|Home | About | Journals | Submit | Contact Us | Français|
There are little data on the relationship between Lewy body disease and mild cognitive impairment syndromes. The Mayo Clinic aging and dementia databases in Rochester, Minnesota, and Jacksonville, Florida were queried for cases who were diagnosed with mild cognitive impairment between 1 January 1996 and 30 April 2008, were prospectively followed and were subsequently found to have autopsy-proven Lewy body disease. The presence of rapid eye movement sleep behaviour disorder was specifically assessed. Mild cognitive impairment subtypes were determined by clinical impression and neuropsychological profiles, based on prospective operational criteria. The diagnosis of clinically probable dementia with Lewy bodies was based on the 2005 McKeith criteria. Hippocampal volumes, rate of hippocampal atrophy, and proton magnetic resonance spectroscopy were assessed on available magnetic resonance imaging and spectroscopy scans. Eight subjects were identified; six were male. Seven developed dementia with Lewy bodies prior to death; one died characterized as mild cognitive impairment. The number of cases and median age of onset (range) for specific features were: seven with rapid eye movement sleep behaviour disorder—60 years (27–91 years), eight with cognitive symptoms—69 years (62–89 years), eight with mild cognitive impairment—70.5 years (66–91 years), eight with parkinsonism symptoms—71 years (66–92 years), six with visual hallucinations—72 years (64–90 years), seven with dementia—75 years (67–92 years), six with fluctuations in cognition and/or arousal—76 years (68–92 years) and eight dead—76 years (71–94 years). Rapid eye movement sleep behaviour disorder preceded cognitive symptom onset in six cases by a median of 10 years (2–47 years) and mild cognitive impairment diagnosis by a median of 12 years (3–48 years). The mild cognitive impairment subtypes represented include: two with single domain non-amnestic mild cognitive impairment, three with multi-domain non-amnestic mild cognitive impairment, and three with multi-domain amnestic mild cognitive impairment. The cognitive domains most frequently affected were attention and executive functioning, and visuospatial functioning. Hippocampal volumes and the rate of hippocampal atrophy were, on average, within the normal range in the three cases who underwent magnetic resonance imaging, and the choline/creatine ratio was elevated in the two cases who underwent proton magnetic resonance spectroscopy when they were diagnosed as mild cognitive impairment. On autopsy, six had neocortical-predominant Lewy body disease and two had limbic-predominant Lewy body disease; only one had coexisting high-likelihood Alzheimer's disease. These findings indicate that among Lewy body disease cases that pass through a mild cognitive impairment stage, any cognitive pattern or mild cognitive subtype is possible, with the attention/executive and visuospatial domains most frequently impaired. Hippocampal volume and proton magnetic resonance spectroscopy data were consistent with recent data in dementia with Lewy bodies. All cases with rapid eye movement sleep behaviour disorder and mild cognitive impairment were eventually shown to have autopsy-proven Lewy body disease, indicating that rapid eye movement sleep behaviour disorder plus mild cognitive impairment probably reflects brainstem and cerebral Lewy body disease.
Mild cognitive impairment refers to the intermediate state between normal ageing and dementia that was initially conceptualized as a prodrome of Alzheimer's disease. Subsequently, it has been recognized that there are both amnestic and non-amnestic forms of mild cognitive impairment. Patients with non-amnestic mild cognitive impairment have deficits in the domains of language, attention/executive functioning or visuospatial functioning. If mild cognitive impairment is an intermediate state between normal ageing and dementia, those with dementia due to other neurodegenerative aetiologies are also likely to pass through an mild cognitive impairment state.
Dementia with Lewy bodies is a syndrome characterized by dementia, plus at least two of the following features: (i) recurrent, fully formed visual hallucinations; (ii) spontaneous parkinsonism; and (iii) fluctuations in cognition and/or arousal (McKeith et al., 2005). Neuropsychological data have shown that those with dementia with Lewy bodies typically have impaired attention/executive functioning and visuospatial skills (Salmon et al., 1996; Ferman et al., 1999, 2002, 2006; Mori et al., 2000). Language skills such as confrontation naming are often preserved (Ferman et al., 2006), while performance on measures of learning and memory is more variable. We hypothesized that patients with a non-amnestic form of mild cognitive impairment, with impairment in attention/executive functioning and/or visuospatial skills would evolve to have other clinical features of dementia of Lewy bodies and ultimately Lewy body disease on autopsy.
Rapid eye movement (REM) sleep behaviour disorder is a parasomnia that is characterized by loss of normal skeletal muscle atonia during REM sleep, with prominent motor activity and dreaming (Schenck et al., 1986; Olson et al., 2000). REM sleep behaviour disorder has been associated with synucleinopathies such as dementia with Lewy bodies, Parkinson's disease, multiple system atrophy and pure autonomic failure, but is far less commonly associated with non-synucleinopathy disorders (Boeve et al., 2001, 2003, 2007b; Iranzo et al., 2006; Postuma et al., 2009). REM sleep behaviour disorder often precedes the onset of cognitive impairment and/or parkinsonism by years or even decades (Schenck and Mahowald, 2002; Iranzo et al., 2006; Boeve et al., 2007b; Postuma et al., 2009). We further hypothesized that patients with mild cognitive impairment and REM sleep behaviour disorder, regardless of the mild cognitive impairment subtype, would represent prodromal Lewy body disease.
We characterized the clinical features, neuropsychological profiles and structural neuroimaging patterns of those who were diagnosed with mild cognitive impairment (any subtype), were prospectively followed, and were subsequently found to have autopsy-proven limbic or neocortical Lewy body disease.
Subjects were identified through the combined databases of the Alzheimer's Disease Research Center at Mayo Clinic Rochester and Mayo Clinic Jacksonville and the Alzheimer's Disease Patient Registry at Mayo Clinic Rochester. Both of these programmes are approved by the Mayo Foundation Institutional Review Board. Written consent for participation was provided by the subjects or their proxies. The databases were queried to identify patients who were diagnosed with mild cognitive impairment between 1 January 1996 and 30 April 2008, were prospectively followed, and were subsequently found to have autopsy-proven limbic- or neocortical-predominant Lewy body disease. The clinical, neuropsychological, neuroimaging and neuropathological features were analysed.
Patients were initially evaluated using a standardized clinical protocol and followed prospectively. The same research protocol was completed at 12–15 month intervals, with additional evaluations in between research assessments depending on active clinical issues. A behavioural neurologist (B.B., D.K., R.P., N.G.R. or the late Emre Kokmen) evaluated each patient by obtaining a medical history from the patient and their corroborating sources and by performing a complete neurological examination (Members of the Department of Neurology, 1998). Features of parkinsonism (e.g. tremor, rigidity, bradykinesia, postural instability, shuffling gait, masked facies, etc.) were noted when present, and Parkinson's disease was diagnosed if a patient fulfilled United Kingdom Brain Bank criteria for the disorder (Hughes et al., 1992). The Clinical Dementia Rating scale (Morris, 1993), Folstein Mini-Mental State Exam (Folstein et al., 1975), and Kokmen Short Test of Mental Status (Kokmen et al., 1991; Tang-Wai et al., 2003), were completed on all cases. The Mayo Fluctuations Scale—a validated operationalized measure for determining the presence of fluctuations in cognition and/or arousal (Ferman et al., 2004)—was completed by the informants. This scale has been used in the standard research protocol at our institution since it was developed in 1999. Affirmative answers on at least three out of the four questions were considered necessary for the presence of fluctuations. All patients underwent a neuropsychological evaluation assessing for memory, language, attention/executive functioning and visuospatial functioning. Structural neuroimaging of the brain with magnetic resonance imaging was performed in a subset of cases that were having data acquired for volumetric magnetic resonance analyses and magnetic resonance spectroscopic analyses.
Testing included assessment of global cognitive functioning [Mattis Dementia Rating Scale (DRS) (Mattis, 1988)]; learning and memory [percent retention on immediate and delayed recall on the Logical Memory subtest of the Weschler Memory Scale-Revised (Wechsler, 1987), learning over trials and percent retention on the Rey-Auditory Verbal Learning Test (Rey, 1964)]; language functioning [Boston Naming Test (Kaplan et al., 1978), Controlled Oral Word Association Test (Benton and Hamsher, 1978), and category/semantic fluency (animals, fruit, vegetables)]; attention/executive functioning [Trail Making Test parts A and B (Reitan, 1958), and Digit Span of the Wechsler Adult Intelligence Scale-Revised (Wechsler, 1981)]; and visuospatial/perceptual functioning [Block Design and Picture Completion subtests of the Wechsler Adult Intelligence Scale-Revised, and copy of Rey-Osterreith Complex Figure (Rey, 1941; Osterrieth, 1944)]. Additional tests were also administered and used by the neuropsychologists in their assessments of individual patients. Mayo Older American Normative Studies norms were used to determine scaled scores for these tests, in which 10 represents the mean and the standard deviation is 3 (Ivnik et al., 1992, 1996, 1997; Lucas et al., 1998a, b). Therefore, a Mayo Older American Normative Studies scaled score of 7 is 1 SD below the mean and 4 is 2 SDs below the mean.
Three of the cases (Cases 3, 7 and 8) had MRI and two had 1H-magnetic resonance spectroscopy (Cases 3 and 7) examinations performed at the time they were diagnosed with mild cognitive impairment. MRI was performed using a General Electric scanner at 1.5 Tesla, and images of the brain were obtained in the sagittal (T1-weighted), axial (proton-density, T2-weighted and fluid attenuation inversion recovery), and coronal (T1-weighted) planes. Hippocampal volume measurements were derived from a T1-weighted 3D volumetric spoiled gradient-recalled echo sequence in the coronal plane by manually tracing their anatomic boundaries for each image slice sequentially from posterior to anterior. Volumes were adjusted for age, gender and head size; normal percentiles are referred to as W scores, using age- and gender-specific normal percentiles based on a previous study (Jack et al., 2000). A value of zero corresponds to the 50th percentile, +1.64 corresponds to the 95th percentile, –1.64 corresponds to the 5th percentile among normal subjects. The rate of change in hippocampal volumes was calculated as the annualized percent change in hippocampal volume, which was computed as the volume in cubic millimeters of scan 2 minus that of scan 1 divided by volume on scan 1, divided by the duration (in years) between the two scans (×100).
T1-weighted images in the sagittal plane were used for localizing the 1H-magnetic resonance spectroscopy voxel. Point resolved spectroscopy pulse sequence with repetition time/echo time = 2000/30 ms was used for the examinations. An 8 cm3 (2 cm × 2 cm × 2cm) voxel, prescribed on a mid-sagittal T1-weighted image, included right and left posterior cingulate gyri and inferior precunei (Kantarci et al., 2000). We quantified metabolite intensities by referencing to an internal standard, the creatine peak, to correct for coil loading, relaxation times and inter subject differences in atrophy (i.e. partial volume averaging of the magnetic resonance spectroscopy voxel).
Weekly consensus meetings were held to review each patient's diagnosis. All information, including neurological, neuropsychological, laboratory and imaging sources, were used. Subjects were diagnosed as having normal cognition, mild cognitive impairment using published criteria (Petersen, 2004), or dementia according to the Diagnostic and Statistical Manual of Mental Disorders, revised third edition (1987). The specific dementia diagnosis for Alzheimer's disease, dementia with Lewy bodies, frontotemporal dementia and other dementia syndromes were made based on established criteria (McKhann et al., 1984; Neary et al., 1998; McKeith et al., 2005).
For the diagnosis of mild cognitive impairment, subjects are required to meet the following criteria: (i) a cognitive complaint, preferably corroborated by an informant; (ii) essentially normal activities of daily living; (iii) normal general cognitive functioning; (iv) abnormal performance in one or more cognitive domains; and (v) not demented. Impairment in cognitive domains was determined by neuropsychological test scores; scores one SD below the mean (i.e. a Mayo Older Adult Normative Scale Standard Score ≤7) were considered borderline to clearly abnormal, but clinical impression based on these and other tests scores and all other available information were also considered, as is routine in clinical practice. Some tests can be abnormal and reflect abnormalities in one and/or another domain (e.g. impaired performance on the Controlled Oral Word Association Test can be due to language and/or attention/executive dysfunction), again underscoring the need for the clinical impression to be based on all data. Based on the domains which were impaired, subjects were further classified as having amnestic mild cognitive impairment, multiple-domain amnestic mild cognitive impairment, single-domain non-amnestic mild cognitive impairment and multiple-domain non-amnestic mild cognitive impairment.
All cases underwent a standardized neuropathologic assessment, with evaluation of gross and microscopic findings and analysis of Alzheimer-type pathology, Lewy body pathology, cerebrovascular pathology and concomitant pathology according to established and published guidelines (Braak and Braak, 1997; Fujishiro et al., 2008). Sections were taken from six regions of the cortex, hippocampus, amygdala, basal ganglia, thalamus, midbrain, pons, medulla and cerebellum. Counts of senile plaques and neurofibrillary tangles were made in six cortical sections, four sectors of the hippocampus, two regions of the amygdala, and the basal nucleus of Meynert with thioflavin-S fluorescent and/or Bielschowsky microscopy. The presence of amyloid angiopathy was assessed. Senile plaques and neurofibrillary tangles were counted at ×100 and ×400, respectively, in cortex, hippocampus and amygdale. A Braak neurofibrillary tangle stage (Braak and Braak, 1991) was assigned to all cases based on the distribution of neurofibrillary tangles with thioflavin-S fluorescent microscopy, as previously described (Togo et al., 2002; Josephs et al., 2004). The severity of senile plaque pathology was also assessed using the Consortium to Establish a Registry for Alzheimer Disease guidelines (Mirra et al., 1991). All cases underwent immunostaining with a monoclonal antibody to phospho-tau (CP13 ; Peter Davies, Albert Einstein College of Medicine, Bronx, NY) and a polyclonal antibody to α-synuclein (Gwinn-Hardy et al., 2000) using immunostaining with a Dako Autostainer. The subtypes of Lewy body pathology (i.e. brainstem, limbic/transitional, or neocortical/diffuse) were determined based on Lewy body counts in five cortical sections and the amygdala. Semiquantitative grading of Lewy body pathology and assignment of Lewy body type were also determined according to recommendations from the Third Consortium of Dementia with Lewy Bodies (McKeith et al., 2005). The two methods gave similar results. The Lewy body density was determined at ×200 magnification from the following regions: middle frontal (Brodmann area 46), superior temporal (Brodmann area 38), inferior parietal (Brodmann area 39), anterior cingulate (Brodmann area 24) and parahippocampal gyri (Brodmann area 35). The recently published and validated Dementia with Lewy Bodies Consensus criteria for the diagnosis of Lewy body disease was used to characterize cases (Fujishiro et al., 2008).
Cerebrovascular pathology was assessed in all cases using a semiquantitative scale similar to that previously reported (Jellinger and Attems, 2003). Briefly, cases with no cerebrovascular lesions were scored 0, those with minimal cerebrovascular pathology (including one to two small lacunes, mild cerebral amyloid angiopathy or mild leukoencephalopathy) were scored 1, those with moderate lesions (including more than 2 lacunes, severe cerebral amyloid angiopathy, or diffuse leukoencephalopathy) were scored as 2, and those with marked cerebrovascular pathology (including old cortical infarcts, multiple microinfarcts or hippocampal sclerosis) were scored 3.
Concomitant pathologies were noted when present. Argyrophilic grains, neuronal threads, oligodendroglial coiled bodies, astrocytic plaques and globose neurofibrillary tangles were assessed using monoclonal antibody to phospho-tau (CP13). Glial intracytoplasmic inclusions were assessed by α-synuclein immunohistochemistry.
The final neuropathologic diagnosis was made according to the Dementia with Lewy Body Consensus criteria for the diagnosis of Lewy body disease (Fujishiro et al., 2008; McKeith et al., 2005) and the National Institute on Aging-Reagan criteria for Alzheimer's disease (Consensus recommendations for the post mortem diagnosis of Alzheimer's disease, 1997).
Eight patients were identified. All had an education level of 10 years or greater, and six of the patients were male. The demographic and clinical features are indicated in Table 1, and the neuropathologic findings are shown in Table 2 . Narrative descriptions of each patient's clinical course are in Supplementary Text E1, and the longitudinal clinical data and neuropsychological profile of impairment at the time of mild cognitive impairment diagnosis for each patient are shown in the Figures 1–8.
Quantitative neuroimaging data were available in a subset of cases. In the three cases who underwent serial magnetic resonance imaging exams, baseline hippocampal volumes at the time of mild cognitive impairment diagnosis and the rate of hippocampal atrophy at follow-up were on average within the range of the cognitively normal subjects, when compared to previously published data (Jack et al., 2000). In two of the cases who underwent baseline 1H-magnetic resonance spectroscopy exams, the choline/creatine ratio and myoinositol/creatine ratio were elevated in both cases, and the n-acetyl aspartate/creatine ratio was decreased in one of the cases compared to previously published data (Kantarci et al., 2000).
Seven patients’ condition declined and they experienced other symptoms and received a diagnosis of clinically probable dementia with Lewy bodies prior to death; one patient died while classified as mild cognitive impairment. All patients had a Clinical Dementia Rating score of 0.5 at the time of the mild cognitive impairment diagnosis. The median age of cognitive symptoms onset was 69 years (range 61–89 years), mild cognitive impairment diagnosis was 70.5 years (range 66–91 years), and dementia onset was 78 (range 67–90 years), with a median Clinical Dementia Rating score of 2 (range 1–3). Mini-Mental State Examination data were available for all patients when initially diagnosed with mild cognitive impairment; the median Mini-Mental State Examination score was 28 (range 23–29).The median number of years between the onset of cognitive symptoms and dementia was 4 years (range 2–6 years) and between the onset of mild cognitive impairment and dementia was 2.5 years (range 1–5 years). The median number of years from a diagnosis of mild cognitive impairment onset to death was 6.5 years (range 5–9 years). Among the seven patients who developed dementia, the median number of years from dementia onset to death was 3 years (range 0–4 years).
In this cohort, both amnestic mild cognitive impairment and non-amnestic mild cognitive impairment subtypes were represented, including three with multi-domain amnestic mild cognitive impairment, three with multi-domain non-amnestic mild cognitive impairment and two with single-domain non-amnestic mild cognitive impairment. Attention/executive functioning (n = 6) and visuospatial functioning (n = 6) were the cognitive domains most frequently affected.
Seven patients had REM sleep behaviour disorder with a median onset age of 60 years (range 27–91 years). Polysomnogram was performed in four cases, in whom REM sleep behaviour disorder was confirmed in all. The onset of REM sleep behaviour disorder features preceded onset of cognitive symptoms in six patients by a median of 10 years (range 2–47 years) and mild cognitive impairment diagnosis by of 12 years (range 3–48 years). One patient developed REM sleep behaviour disorder 2 years after the onset of cognitive symptoms and at the same time as the diagnosis of mild cognitive impairment.
Six patients had visual hallucinations, with a median onset age of 72 years (range 64–90 years). Among these patients, five developed hallucinations at a median of 3 years after the onset of cognitive symptoms (range 1–4 years); one patient developed hallucinations 3 years prior to the onset of cognitive impairment.
Seven patients had parkinsonism, with a median onset age of 71 years (range 71–92 years). In five patients, parkinsonism occurred at a median of 3 years after onset of cognitive symptoms (range 2–5 years). Two patients developed parkinsonism at the same time as cognitive symptoms onset, and one patient developed parkinsonism 1 year prior to onset of cognitive symptoms. None of the patients fulfilled criteria for the diagnosis of Parkinson's disease prior to the onset of dementia.
Fluctuations were classified as present in six patients based on Mayo Fluctuations Scale data, with a median onset age of 76 years (range 68–92 years). Fluctuations evolved after onset of cognitive symptoms in all six cases, at a median of 4.5 years (range 3–6 years). Mayo Fluctuations Scale data were not available in one patient who died prior to the development of the scale; however, the clinical notes did not suggest the presence of fluctuations.
All three core features of dementia with Lewy bodies eventually developed in five patients. Both patients with two core features had spontaneous parkinsonism. The patient who died with a classification of mild cognitive impairment had parkinsonism and REM sleep behaviour disorder, but no hallucinations or fluctuations. Visual hallucinations and fluctuations tended to occur after the onset of cognitive impairment at a median of 3 and 4.5 years, respectively. Parkinsonism tended to manifest after the onset of cognitive symptoms, though also did occur before and at the same time as the cognitive symptoms onset in a few patients. Seven patients had clinical evidence of REM sleep behaviour disorder, with six patients having dream enactment behaviour preceding the onset of mild cognitive impairment.
The neuropathological data is shown in Table 2. Six patients had a pathological diagnosis of neocortical Lewy body disease, and two patients had limbic Lewy body disease. Concomitant pathology was present in five patients; one met high likelihood National Institute of Aging-Reagan criteria for Alzheimer's disease, another had agyrophilic grain disease, another had mild cerebrovascular disease (none had cerebrovascular pathology beyond lacunes, mild amyloid angiopathy, or a small focal infarct in the caudate head), and three had pathological ageing with diffuse plaques but sparse or no neuritic plaques.
Our principal findings were that (i) Lewy body disease can evolve through a mild cognitive impairment intermediate state; (ii) amnestic or non-amnestic mild cognitive impairment subtypes can evolve into dementia with Lewy bodies; and (iii) mild cognitive impairment subtypes with impairment in attention-executive function and/or visuospatial functioning may be particularly suggestive of underlying Lewy body disease.
Results from previous research have shown that those who died with an amnestic mild cognitive impairment subtype (n = 15) may have some degree of medial temporal lobe changes in essentially every case (e.g. neurofibrillary tangles, argyrophilic grain disease, hippocampal sclerosis, etc), with only a few meeting criteria for fully expressed Alzheimer's disease (Petersen et al., 2006). While data in those who evolved from amnestic mild cognitive impairment to dementia did show that most met criteria for Alzheimer's disease (Jicha et al., 2006), a minority of patients had other pathological diagnoses (note the single case with Lewy body disease pathology in that series was not included in the current series of eight cases as she developed dementia prior to 1996). Three of the patients in our series had an amnestic mild cognitive impairment subtype and Lewy body disease on autopsy. Though it has been suggested that all cases of amnestic mild cognitive impairment reflect evolving Alzheimer's disease (Morris et al., 2001; Dubois and Albert, 2004; Sarazin et al., 2007), our findings suggest that the neuropathological substrate for amnestic mild cognitive impairment is more heterogeneous.
The cognitive domains most frequently affected in the mild cognitive impairment of dementia with Lewy bodies were attention/executive functioning and visuospatial functioning. Moreover, seven patients had a history of dream enactment behaviour during sleep, even if it was no longer an active problem. Our experience suggests that patients with a history of dream enactment behaviour during sleep who present with attention/executive and/or visuospatial difficulty despite preserved complex activities of daily living may represent underlying Lewy body disease. Indeed, seven of the eight patients were followed longitudinally and eventually developed a dementia and corresponding core clinical features of dementia with Lewy bodies.
Several of the subjects in this series were relatively easy to characterize, with clear cognitive complaints that were corroborated by their informants and neuropsychological features that permitted a straight-forward determination of which domains were impaired.
However, the neuropsychological features of mild cognitive impairment associated with Lewy body disease are variable. Determining which domains are impaired and classifying them into a specific mild cognitive impairment subtype can be challenging in some patients. Case 3 exemplifies this challenge, as he repeatedly voiced concerns over his cognitive symptoms and used notes obsessively, yet his performance on neuropsychological testing over the initial 4 years of his symptoms was normal or only mild impaired. Even on testing just prior to the diagnosis of dementia with Lewy bodies, performance was in the average range across most tests. To what degree this represents cognitive fluctuations, or the ability of some patients to ‘rise to the occasion’ and perform better than expected on formal testing despite florid symptoms and modest functional changes, is unclear. This case also underscores that not all patients who voice strong cognitive concerns yet perform normally on neuropsychological testing are ‘worried well’; indeed, this patient had autopsy-proven Lewy body disease.
We again emphasize the need to incorporate all clinical and neuropsychological data when making determinations on mild cognitive impairment diagnoses; following a strict algorithmic approach to interpreting neuropsychological data does not capture the complexity of some cases. On the other hand, it may be argued that the more impaired multi-domain mild cognitive impairment cases may have been much closer to a diagnosis of dementia than those with more restricted patterns of cognitive difficulty. Nonetheless, adamant assurances from family members that their activities of daily living continued to be unchanged at the time of the evaluation warranted that a diagnosis of mild cognitive impairment be made, rather than dementia. In either case, the family's observations are critically important when making a determination of functional status. As noted in the ‘Methods’ section, many of the tests included in the study cannot be completely compartmentalized to a single cognitive domain, so poor performance on some tests can be interpreted as reflecting impairment in more than one domain. Also, we purposefully gave more weight to delayed recall than immediate recall of paragraph material, given data that shows delayed recall is a useful discriminator of Alzheimer's disease from dementia with Lewy bodies (Ferman et al., 2006), realizing that results on immediate recall are also similarly different between mild cognitive impairment associated with Alzheimer's disease pathology compared with that associated with Lewy body disease pathology (Jicha et al., 2008).
Our observations should be considered preliminary since we had structural MRI on only three subjects and 1H-magnetic resonance spectroscopy on only two. Hippocampal volumes and atrophy rates were within the normal range of values in all three of the mild cognitive impairment cases. This result is consistent with results from studies of patients who have dementia with Lewy bodies, in whom hippocampal volumes are notably larger than patients with Alzheimer's disease (Whitwell et al., 2007). 1H-magnetic resonance spectroscopy analysis showed significantly elevated choline/creatine ratios in the two mild cognitive impairment cases we studied. Dementia with Lewy body patients are characterized by significantly elevated choline/creatine ratios, which tend to be higher than the choline/creatine elevation in patients with Alzheimer's disease (Kantarci et al., 2004). Although the number of subjects with 1H-magnetic resonance imaging was very small, our results in mild cognitive impairment are consistent with those of dementia with Lewy bodies.
This series is small, but a relatively consistent evolution of clinical features occurred. All but one had REM sleep behaviour disorder preceding the onset of mild cognitive impairment, and most developed parkinsonism concurrently or after the onset of their cognitive symptoms. None of these patients fulfilled criteria for Parkinson's disease prior to the onset of cognitive symptoms, nor during their phase of mild cognitive impairment; and hence none could be considered as representing Parkinson's disease with dementia. Visual hallucinations tended to evolve after both cognitive impairment and parkinsonism, and fluctuations, at least as measured by the Mayo Fluctuations Scale (Ferman et al., 2004), tended to be the final core feature to evolve. Dementia onset along with two or more of the other core features led to the diagnosis of clinically probable dementia with Lewy bodies 2–6 years after a diagnosis of mild cognitive impairment was made. These data are consistent with prior work that has shown that dementia plus REM sleep behaviour disorder in the absence of other core features of dementia with Lewy bodies probably reflects underlying Lewy body disease (Ferman et al., 2002).
There were exceptions to the evolution of features noted above. One patient initially experienced recurrent fully formed visual hallucinations and was thought to have Charles–Bonnet syndrome due to the absence of any other neurological signs or symptoms. The hallucinations spontaneously remitted but were followed 3 years later by cognitive decline, with the hallucinations returning years later. Another patient complained vehemently about his memory problems, urinary incontinence and erectile dysfunction, though neuropsychological performance early in his course was minimally abnormal. And one patient underwent a polysomnogram to evaluate obstructive sleep apnoea as a possible contributor to her cognitive symptoms. While REM sleep without atonia—the electrophysiologic substrate for REM sleep behaviour disorder—was evident on her polysomnogram, she did not begin exhibiting recurrent dream enactment behaviour until 2 years later.
In addition, it may be challenging to elicit a history of fluctuations in a clinical setting. For this study, the Mayo Fluctuations Scale was used to operationalize a more consistent determination of fluctuations. Asking whether the patient fluctuates does not distinguish dementia with Lewy bodies from Alzheimer's disease (Ferman et al., 2004). In order to make more definitive generalizations about the early characteristics and evolution of dementia with Lewy bodies/Lewy body disease, more subjects need to be assessed in a comprehensive manner and followed longitudinally using scales that adequately capture the concept of fluctuations in cognition and/or arousal.
A clinical feature or sign in neurodegenerative disease reflects sufficient neuronal/glial/neurotransmitter dysfunction in a critical neuronal network. While some of these features or signs have known or suspected networks of dysfunction, the underlying substrate for other features or signs is less clear. Parkinsonism is most likely to be associated with dopaminergic depletion due to sufficient degeneration in the nigrostriatal system, and cognitive impairment is likely to be associated, at least in part, with cholinergic depletion due to basal forebrain/limbic system/neocortex degeneration. It is highly likely that REM sleep behaviour disorder reflects sufficient degeneration in brainstem networks, although the precise structures have yet to be identified in humans (for review, see Boeve et al., 2007b). The underlying pathology of visual hallucinations and fluctuations is not well understood.
The evolution of clinical features must reflect the topography of degeneration over time. Many of our patients tended to experience REM sleep behaviour disorder prior to cognitive impairment, with subsequent development of parkinsonism. This evolution suggests that dysfunction in the pontomedullary circuitry precedes dysfunction in the basal forebrain, limbic system and neocortical circuitry. Dysfunction in the nigrostratial circuitry appears to occur subsequently. One could argue that this evolution of features and presumed pathophysiologic basis does not support the Braak staging scheme (Braak et al., 2003, 2004) of Parkinson's disease (Burke et al., 2008), yet it may provide some clues into Lewy body disease pathology of the dementia with Lewy bodies phenotype.
A ‘bottom-to-top’ or ascending progression of Lewy neurites, Lewy bodies and neuronal degeneration as proposed in Stages 1–6 of the Braak scheme explains the evolution of features in typical Parkinson's disease quite well (Braak et al., 2003, 2004) but is less satisfactory for dementia with Lewy bodies. Findings on many ancillary tests support the contention that those with idiopathic REM sleep behaviour disorder (i.e. REM sleep behaviour disorder not associated with any other neurological symptoms or disorders) reflects underlying Lewy body disease (Albin et al., 2000; Eisensehr et al., 2000, 2003; Ferini-Strambi et al., 2004; Stiasny-Kolster et al., 2005; Caselli et al., 2006; Mazza et al., 2006; Terzaghi et al., 2008; Postuma et al., 2009) and brainstem-predominant Lewy body disease has been documented in two cases of idiopathic REM sleep behaviour disorder (Uchiyama et al., 1995, Boeve et al., 2007a), which also supports this staging system.
A similar evolution of degenerative changes may occur in both Parkinson's disease and dementia with Lewy bodies, and the timing of clinical features may reflect when the critical thresholds of neuronal network degeneration are reached. Consider a hypothetical example (this example is for illustrative purposes and does not imply these percent depletions are entirely accurate), in which an 80% depletion of dopaminergic neurons may be needed to manifest parkinsonism, and a 50% depletion of cholinergic neurons may be needed to manifest cognitive impairment; it would be reasonable to suggest a ‘bottom-to-top’ progression of Lewy body disease pathology could still explain the onset of cognitive symptoms prior to parkinsonism if both systems are affected gradually over years but the thresholds for the expression of clinical deficits are different such that cognitive impairment becomes evident prior to parkinsonism. In other words, the evolution of features in the dementia with Lewy bodies phenotype does not necessarily refute the Braak staging system for Lewy body disease progression.
Another explanation is that, at least in some cases, a ‘top-to-bottom’ or descending progression of degenerative changes from the neocortex/limbic system to the nigrostriatal system may better explain the onset of cognitive impairment prior to parkinsonism in many dementia with Lewy bodies cases. Or a more patchy and discontinuous progression could evolve (Frigerio et al., 2009), with the cholinergic neurotransmitter system reaching a critical threshold of degenerative changes prior to the nigrostriatal neuronal networks. This is similar to variations of neurodegenerative disease progression in Alzheimer's disease. Most cases of evolving Alzheimer's disease tend to follow the Braak stages of neurofibrillary tangle deposition and develop amnestic mild cognitive impairment prior to language, attention/executive and visuospatial dysfunction; however, there are certainly cases of pathologically confirmed atypical Alzheimer's disease that have presented as focal cortical syndromes such as primary progressive aphasia (Josephs et al., 2008), corticobasal syndrome (Boeve et al., 1999), and posterior cortical atrophy (Tang-Wai et al., 2004). As a result, exceptions to any model of neurodegenerative disease progression will always exist, and surely Lewy body disease will evolve in more than one manner.
These cases add to growing evidence that REM sleep behaviour disorder in the setting of neurodegenerative disease strongly suggests an underlying synucleinopathy—particularly Lewy body disease or multiple system atrophy (Boeve et al., 2001, 2003, 2007b; Gagnon et al., 2006; Iranzo et al., 2006; Postuma et al., 2009). Furthermore, REM sleep behaviour disorder tends to precede the onset of cognitive and motor features by years or decades in the synucleinopathies, whereas REM sleep behaviour disorder typically evolves concurrently or after the onset of motor and cognitive features in the non-synucleinopathy disorders. Hence, REM sleep behaviour disorder preceding the motor and cognitive features of a neurodegenerative disorder may be particularly specific for synucleinopathies (Boeve et al., 2003, 2007b; Gagnon et al., 2006; Iranzo et al., 2006; Postuma et al., 2009).
The almost complete absence of REM sleep behaviour disorder in autopsy-proven cases of amyloidopathies and tauopathies convincingly suggests that selective vulnerability of key brainstem networks underlying REM sleep behaviour disorder occurs in the synucleinopathies. There are virtually no features in clinical neurology which are 100% specific for an aetiological category of disease, and such degeneration therefore is not specific to a particular proteinopathy, but rather the same neuronal network(s) is (are) involved in patients with REM sleep behaviour disorder associated with any neurodegenerative disorder. Yet the REM sleep behaviour disorder-synucleinopathy association is strong. Knowledge about the associated neuronal networks of REM sleep behaviour disorder in humans may provide insights into why these and other key structures such as the dorsal motor nucleus of the vagus, olfactory bulb and substantia nigra pars reticularis are consistently affected in the synucleinopathy spectrum disorders.
REM sleep behaviour disorder was incorporated into the dementia with Lewy bodies diagnostic criteria and functions as a feature that increases the diagnostic confidence from one of clinically possible, to clinically probable dementia with Lewy bodies (McKeith et al., 2005). This feature has been added as a variable of interest by the National Alzheimer's Coordinating Center. Currently, the presence of recurrent dream enactment behaviour suggests a clinical diagnosis of dementia with Lewy bodies in the setting of dementia, particularly when REM sleep behaviour disorder can be verified by polysomnogram. Given that all seven patients with REM sleep behaviour disorder and mild cognitive impairment developed Lewy body disease on autopsy, the presence of REM sleep behaviour disorder—regardless of the mild cognitive impairment subtype—is likely to reflect evolving Lewy body disease.
The strengths of the study include well-characterized and prospectively followed patients who developed mild cognitive impairment and had autopsy-proven Lewy body disease. Importantly, mild cognitive impairment diagnoses were made in real-time. Subjects were not assigned mild cognitive impairment diagnoses retrospectively after a diagnosis of dementia with Lewy bodies had been made. A weakness of this clinicopathologic series is that it involves only eight cases. As a result, generalizations about the progression of symptoms in mild cognitive impairment to dementia with Lewy bodies will require further longitudinal studies. Another limitation is that due to slight changes in the neuropsychological battery over time, an identical set of neuropsychological data was not obtained on all patients. Also, at least one patient was not testable, and another was unable to complete testing due to fatigue.
Our findings indicate that Lewy body disease can pass through a mild cognitive impairment intermediate state, with any mild cognitive impairment subtype potentially evolving into dementia with Lewy bodies. All cases with REM sleep behaviour disorder and mild cognitive impairment eventually were shown to have autopsy-proven Lewy body disease, indicating that REM sleep behaviour disorder plus mild cognitive impairment probably reflects brainstem and cerebral Lewy body disease. These findings also underscore the importance of expanding the characterization beyond cognition in the dementing neurodegenerative disorders; attempts should be made to perform longitudinal assessments of the cognitive/neuropsychological, neuropsychiatric/behavioural, motor, sleep, autonomic and other features to improve the understanding of early clinical manifestations of Lewy body disease and other neurodegenerative disorders.
National Institutes of Health [AG15866 to B.F.B., T.J.F. (PI), G.E.S., V.S.P., D.W.D.; AG16574 to B.F.B., T.J.F., G.E.S., J.E.P., D.W.D., D.S.K., N.G.R., J.A.L., K.K., C.J., V.S.P., R.C.P. (PI);NS40256 to B.F.B., T.J.F., D.W.D. (PI); AG06786 to B.F.B., G.E.S., J.E.P., D.S.K., Y.E.G., V.S.P., R.C.P. (PI); and AG23195 to K.K., M.S., C.R.J. (PI)], and the Robert H. and Clarice Smith and Abigail Van Buren Alzheimer's Disease Research Program of the Mayo Foundation.
Supplementary material is available at Brain online.
The authors thank their staff at the Mayo Alzheimer's Disease Research Center and Mayo Center for Sleep Medicine for their evaluation and education/counselling for many of the patients and families included in this report. They particularly extend their gratitude to the patients and their families for participating in research on ageing, cognition and sleep.