"Untangling" the relationship between Alzheimer disease and dementia with Lewy bodies.

Neurology 2012;79:1938–1939 Lewy bodies, protein aggregations classically associated with Parkinson disease (PD) substantia nigra neurons, are not limited to the brainstem.1 Over time, pathologists developed terms such as “diffuse Lewy body disease” (DLBD) to describe autopsy cases in which Lewy bodies even extended to the neocortex.2 Efforts soon followed to define a clinical syndrome that predicted histopathology-defined entities such as DLBD.3 Eventually, formal international consensus criteria that emphasized dementia with varying combinations of parkinsonism, visual hallucinations, and cognitive/level of consciousness fluctuations were proposed.4 Clinicians embraced the new disease entity these criteria defined, dementia with Lewy bodies (DLB), applying it mostly to patients with concomitantly evolving dementia and parkinsonism. Not surprisingly, even though parkinsonism is not a requisite feature, neurologists have largely come to think of DLB as an Alzheimer disease (AD) and PD overlap syndrome. AD and DLB histologic features also overlap. Brains of patients who have been clinically diagnosed with DLB often have, in addition to brainstem and cortical Lewy bodies, neurofibrillary tangles and (especially) -amyloid plaques. Pervasive clinical and histologic overlap necessarily raises the question of whether DLB truly constitutes an independent disorder,5 and if not, does it more likely represent an AD or PD variant? In this issue of Neurology, Tsuang et al.6 present a study that informs this debate. The authors considered brain pathology data from 562 individuals who experienced dementia during life and 267 individuals who had not. Before death, subjects had been followed at AD centers, and no one carried a clinical PD diagnosis. Brains were categorized according to the presence or absence of substantial Lewy body and AD pathology. The 267 autopsied control brains showed at most limited Lewy body or AD pathology. Also included as controls were an additional 124 living subjects without dementia presumed to lack significant Lewy body or AD-associated protein aggregations. The GBA gene that codes for glucocerebrosidase, causing Gaucher disease when both copies are mutated,7 and that increases PD risk when one copy is mutated,8 was sequenced in 562 subjects with dementia and 391 control subjects. Within the dementia group, 14% were considered “pure” DLB (pDLB) cases that had Lewy body but not AD pathology, 41% had Lewy body disease (LBD) extending beyond the brainstem and AD pathology (LBD-AD group), and 45% had AD but not Lewy body pathology. Twenty-five GBA mutations were found within the 953 sequenced subjects. These mutations were not distributed equally among the groups. The pDLB group had the highest GBA mutation frequency, both the AD and control groups had low GBA mutation frequencies, and the LBD AD group had an intermediate GBA mutation frequency. These data reveal that in addition to increasing PD risk, heterozygous GBA mutations also increase DLB risk. The authors reasonably conclude, therefore, that at a molecular-pathology level, DLB is more closely related to PD than it is to AD. This conclusion is further supported by the fact that sex distributions differed between the groups. As is the case with most PD but not AD cohorts, the pDLB group was predominantly male.9 Also, because study subjects derived from AD centers, and PD-diagnosed patients were not included, the pDLB group was probably not contaminated by PD that had simply progressed to a PD dementia syndrome. Study limitations include the fact that subjects considered to have DLB in this study were not diagnosed with DLB during life. The authors state this is because most subjects died before current DLB consensus criteria were published.10 Thus, whereas subjects with pDLB and LBD AD by definition certainly had LBD and usually DLBD, it is unclear whether individuals in these groups had clinically di-