Association between visual hallucinations and cognitive performance in Lewy body dementia and Alzheimer's disease: A cross-sectional study

Abstract

Background: Visual hallucinations (VH) are an important neuropsychiatric feature of dementia. The association between VH and cognition remains controversial.

Objective: To investigate the differences in clinical correlates of VH and explore the associations between VH and cognitive functional decline in individuals with dementia with Lewy bodies (DLB) and Alzheimer's disease (AD).

Methods: Outpatient medical records of 154 patients with DLB and 297 patients with AD between January 2017 and December 2023 were reviewed. We collected demographic characteristics and used neuropsychological assessments and semi-structured detailed interviews to evaluate cognition and VH. Multiple linear regression and mediation analyses were employed to analyze the data, adjusting for confounding variables.

Results: DLB patients had a higher prevalence of VH than AD patients (p < 0.01). The presence of VH predicted lower Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA) scores in both DLB and AD patients (p < 0.01). In DLB patients, VH were associated with lower attention function scores after adjustment (p = 0.027). In AD patients, VH were related to worsened orientation ability after adjustment (p = 0.033). Attention function partially mediated the association between VH and cognition in DLB patients (p < 0.01), whereas orientation function partially mediated this association in AD patients (p < 0.01).

Conclusions: VH may independently correlate with deterioration in global cognitive performance. In DLB patients with VH, attentional function appears to be more impaired, whereas in AD patients, orientation function is the most affected. Different cognitive domains may help distinguish between DLB and AD patients with VH.

Keywords: Alzheimer's disease; Lewy bodies; cognitive impairment; dementia; visual hallucinations.

Figure 1.

Neuropsychological assessments scores in DLB and AD patients with and without VH. Neuropsychological assessment scores in AD and DLB patients. The results of non-memory cognition neuropsychological (ADL and CDR) assessment scores between those with VH and NVH in AD and DLB groups are described in (A). The results neuropsychiatric disturbances (NPI) assessment scores between those with VH and NVH in AD and DLB groups are described in (B). The results of global cognition (MMSE and MoCA) assessment scores between those with VH and NVH in AD and DLB groups are described in (C). The results of different cognitive domains assessment scores between those with VH and NVH in AD and DLB groups are described in (D). AD: Alzheimer's disease; ADL: Activity of Daily Living Scale; CDR: Clinical Dementia Rating; DLB: dementia with Lewy bodies; MMSE: Mini-Mental State Examination; MoCA: Montreal Cognitive Assessment; NPI: Neuropsychiatric Inventory; NVH: non-visual hallucination; VH: visual hallucinations. *p < 0.01, **p < 0.05, ***p < 0.001.

Figure 2.

Association between neuropsychological test scores and VH in patients with DLB or AD. The standardized coefficient and 95% confidence interval results of the association of cognitive scores between VH and DLB groups are described. AD: Alzheimer's disease; DLB: dementia with Lewy bodies; MMSE: Chinese Mini-Mental State Examination; MoCA: Montreal Cognitive Assessment; VH: visual hallucination. *p < 0.05, **p < 0.01, ***p < 0.001. Model 1: unadjusted model. Model 2: adjustment for sex, age, education, disease duration, stroke, DM, heart disease, hypertension, RBD, Parkinsonism and Cognitive fluctuation. Model 3 in DLB: adjustment for sex, age, education, disease duration, stroke, DM, heart disease, hypertension, RBD, Parkinsonism, Cognitive fluctuation, Attention, Language, Clock drawing, and Visuo-spatial ability of MMSE. Model 3 in AD: adjustment for sex, age, education, disease duration, stroke, DM, heart disease, hypertension, RBD, Parkinsonism, Cognitive fluctuation, Attention, Orientation Ability and Visuo-spatial ability of MMSE.

Figure 3.

Mediation factor of VH and global cognition functional in DLB and AD patients. Mediation analysis model adjusted for the common risk factors of dementia, including sex, age, education, disease duration, stroke, diabetes mellitus, heart disease, hypertension, rapid eye movement sleep behavior disorder, parkinsonism, and cognitive fluctuation. (a) DLB: Mediation models of VH, attention, and global cognitive function (MMSE/MoCA scores): Direct effect (−0.352, p = 0.654) / (−2.097, p < 0.001) of VH on MMSE/MoCA scores. Indirect effect (−0.442, p = 0.574) / (−2.321, p < 0.001) of VH on MMSE/MoCA scores through attention. (b) AD: Mediation models of VH, orientation (MoCA), and global cognitive function (MMSE and MoCA scores): Direct effect (−0.682, p = 0.420) / (−0.760, p = 0.184) of VH on MMSE/MoCA scores. Indirect effect (−1;704, p = 0.004) / (−1.819, p = 0.002) of VH on MMSE/MoCA scores through orientation. (c) AD: Mediation models of VH, orientation (MMSE), and global cognitive function (MMSE and MoCA scores): Direct effect (−0.456, p = 0.390) / (−1.001, p = 0.090) of VH on MMSE/MoCA scores. Indirect effect (−1.993, p < 0.001) / (−1.574, p = 0.002) of VH on MMSE/MoCA scores through orientation (MMSE). (d) DLB: Mediation models of cognitive fluctuation, VH, and global cognitive function (MMSE and MoCA scores): Direct effect (−1.249, p = 0.168) / (−0.790, p = 0.430) of VH on MMSE/MoCA scores. Indirect effect (−0.559, p = 0.024) / (−0.610, p = 0.030) of VH on MMSE/MoCA scores through cognitive fluctuation. (e) AD. Mediation models of cognitive fluctuation, VH, and global cognitive function (MMSE and MoCA scores): Direct effect (−1.205, p = 0.10) / (−2.767, p = 0.016) of VH on MMSE/MoCA scores. Indirect effect (−0.650, p = 0.020) / (−0.653; p = 0.020) of VH on MMSE/MoCA scores through cognitive fluctuation. AD: Alzheimer's disease; DLB: dementia with Lewy bodies; MMSE: Mini-Mental State Examination; MoCA: Montreal Cognitive Assessment; VH: visual hallucinations. *p < 0.01, **p < 0.05, ***p < 0.001.