Sensitivity and specificity of three clinical criteria for dementia with Lewy bodies in an autopsy-verified sample
✍ Scribed by Cheryl A. Luis; Warren W. Barker; K. Gajaraj; Dylan Harwood; Randy Petersen; Alice Kashuba; Carol Waters; Pat Jimison; Gary Pearl; Carol Petito; Dennis Dickson; Ranjan Duara
- Publisher
- John Wiley and Sons
- Year
- 1999
- Tongue
- English
- Weight
- 97 KB
- Volume
- 14
- Category
- Article
- ISSN
- 0885-6230
No coin nor oath required. For personal study only.
✦ Synopsis
Objective. To evaluate the sensitivity and speci®city of the clinical features of three published diagnostic criteria for diuse Lewy body disease (DLBD) using autopsy-con®rmed Alzheimer's (AD), DLBD and AD DLBD (mixed) dementia cases.
Design. Retrospective chart review of an autopsy series of 56 patients selected from the State of Florida Brain Bank on the basis of a pathological diagnosis of either pure AD, DLBD ( pure and common forms) or AD DLBD (mixed) dementia. Clinical features were assessed by three raters blind to the pathological diagnosis.
Results. The existing criteria for a clinical diagnosis of DLBD were highly speci®c (90±100%) but not very sensitive (49±63%) in the dierential diagnosis of DLBD versus AD; sensitivity did improve (61±74%) when mixed AD DLBD cases were eliminated. Clinical features that occur more frequently in DLBD than in AD were unspeci®ed hallucinations, unspeci®ed EPS, ¯uctuating course and rapid progression. Post-hoc analysis also indicated that hallucinations and EPS were more common early in the disease course of DLBD than in AD. Empirically derived criteria, formulated using the most prevalent clinical features, demonstrated sensitivity values of 57±96% for pure forms and 43±91% for mixed forms.
Conclusions. This study demonstrated that additional improvements in the established criteria for DLBD are needed. Our empirically derived criteria enhanced the distinction of DLBD from AD while allowing the clinician the choice of maximizing sensitivity with acceptable speci®city, and vice versa.