) | ||
| Toxic–metabolic disorders | 60.0 (10.49) | 0.554 | 98.7 (104.43) | 0.254 |
| Other | 63.9 (15.83) | 193.5 (200.14) | ||
| Normal pressure hydrocephalus | 75.0 (7.94) | 0.202 | 303.3 (106.81) | 0.306 |
| Other | 63.3 (15.62) | 184.6 (198.16) | ||
| Autoimmune disorders | 53.0 (13.43) | 0.043 ∗ | 55.5 (57.32) | < 0.001 ∗∗ |
| Other | 64.6 (15.45) | 199.1 (200.41) | ||
| Neoplasia | 66.0 (15.93) | 0.684 | 177.1 (142.30) | 0.880 |
| Other | 63.5 (15.60) | 188.9 (200.74) | ||
| Dementia associated with traumatic brain injury | 55.3 (23.97) | 0.177 | 90.7 (177.59) | 0.213 |
| Other | 64.2 (14.93) | 194.0 (197.21) | ||
| Infectious disorders | 40.0 (13.18) | < 0.001 ∗∗ | 97.9 (95.44) | 0.179 |
| Other | 65.6 (14.09) | 195.6 (201.48) | ||
| Mixed dementia | 63.7 (17.09) | 0.990 | 309.8 (226.42) | 0.029 ∗ |
| Other | 63.7 (15.47) | 173.7 (189.27) |
- ∗Statistical significance for p < 0.050.
- ∗∗Statistical significance for p < 0.010.
- †Data shown as “mean (standard deviation).”
- ‡Student’s t-test.
There was positive association between prion diseases and involuntary movements (odds ratio (OR) = 21.19; p < 0.001); nonprion degenerative diseases and aggressiveness (OR = 3.10; p = 0.014), psychosis (OR = 3.10; p = 0.014), and disinhibition (OR = 3.25; p = 0.016); dementia of vascular origin and agitation (OR = 2.48; p = 0.031); NPH and sphincter incontinence (OR = 39.06; p = 0.017); neoplasia and headache (OR = 8.34; p = 0.025); dementia associated with traumatic brain injury and consciousness impairment (OR = 11.25; p = 0.015); and infectious disorders and headache (OR = 25.00; p < 0.001) and new-onset epileptic seizures (OR = 5.40; p = 0.035). Nonprion degenerative disease associated negatively with new-onset epileptic seizures (OR = 0.10; p = 0.006). Regarding cognitive and physical examination of RPD patients, there was positive association between prion diseases and spasticity (OR = 5.80; p = 0.048) and hyperreflexia (OR = 10.71; p = 0.004); dementia of vascular origin and pyramidal signs (OR = 2.75; p = 0.033) and spasticity (OR = 7.62; p = 0.002); autoimmune disorders and memory impairment (OR = 5.40; p = 0.035), pyramidal signs (OR = 7.22; p = 0.012), and hypoesthesia (OR = 5.80; p = 0.048); and dementia associated with traumatic brain injury and confusion (OR = 8.32; p = 0.021) and inattention (OR = 15.82; p = 0.004). Alternatively, there was negative association between nonprion degenerative diseases and pyramidal signs (OR = 0.16; p = 0.010) and muscular weakness (OR = 0.18; p = 0.001) and dementia of vascular origin and appendicular ataxia (OR = 0.22; p = 0.049). Other RPD clinical characteristics did not associate significantly with revised diagnosis (data not shown).
Some brain CT findings were associated with etiological group. There was positive association between nonprion degenerative diseases and diffuse atrophy (OR = 3.74; p = 0.014), dementia of vascular origin and gliosis (OR = 5.25; p < 0.001) and microangiopathy (OR = 4.62; p = 0.001), and dementia associated with traumatic brain injury and gliosis (OR = 8.39; p = 0.039). Infectious disorders were negatively associated with diffuse atrophy (OR = 0.05; p = 0.048). On brain MRI, there was positive association between nonprion degenerative dementia and diffuse atrophy (OR = 4.88; p = 0.021) and temporal lobe atrophy (OR = 7.18; p = 0.006); dementia of vascular origin and diffuse atrophy (OR = 4.26; p = 0.040), gliosis (OR = 6.38; p = 0.008), microangiopathy (OR = 7.64; p = 0.007), and ischemic lacunae (OR = 10.00; p = 0.010); NPH and ventriculomegaly (OR = 86.33; p = 0.005); and mixed dementia and gliosis (OR = 7.83; p = 0.022). Other imaging findings showed no statistically significant association with etiological group (data not shown). CSF pleocytosis associated positively with autoimmune disorders (OR = 10.36; p = 0.012) and infectious disorders (OR = 6.67; p = 0.028). There was no statistically significant association between EEG findings and etiological group (data not shown).
Of the 11 patients with disease duration superior to 365 days and equal to or inferior to 730 days, six (54.4%) had nonprion degenerative diseases (one probable AD, one bvFTD, one CBD, two possible AD or possible DLB, and one possible AD or possible LBD or possible CBD), two (18.2%) had dementia of vascular origin, and three (27.3%) had mixed dementia (two with the combination of VD and possible AD and one with the combination of VD and posttraumatic dementia). The comparison of the distribution of revised diagnoses of RPD according to disease duration is shown in Table 4.
| Etiologic group | Duration ≤ 365 days (%) | Duration > 365 and ≤ 730 days (%) |
|---|---|---|
| Prion diseases | 8 (8.6) | — |
| Nonprion degenerative diseases | 18 (19.4) | 6 (54.5) |
| Dementia of vascular origin | 25 (26.9) | 2 (18.2) |
| Toxic–metabolic disorders | 6 (6.4) | — |
| Normal pressure hydrocephalus | 2 (2.2) | — |
| Autoimmune disorders | 8 (8.6) | — |
| Neoplasia | 7 (7.5) | — |
| Dementia associated with TBI | 5 (5.4) | — |
| Infectious disorders | 8 (8.6) | — |
| Mixed dementia | 6 (6.4) | 3 (27.3) |
| Total | 93 (100.0) | 11 (100.0) |
- Abbreviation: TBI, traumatic brain injury.
In total, 10 (9.6%) RPD patients showed cognitive improvement. Two (1.9%) patients, diagnosed, respectively, with anti-LGI1 limbic encephalitis and brain abscess, had a complete or near-complete recovery. The remaining eight (7.7%) showed partial improvement: one with PRES, one with defined anti-NMDAR encephalitis, one with probable seronegative autoimmune encephalitis, one with possible autoimmune encephalitis, one with meningioma, one with anaplastic astrocytoma, one with neurocryptococcosis, and one with herpetic encephalitis. Table 5 shows the percentage of patients of each etiological group that showed any cognitive improvement.
| Etiologic group | Improvement | No improvement |
|---|---|---|
| Prion diseases | 0 (0.0) | 8 (100.0) |
| Nonprion degenerative diseases | 0 (0.0) | 24 (100.0) |
| Dementia of vascular origin | 1 (3.7) | 26 (96.3) |
| Toxic–metabolic disorders | 0 (0.0) | 6 (100.0) |
| Normal pressure hydrocephalus | 0 (0.0) | 2 (100.0) |
| Autoimmune disorders | 4 (50.0) | 4 (50.0) |
| Neoplasia | 2 (28.6) | 5 (71.4) |
| Dementia associated with traumatic brain injury | 0 (0.0) | 5 (100.0) |
| Infectious disorders | 3 (37.5) | 5 (62.5) |
| Mixed dementia | 0 (0.0) | 9 (100.0) |
| Total | 10 (9.6) | 94 (90.4) |
- ∗Results shown as “absolute number (percentage).”
We conducted a post hoc analysis to identify which clinical characteristics were associated with cognitive improvement among patients with RPD. Improvement was associated with female gender (OR = 4.52; p = 0.038), an autoimmune disorder etiological diagnosis (OR = 15.00; p = 0.003), an infectious disorder etiological diagnosis (OR = 7.63; p = 0.028), hyperorality or dietary changes (OR = 6.28; p = 0.040), headache (OR = 26.70; p < 0.001), nausea and vomiting (OR = 6.28; p = 0.040), fever (OR = 11.50; p = 0.045), nystagmus (OR = 23.25; p = 0.024), and Broca aphasia (OR = 13.00; p = 0.011). RPD patients who improved were also younger at symptom onset (46.5 ± 12.99 years × 65.60 ± 14.71 years; p < 0.001). Alternatively, lower improvement odds were associated with a nonprion degenerative disease diagnosis (OR = 0.09; p = 0.029) and diffuse brain atrophy on CT (OR = 0.07; p = 0.013). Other RPD characteristics were not associated with odds of improvement (data not shown).
4. Discussion
This study found a 2.2% prevalence of RPD among the 4714 analyzed medical records and a prevalence of improvement of 9.6% among RPD patients. All cases of potentially reversible diseases had a duration of 1 year or less. The most prevalent RPD etiologic groups in this sample were dementia of vascular origin, nonprion degenerative diseases, and mixed dementia. The combination of VD and nonprion degenerative diseases represented most diagnoses of mixed dementia. Therefore, 60 (57.7%) patients had either dementia of vascular origin, nonprion degenerative disease, or a combination of both, disorders that usually display a slowly progressive course. Prion diseases, autoimmune disorders, and infectious disorders represented 7.7% of our sample each, while neoplasia, toxic–metabolic disorders, dementia associated with traumatic brain injury, and NPH were less common.
Prion diseases were considerably less frequent in our study than in studies from reference centers for prion diseases [3, 6–9]. Studies on RPD conducted in centers of reference for prion diseases found prion diseases to be the most frequent etiology of RPD, corresponding to 33.7%–68.2% of all RPD cases [3, 6–9]. Those studies, however, included mostly patients with suspected CJD or other prion diseases [3, 6–9], which may have biased patient selection toward a higher prevalence of prion diseases. The frequency of prion diseases identified in our sample was closer to that reported by other tertiary center studies, which conducted patient selection according to clinical criteria including a temporal cutoff point [10–14, 31]. Such studies reported frequencies of prion diseases among RPD patients ranging from 7.1% to 30.6% [10–14, 31]. Table 6 compares the relative frequencies of different RPD etiologic groups observed in tertiary center studies.
| Study | Screened patients (N) | RPD patients (N) | Frequency of different etiological groups of RPD (%) | |||||
|---|---|---|---|---|---|---|---|---|
| PD | NPDD | DVO | AID | ID | Other | |||
| Papageorgiou et al. [10] | 279 | 68 | 13.2 | 47.0 | 13.2 | 8.8 | 5.9 | 10.4 |
| Sala et al. [11] | 80 | 49 | 30.6 | 36.7 | 8.2 | 2.0 | 2.0 | 12.2 |
| Anuja et al. [12] | 693 | 187 | 7.5 | 14.4 | 9.6 | 26.7 | 20.8 | 18.2 |
| Zhang et al. [13] | ∗ | 310 | 7.1 | 24.8 | — | 9.7 | 25.2 | 15.4 |
| Studart Neto et al. [14] | 1,648 | 61 | 11.5 | 8.2 | 3.3 | 45.9 | 19.7 | 6.5 |
| Acosta et al. [31] | 144 | 104 | 29.8 | 26.9 | 2.9 | 23.1 | 2.9 | 13.5 |
| Our study | 4,714 | 104 | 7.7 | 23.0 | 26.0 | 7.7 | 7.7 | 23.1 |
- Abbreviations: AID, autoimmune disorders; DVO, dementia of vascular origin; ID, infectious disorders; NPDD, nonprion degenerative diseases; PD, prion diseases; RPD, rapidly progressive dementia; TMD, toxic–metabolic disorders.
- ∗Number of screened patients not informed.
Nonprion degenerative diseases were our study’s second most frequent RPD etiology. This is in accordance with the results of other tertiary center studies that found high frequencies of nonprion degenerative diseases among RPD patients [10, 11, 13, 31]. Nonprion degenerative diseases were also commonly identified in studies from reference centers for prion diseases, ranging from 13.7% to 29.0% of RPD cases [3, 6–9]. One study conducted in the memory clinic of a tertiary center found an even higher frequency of nonprion degenerative diseases (89.6% of RPD cases) [32]. Alternatively, the frequency of dementia of vascular origin found in our study was significantly larger than that reported by other studies on RPD [3, 6, 7, 9–12, 14, 31, 32]. Methodological or patient referral differences between diverse centers may have caused this divergence.
Potentially reversible RPD etiologies comprised a significant portion of our sample (29.8%). This finding was in accordance with what has been reported by other tertiary center studies [10–14, 31], albeit three groups found that potentially reversible disorders represented the majority of RPD cases [12–14]. This higher frequency of potentially reversible RPD etiologies may be explained by the inclusion of younger patients in those studies since reversible causes of RPD appear to be associated with a younger age of onset [6]. In our sample, nearly 10% of patients showed cognitive improvement, which is much lower than the favorable outcome frequency of 59% found in another Brazilian study [14]. That difference is probably due to the higher frequency of autoimmune and infectious disorders in that sample (7.7% vs. 45.9% and 7.7% vs. 19.7%, respectively) [14]. In both studies, autoimmune disorders, infectious disorders, and younger age were associated with better prognosis, while degenerative diseases had worse prognosis [14]. We found an association between female gender and RPD improvement. This association does not seem to be explained by differences in etiological diagnoses between genders and must be further investigated.
The subgroup of patients who had been submitted to a more thorough investigation with both MRI and CSF showed significantly higher frequencies of prion diseases, toxic–metabolic disorders, autoimmune disorders, and infectious disorders in comparison to our total sample, while the frequency of nonprion degenerative diseases and dementias of vascular origin was much lower. This difference must be interpreted with caution because our data were collected from a clinical setting, in which the choice of complementary tests for the investigation of dementia usually does not follow an established protocol but rather clinical suspicion, which probably makes some RPD patients more prone to be thoroughly investigated than others. Unfortunately, some complementary tests like MRI, EEG, and CSF protein 14-3-3 and tau determinations are not widely available in Brazilian public tertiary centers. In our sample, many patients were not submitted to those complementary tests, and therefore, the prevalence of prion diseases may have been underestimated in this study.
When the patients were stratified according to disease duration, it was found that all cases of prion diseases, toxic–metabolic disorders, NPH, autoimmune disorders, neoplasia, and infectious disorders had a duration inferior to or equal to 365 days. No patient with a disease duration superior to 365 days and inferior or equal to 730 days had a potentially reversible cause of RPD, and all of them had at least a nonprion degenerative disease or dementia of vascular origin, conditions that usually progress slowly [2–4]. In addition, this study found that prion diseases and autoimmune disorders had shorter disease duration in comparison to other causes of RPD, while nonprion degenerative diseases and mixed dementia had longer disease duration. These findings suggest that classifying cases of dementia between 1 and 2 years of duration as RPD may be futile, since the etiological diagnoses in that group of patients may not be significantly different from those of patients with dementia with a duration longer than 2 years. Likewise, one study found a high accuracy for diagnosing nonchronic neurodegenerative RPD when a 6-month disease duration cutoff point was used [31].
An additional finding of our study was the association of nonprion degenerative diseases and dementia of vascular origin with older age, while autoimmune and infectious disorders were associated with younger age. This is in accordance with what is reported in literature [6, 8, 10, 13, 32]. Likewise, the literature has reported shorter disease duration in patients with prion disease [10] and longer in patients with nonprion degenerative diseases [8], findings replicated in our study.
In addition to younger age of onset, potentially reversible causes of RPD show certain clinical and complementary test characteristics that may aid the diagnosis. Patients with metabolic disorders often have a history of substance abuse or toxic agent exposure [2, 3, 5] or malnutrition [3]. Certain clinical manifestations may suggest a specific diagnosis; for instance, the triad of dementia, diarrhea, and dermatitis is associated with pellagra (Vitamin B3 deficiency), while the combination of acute dementia, ophthalmoparesis, and ataxia suggests Wernicke–Korsakoff syndrome [3]. A thorough metabolic investigation including electrolytes, thyroid hormones, renal and liver function, vitamins, and toxicologic tests may help diagnose metabolic causes of RPD [2–5]. Thalamic hyperintensity on brain MR is seen in Wernicke–Korsakoff syndrome [3, 5], while pontine and extrapontine myelinolysis suggests osmotic demyelination syndrome [3]. Autoimmune disorders causing RPD frequently display memory impairment [3, 5], which may be accompanied by behavioral symptoms [3, 4] and seizures [4, 5]. Brain MR often displays hyperintensities on limbic structures [2, 5] and CSF shows inflammatory findings: pleocytosis, high protein count, and oligoclonal bands [2–5]. The presence of specific neural surface antibodies is helpful for the diagnosis of several autoimmune encephalopathies like anti-NMDAR and anti-LGI1 encephalitis [2–5]. RPD associated with neoplasia can be diagnosed with brain imaging [3–5] and CSF cytology [2, 4, 5]. Features suggestive of infectious RPD include flu-like symptoms [3], fever [3, 4], meningeal signs [3], accelerated course, and seizures [5]. CSF usually shows pleocytosis, high protein count, and oligoclonal bands [2–5]. Specific testing through blood and CSF cultures, PCR, and serologies confirms the diagnosis [2–5].
Some study design differences between our study and other research from other tertiary centers must not be overlooked. We employed a disease duration cutoff point of 2 years for the definition of RPD, which was also used in three other studies [13, 14, 31]. Alternatively, some groups have defined RPD with a cutoff point of 1 year of disease duration [10–12]. One study included an additional criterion based on dementia severity for the definition of RPD [14]. Such methodological differences may partially account for observed variations of RPD etiology frequencies between studies.
The main advantages of this study were its multicentric design, the employment of well-established clinical criteria for diagnosing different causes of RPD, and the broad time interval researched. Additionally, using ICD-10 codes in the search for medical records allowed the analysis of a large number of medical records and excused patient selection from hospital staff recollection, which could be prone to memory bias.
Nonetheless, our research has some limitations. The retrospective design based on medical record review precluded direct data acquisition by the researchers. Accordingly, we were not able to rule out the possibility that certain signs and symptoms may have been selectively investigated based on suspected diagnosis (for instance, urinary symptoms may have been better investigated in patients with ventriculomegaly on brain scans); therefore, the analysis of potential association between clinical findings and etiological group should be interpreted with caution. The multicentric design and the broad time interval researched resulted in a highly heterogeneous sample in terms of information available on medical records and clinical workup both inside and between centers. The selection of medical records based on ICD-10 codes may have missed RPD patients admitted with diseases not included in our list, such as pneumonia and urinary tract infection. Finally, since only inpatients were selected, there is a risk of selection bias toward a higher inclusion of diseases that more frequently demand inpatient care, like cerebrovascular disease, autoimmune disorders, and infectious disorders, and the potential underestimation of disorders such as nonprion degenerative diseases.
5. Conclusions
In conclusion, the relative frequencies of different causes of RPD in our sample were generally comparable to those observed in other tertiary center studies, validating those studies in a geographic region not studied before, with eventual differences possibly related to methodological differences between studies and demographical, patient referral, and disease prevalence variations between centers. Several patients with RPD show improvement with specific treatment, underscoring the importance of thorough investigation of RPD to correctly and promptly identify potentially reversible cases. RPD investigation should be guided by history, physical, and cognitive examination, and brain imaging, CSF analysis, and EEG are of great importance for differential diagnosis. Patients with dementia duration between 1 and 2 years seem to mostly have disorders usually associated with slow progression, and therefore, classifying such patients as having RPD may be of little use. Future studies would benefit from prospective designs that allow a better standardization of clinical workup, therefore producing richer clinical data and more accurate etiological diagnoses. In addition, including patients from an outpatient setting could make samples more representative of the general population with RPD.
Conflicts of Interest
The authors declare no conflicts of interest.
Funding
KL holds a National Council for Scientific and Technological Development–CNPq PQ2 Research Fellowship (Process Number 313205/2020-5).
Open Research
Data Availability Statement
The data supporting this study’s findings are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
