Systematic review and meta-analysis show that dementia with Lewy bodies is a visual-perceptual and attentional-executive dementia

Abstract

To resolve differences in the literature, we have systematically reviewed 21 controlled comparisons of the cognitive performance of patients with dementia with Lewy bodies (DLB) These were identified by end May 2002 by Medline and PsycInfo searches, checking reference lists and contacting authors. Nine had comparisons between DLB patients (total n = 180) and age-matched controls (n = 172). Sixteen had comparisons between DLB (n = 312) and Alzheimer's disease (AD, n = 380). Three compared DLB (n = 48) with Parkinson's disease (PD, n = 65). Two raters independently scored the methodological quality. This was variable with a lack of high-quality studies (median rating 3 on a 0-7 scale, Kw = 0.41). There was a significant heterogeneity in results with marked discrepancies between studies. In a meta-analysis, DLB patients were more cognitively impaired than were AD or PD patients (95% CI of inverse variance weighted average of effect size relative to controls DLB 2.0-2.2; AD 1.4-1.6; PD 0.7-1.0). To permit an analysis of impairments in specific cognitive areas, the cognitive abilities underpinning the wide variety of tasks used were classified by a group of experienced neuropsychologists. Reducing overlapping task classifications using factor analysis showed large effect sizes relative to controls, AD and PD on two factors (combined variance 30%): attentional/executive impairment (effect sizes 1.1-2.9) and visual-perceptual impairment (0.7-3.6). There were small differences on two other factors (combined variance 39%): general verbal/non-verbal impairment (-0.12 to -0.5) and relative verbal memory impairment (-0.33 to 0.21). The cognitive performance is also more variable in DLB than in controls or in AD, but not PD (ratio of DLB to comparator standard deviations estimated from linear regression: DLB/controls 2.5-3.6; DLB/AD 2.1-2.6; DLB/PD 0.8-1.0). The greater variability of patients with DLB is seen only on tasks needing timed or motor responses, visual learning, executive or attentional abilities, or with visual content. Further stratification indicated that recent consensus diagnostic criteria, clinical diagnoses, and milder dementia were all associated with a more distinctive cognitive profile. The uniquely profound visual-perceptual and attentional-executive impairments that characterize DLB are consistent with the most frequent locations of Lewy bodies in frontal, cingulate, and inferior temporal cortex and may be related to the characteristic visual hallucinations and clinical fluctuations of this disease. These findings need to be confirmed in prospective, longitudinal, clinicopathological studies.