La maladie bipolaire (MB) s’accompagne d’un déficit cognitif léger, pouvant devenir démentiel chez le sujet âgé, rappelant la notion de vésanie. L’objectif de l’étude était de caractériser cette démence. Treize patients aux antécédents de MB avant 50 ans, répondant aux critères de démence ont consulté en centre mémoire et ont été suivis au minimum deux ans. Ils étaient en euthymie, leur âge moyen était de 70,8 ans (±7,7), le début de l’état démentiel survenait en moyenne 29,2 ans (±10,1) après le début de la MB, le score moyen au minimental state examination (MMSE) était de 24,0 (±4,3). Après un suivi moyen de 6,1 ans (±2,8), le MMSE moyen était de 23,5 (±3,2). La perte moyenne annuelle au MMSE était de 0,5 (±4,4). L’imagerie morphologique ne mettait pas en évidence d’atrophie focale ou de lésions vasculaires. L’imagerie fonctionnelle montrait dans tous les cas un hypodébit frontotemporal souvent associé à un hypodébit pariétal plus modéré. Le suivi neuropsychologique notait un syndrome dyséxécutif, un trouble mnésique fronto-sous-cortical et souvent un déficit visuospatial. Un syndrome frontal comportemental modéré était observé, tandis que les hallucinations étaient rares. Même après six ans, aucun patient ne répondait aux critères de probabilité des principales démences, faisant suspecter l’hypothèse d’une démence spécifique à la MB qu’une étude prospective devrait explorer.

Converging evidence suggests that people with bipolar disorder (BPD) exhibit persistent cognitive impairment independently from the emotional state. In old age BPD, the cognitive decline is more severe and can fulfill the criteria of dementia. However, the characteristics of bipolar disorder dementia are still unknown.

The aim of the study was to characterise the cognitive and imaging profile of the dementia following bipolar disorder.

Patients fulfilling criteria of dementia and followed-up in the memory unit for at least two years were included. Patients with substance abuse were excluded. A battery of specific (assessing verbal memory, attention, frontal executive function, construction and visuospatial impairment), and global (MMSE and Mattis dementia rating scale) neuropsychological tests, behavioural assessment using the frontotemporal behavioural scale, MRI and HMPAO–SPECT imaging were performed in all patients during euthymic state.

We included 13 patients with bipolar disorder (9W/4M). The mean age was 70.8 years (±7.7). Dementia began in average 29.2 years (±10.1) after the onset of the bipolar disorder. The mean score of MMSE was 24.0 (±4.3). The mean score of the Mattis dementia rating scale was 122.5 (±8.9). After an average of 6.1 years (±2.8) of follow-up, the mean score of MMSE was 23.5 (±3.2). The annual MMSE score decrease was of 0.5 (±4.4) per year. In more than 75% of the patients, Trail-Making Test-part B, Go-nogo test, Stroop test, delayed free recall in verbal explicit long-term memory test, category fluency tasks and code test were impaired. In more than 50% of patients, free recall, delayed cued recall, clock test, visuospatial battery and temporal orientation were impaired. On the other hand, spatial orientation and recognition were within the standards. The mean of the BREF score was 10.6 (±3.2). A moderate frontal behavioural syndrome was observed, but never persistent hallucinations. Seven patients had been treated with lithium and seven with antipsychotics, but none during the neurological assessment. Moderate extrapyramidal signs were reported in 10 patients, of which the seven patients treated in the past with antipsychotics. MRI showed no focal atrophy and no vascular lesions. Functional imaging conducted in 10 patients always showed uptake decrease in the frontotemporal regions and sometimes in the parietal region too. After six years of follow-up, no patient fulfilled the probable criteria for the main dementia, Alzheimer disease, vascular dementia, frontotemporal dementia and dementia with Lewy bodies.

The data of this study support a possible specific dementia postbipolar disorder and not only mild cognitive decline. This hypothesis could be tested in a prospective study. Such dementia could be a main differential diagnosis from long lasting frontotemporal dementia. The pathogenic process of this dementia could also be determined.