History

Dementia with Lewy bodies (DLB) is a relative newcomer to the field of late-life dementia, and its late emergence as an independent disease entity can in part be explained by its history. Friedrich Lewy initially noted the presence of spherical neuronal inclusions in autopsy cases of paralysis agitans, which were later coined Lewy bodies (LBs). Later in 1961, a case report by Okazaki and colleagues described two curious cases of elderly men who presented with dementia but who died shortly after with severe extrapyramidal rigidity, and who at autopsy had widespread LBs in the cerebral cortices. Over the next 20 years, further cases of dementia with parkinsonism were reported by Japanese researchers. Nevertheless, LB dementia was considered a rare condition, with the frequent co-occurrence of Alzheimer’s pathology (see later discussion) evident in the autopsy cases. Several competing labels were used, including diffuse LB disease, dementia associated with cortical LBs, the LB variant of Alzheimer’s disease (AD), senile dementia of LB type, and LB dementia.

Real interest in DLB, however, emerged in the late 1980s with advances in neuropathologic techniques including ubiquitin staining, which made the visualization of LBs easier. Seminal work done in a series of careful autopsy studies identified that LBs were actually a common feature of late-life dementia, affecting between 15% and 20% of elderly demented cases, and thus establishing DLB as the second most common form of degenerative dementia in old age. More recently, specific α-synuclein immunohistochemical staining has shown α-synuclein to be the core element of LBs, therefore placing DLB at the heart of the pathologic family of α-synucleinopathies that includes diseases such as Parkinson disease (PD), the associated Parkinson disease dementia (PDD), multisystem atrophy, pure autonomic failure, and rapid eye movement (REM) sleep behavior disorder. Further advances have shown that small aggregates of α-synuclein, not necessarily associated with LBs, are widespread in the brains of people with DLB, and it has been speculated that these small deposits may directly impair synaptic function, upset neuronal homeostatic energy processing, and lead ultimately to neuronal cell death. Nevertheless, linking how the pathologic changes lead to cognitive dysfunction and the other deleterious symptoms in DLB remains unclear.

Clinical characterization

Concordant with advances in neuropathology, research into DLB has led to the clearer clinical characterization of DLB, with the development of operationalized consensus criteria initially published in 1996 and revised in 2005 ( Box 1 ). The criteria clinically define 3 core symptoms for DLB: prominent (vivid) and recurrent visual hallucinations, spontaneous extrapyramidal symptoms, and fluctuating cognitive impairment. Other important clinical features are REM sleep behavior disorder and neuroleptic (antipsychotic) sensitivity. The presence of at least one core symptom plus one of the other supportive features is sufficient to make a diagnosis of probable DLB in the absence of obvious other causes such as cerebrovascular disease. Of note is the significance of neuroimaging in the revised criteria (discussed more in relevant sections in this article). Deficits in striatal dopamine transporter function are strongly suggestive of a diagnosis of DLB; that is, the presence of this plus only one of the core symptoms would be sufficient to make a diagnosis of probable DLB. Supportive neuroimaging features in making a diagnosis of DLB include relative preservation of medial temporal lobe structures on neuroimaging, generalized reductions in perfusion/metabolic imaging with reduced occipital activity, and abnormal cardiac imaging with iodine 123 meta -iodobenzylguanidine ([ ¹²³ I]MIBG).

Clinical characterization

Concordant with advances in neuropathology, research into DLB has led to the clearer clinical characterization of DLB, with the development of operationalized consensus criteria initially published in 1996 and revised in 2005 ( Box 1 ). The criteria clinically define 3 core symptoms for DLB: prominent (vivid) and recurrent visual hallucinations, spontaneous extrapyramidal symptoms, and fluctuating cognitive impairment. Other important clinical features are REM sleep behavior disorder and neuroleptic (antipsychotic) sensitivity. The presence of at least one core symptom plus one of the other supportive features is sufficient to make a diagnosis of probable DLB in the absence of obvious other causes such as cerebrovascular disease. Of note is the significance of neuroimaging in the revised criteria (discussed more in relevant sections in this article). Deficits in striatal dopamine transporter function are strongly suggestive of a diagnosis of DLB; that is, the presence of this plus only one of the core symptoms would be sufficient to make a diagnosis of probable DLB. Supportive neuroimaging features in making a diagnosis of DLB include relative preservation of medial temporal lobe structures on neuroimaging, generalized reductions in perfusion/metabolic imaging with reduced occipital activity, and abnormal cardiac imaging with iodine 123 meta -iodobenzylguanidine ([ ¹²³ I]MIBG).

Diagnosis of dementia with LBs versus PD with dementia

An area that has been subject to controversy is the separation of DLB from PDD. Consensus criteria for DLB and PDD recommend that for a diagnosis of PDD the extrapyramidal motor features need to be present for at least 12 months or more before the onset of the dementia, but if the dementia precedes the motor symptoms or has occurred within 12 months of the motor features, the diagnosis should be DLB. While subtleties exist between DLB and PDD, both conditions share cognitive, neuropsychiatric, neurochemical, pathologic, and imaging similarities, and this has led to the argument that both PDD and DLB should be considered the same disorder, or at the least two different points on an LB disease spectrum. The arbitrary one-year rule continues to allow distinction for research purposes and understanding the interrelationship between the two disease entities, but clinically, particularly where there is uncertainty concerning the onset of the motor symptoms relative to the cognitive symptoms, unitary approaches should be considered, and the catch-all term of LB dementias that includes both DLB and PDD can be useful in this regard.

Diagnosis of DLB versus AD

Recognition of DLB is important diagnostically because of the potentially dangerous adverse reaction to neuroleptics, the misuse of antiparkinsonian drugs, and the positive response to cholinesterase inhibitors. It has also been recognized that levels of neuropsychiatric symptomatology, functional disability, impaired quality of life, increased hospitalization, and annual care costs in DLB are significantly greater than those associated with AD, and thus early and accurate diagnosis is likely to be an essential component of effective treatment and management.

However, despite the development of clinical criteria, it is still apparent that DLB is underdiagnosed. Whereas 15% to 20% of dementia cases demonstrate neuropathologic features of DLB, epidemiologic surveys have tended to show rates of DLB to be much lower, at between 1% and 5%. This discrepancy has in part been driven by the lack of awareness of DLB as an independent clinical entity, difficulties in making the diagnosis clinically, and the interaction between coexistent AD and LB pathologies. Neuropathologically, up to 90% of DLB cases have high senile plaque counts, a level comparable with those found in pure AD, although only a minority of DLB cases show typical AD-associated neocortical neurofibrillary tangles, which are now thought to be integral to the pathophysiology of AD. Amyloid-β in the neocortex has also been associated with more extensive α-synuclein lesions, and thus it has been speculated that amyloid-β may have a role in the genesis of α-synuclein deposition in DLB.

Recent clinicopathologic data have suggested that the expression of a DLB clinical phenotype is related to the severity of LB pathology and inversely related to the severity of Alzheimer-type pathology ( Table 1 ). This finding has significant ramifications in making it difficult to diagnose DLB on clinical grounds alone in all cases.

Neuroimaging in DLB

One of the main areas for advancing research and improving the diagnosis of DLB has been the use of imaging, and the following sections explore the use of several imaging modalities that have been applied in DLB. These have focused mainly on imaging of brain structures and function although, given the increasing recognition that DLB is a multisystem disorder, a discussion is included on the use of cardiac imaging with [ ¹²³ I]MIBG to differentially diagnose DLB.

Structural magnetic resonance imaging

Volumetric changes in dementia, whether global or region specific, can be elucidated using structural magnetic resonance (MR). Although the largest body of literature exists for Alzheimer’s dementia, volumetric studies in DLB have helped in defining some of the neurobiological underpinnings of the disease, as well as making important contributions to the differential diagnosis of the condition and providing insights into its longitudinal progression. Analysis of structural data has included several methods ranging from simple visual inspection, through to region of interest (ROI) analysis and voxel-based morphometry (VBM). ROI and visual rating methods have been very much human led with operator-defined structures of interest, whereas VBM analysis has no a priori biases and makes statistical comparisons on a point-by-point basis (voxel level) across data sets.

Atrophic Changes in DLB

One of the main findings in AD is that of widespread global atrophy. In DLB the findings are less consistent; several studies have reported increased atrophy in DLB ( Fig. 1 ) in insular, frontal, inferior parietal, temporal, and occipital cortices compared with similarly aged controls, with relative sparing of the sensorimotor cortex. A large study by Whitwell and colleagues found very little cortical atrophy in patients with DLB aside from some subtle cortical loss in frontal and parietal lobes as well as volumetric reductions in hypothalamus, basal forebrain, and midbrain regions. The latter findings have also been reported in DLB by other groups, and may explain the marked cholinergic deficits apparent in DLB given that the nucleus basalis of Meynert is contained in the substantia innominata. However, in the studies by Whitwell and colleagues and Hanyu and colleagues, midbrain atrophy was also noted in AD patients, so these changes may not be specific to DLB.

Regions where atrophy is seen on MR imaging in DLB relative to healthy control subjects of similar age.

Putamen atrophy has been reported in DLB in comparison with AD by Cousins and colleagues, although no significant structural differences have been found in the caudate nucleus of DLB patients in contrast to PDD, where atrophy of the right caudate tail was found in one structural VBM study. Overall, striatal structural atrophic changes may reflect increased α-synuclein pathology but they probably are less indicative of functional dopaminergic loss in DLB, probably for 2 main reasons. First, Cousins and colleagues did not find a link between striatal atrophy and the severity of motor symptoms, and second, the magnitude of atrophy (5%–10%) is much less marked than the large (50%) reductions that occur in DLB in the uptake of radioligands, which bind to the striatal dopamine transporter (see later discussion).

Overall, therefore, global and subcortical atrophic changes have not proved to be particularly useful in the differential diagnosis of DLB. Perhaps the most consistent finding has been the lack of medial temporal lobe atrophy in DLB compared with AD on both visual inspection ( Fig. 2 ) and volumetric analysis, although the presence of medial temporal atrophy does not rule out a diagnosis of DLB. Instead medial temporal lobe atrophy, if present, tends to rule in AD rather than other dementias; this is supported by an autopsy study of 46 patients with dementia (AD, DLB, and vascular) by Burton and colleagues, who noted that antemortem MR of the medial temporal lobe was robustly associated with Alzheimer tangle pathology rather than plaque or LB pathology within the temporal lobe.

Coronal T1-weighted MR imaging. ( Left ) Patient with AD showing marked hippocampal atrophy. ( Right ) Patient with DLB with normal hippocampi.

Nevertheless, more recent reports using high-resolution structural MR imaging have suggested that subtleties exist between AD and DLB with regard to hippocampal subfield atrophy. Sabattoli and colleagues noted differential atrophy of more anterior regions in DLB compared with AD, consistent with CA2 and CA3 atrophy, and Firbank and colleagues reported reduced subiculum thickness, selective CA1 neuronal loss, and loss of distinction in a hypointense line that is visible between CA1 and CA3/4 in AD subjects in comparison with controls. This evidence is preliminary, but if replicated in larger samples high-resolution hippocampal imaging may become an additional tool in making a differential diagnosis between AD and DLB.