Comparison of extrapyramidal features in 31 pathologically confirmed cases of diffuse lewy body disease and 34 pathologically confirmed cases of parkinson's disease

For the most part, literature on the clinical features of diffuse Lewy body disease (DLBD) focuses on distinguishing DLBD from Alzheimer's disease (AD) rather than Parkinson's disease (PD). ^[1-7] The extrapyramidal syndrome, and more specifically, the parkinsonism of DLBD has been difficult to characterize; the literature largely consists of single case reports or small case series. ^[8-30] There has not been a consensus regarding the proportion of DLBD cases that demonstrate parkinsonian signs; estimates include 45%, ^[31] 84%, ^[2] and 100%. ^[4,5] Moreover, there is little agreement about the type of parkinsonian syndrome seen in DLBD. Some regard it as an akinetic-rigid syndrome without rest tremor ^[1]; others report rest tremor in as many as 66%. ^[4] Some regard the syndrome as indistinguishable from PD; ^[4] others report that gait instability may be the only sign of parkinsonism. ^[1] The proportion of pathologically confirmed cases of DLBD diagnosed clinically as PD varies from 20 to 88%, ^[2,4,5,31] and further characterization of the extrapyramidal features of DLBD might facilitate antemortem diagnosis; in particular, distinguishing it from PD. Previously, ^[5] we compared the parkinsonism of eight cases of DLBD with 20 cases of PD, all pathologically diagnosed. ^[5] While we noted several trends, sample size was a limitation. To overcome this limitation, we designed the present study, examining prospective and retrospective clinical data on a large series of pathologically confirmed cases of DLBD (N = 31) and PD (N = 34) from two centers.

Methods.

There were 31 pathologically diagnosed cases of DLBD; 14 from the Columbia-Presbyterian Medical Center (CPMC) Brain Bank (1990-1995), and 17 from the University of Rochester Medical Center (URMC) Brain Bank (1984-1994). Nine (52.9%) of 17 from the URMC Brain Bank had been evaluated prospectively in the Rochester Alzheimer's Disease Project (RAPD) at URMC. These nine were recruited through affiliated clinics, hospitals, nursing homes, and community service organizations, and the importance of postmortem examination was discussed during enrollment. Six (66.6%) of these nine were diagnosed clinically as having dementia, and three were diagnosed as having both dementia and PD. There were 34 pathologically diagnosed cases of PD, representing all pathologically diagnosed cases of PD from the CPMC Brain Bank (1990-1995). Nineteen (39.6%) of 48 cases from the CPMC Brain Bank had been evaluated prospectively for their dementia or parkinsonism by neurologists associated with the CPMC Brain Bank. The remainder were referred as outside cases.

Hospital records, physicians' notes, and standardized data forms (in the case of prospectively enrolled subjects) were reviewed to assess the clinical features of all cases. The following clinical features were categorically encoded as present, absent, or not commented on: any tremor (either tremor at rest or with action), rest tremor, rigidity (increased muscle tone with passive motion), bradykinesia (movements with slow or diminished frequency or diminished amplitude), gaze palsy (conjugate inability to move both eyes in a vertical or horizontal direction, in the absence of peripheral etiologies), dystonia (sustained twisting movements), myoclonus (sudden, brief, shock-like jerks), and a documented clinical response to levodopa (clinical improvement in either tremor, rigidity, bradykinesia, or gait). Data on postural instability were not consistently present in the records. A cognitive or psychiatric abnormality was defined as a consistent demonstration of either disorientation, confusion, memory impairment, or hallucinations. ^[5] A sign was coded as "present" if it was specifically noted in the record on one or more occasions. Data were not complete for all individuals, and the denominators of all proportions excluded those cases without sufficient data. Age of onset was the age of first reported cognitive, psychiatric, or motor symptom.

Detailed neuropathologic methods employed on all autopsy cases enrolled at CPMC and URMC have been reported previously and will not be described in detail here. ^[5] Cases were divided into two pathologic diagnostic categories, DLBD and PD, based on criteria from Kosaka et al. ^[32] DLBD cases corresponded to their type A or DLBD (numerous Lewy bodies widely distributed in brainstem, diencephalic nuclei, cerebral cortex, and basal ganglia). Two additional cases with Kosaka's type B or transitional Lewy body disease were not included in this analysis. The pathologic diagnosis of PD was based on finding a clear depletion of pigmented neurons in the substantia nigra without another known cause. Presence of Lewy bodies was required for the diagnosis of PD. In cases where Lewy bodies were difficult to find, four hematoxylin-eosin-stained 7-micro m sections containing midbrain were examined. All of the DLBD cases and all but two PD cases had been stained with anti-ubiquitin antibody.

The chi squared test, Fisher's exact test, and two-tailed Student's t test or its standard normal approximation (z) were used to determine statistical significance. Odds ratio (OR) was used to assess associations between clinical features and diagnoses. Sensitivity was the probability of screening positive if disease is truly present; specificity, the probability of screening negative if disease is truly absent; and positive predictive value, the probability of disease given a positive screen.

Results.

DLBD had an older mean age of onset than PD, a shorter mean disease duration, and a similar mean age at death (Table 1). For those with DLBD who initially presented with parkinsonism, the mean age of onset was 60.0 years. In comparison, for those with DLBD who initially presented with cognitive or psychiatric changes, the mean age at onset was 70.9 years (z = 1.97, p = 0.05). There were no gender differences between DLBD and PD (see Table 1).

The majority (72.4%) of those with DLBD initially presented with cognitive or psychiatric changes, while the majority of those with PD (93.9%) presented with parkinsonism (see Table 1). Parkinsonian features were common in DLBD. Of those with pathologically confirmed DLBD, 92.3% had at least one sign of parkinsonism, while 92.8% had either rigidity or bradykinesia. A total of 45.2% of those with pathologically confirmed DLBD had received clinical diagnoses of PD (12.9% without AD and 32.3% with AD) (see Table 1). In comparison, 94.4% of those with pathologically confirmed PD had received clinical diagnoses of PD (83.3% without AD and 11.1% with AD) (see Table 1). PD (with or without dementia) was diagnosed in eight (73%) of 11 of those with DLBD and rest tremor, compared with one (11%) of nine of those with DLBD and no rest tremor (chi squared = 12.7, p < 0.001).

Tremor at rest was more common in PD than DLBD. Any tremor (including either rest or action tremor) was present in a similar number (see Table 1). Myoclonus was more common in DLBD. There were no differences in rigidity, bradykinesia, dystonia, or gaze palsies (see Table 1). There was a difference in perceived need for levodopa; almost all of those with PD compared with fewer than one-half of those with DLBD had received a trial of levodopa therapy at some point (see Table 1). A documented clinical response to levodopa may have been more common in PD (see Table 1).

A total of 87.1% of those with DLBD received clinical diagnoses of dementia compared with 16.7% of those with PD (see Table 1). Detailed analysis of the psychiatric features is the subject of another manuscript. ^[33] Briefly, psychiatric features were more common in DLBD than PD (chi squared = 4.6, p = 0.03), and most notably, delusions were more common in DLBD (57%) than PD (15%) (chi squared = 10.1, p = 0.00015).

The seven DLBD cases who had a response to levodopa may have had more typical parkinsonism than those who did not; however, numbers are small. Rigidity was present in 100%, bradykinesia in 100%, "any tremor" in six (100%) of six commented on, and rest tremor in all four (100%) commented on. Finally, six (86%) of seven were diagnosed as having either parkinsonism or parkinsonism and dementia.

Four clinical features (myoclonus, absence of rest tremor, no documented clinical response to levodopa, no perceived need to treat with levodopa) distinguished PD from DLBD (see Table 1). The occurrence of any one of these four features distinguished DLBD from PD (chi squared = 13.17, p < 0.001), and was 10 times more likely to occur in DLBD (OR = 10.29, 95% confidence interval [CI] = 2.58-41.11). The likelihood of having DLBD given any one of these four features was 85.7% (positive predictive value). Any one of these four features occurred in 66.7% of those with DLBD (sensitivity). None of these features occurred in 85.7% of those with PD (specificity).

Discussion.

We compared the extrapyramidal features of pathologically confirmed cases of DLBD and PD. Ours is the first multicenter study focusing on the extrapyramidal features of DLBD and is the largest series that focuses on these features in DLBD. All cases were pathologically confirmed. Finally, we compared DLBD cases with a large comparison group with PD.

Those with DLBD had an older mean age of onset than PD. Our figure for PD (62.0 years) is similar to that reported by the United Kingdom Parkinson's Disease Society Brain Bank (62.5-63.6 years). ^[34,35] The figure for DLBD (67.9 years), similar to Kosaka's (66.7 years), ^[36] is greater than we had reported previously (57 years). ^[5] In the present study, 24.1% with DLBD initially presented with parkinsonism rather than cognitive or psychiatric changes, in comparison with 80% in the previous study. ^[5] As demonstrated in this study, as well as others, ^[4,5] those with DLBD initially presenting with parkinsonism may have a younger age of onset or their disease may be diagnosed at an earlier stage than those initially presenting with cognitive changes. This finding is consistent with the notion that PD and DLBD represent two ends of the same disease spectrum, and that those presenting with parkinsonism have an age of onset that is closer to that in PD.

Most patients with DLBD either had bradykinesia or rigidity (92.8%), and 45.2% received clinical diagnoses of PD (see Table 1). Estimates of the proportion of DLBD cases that demonstrate signs of parkinsonism include 45%, ^[31] 84%, ^[2] and 100%. ^[4,5] There are several possible explanations for this wide range of estimates. First, variable numbers of patients may have neuroleptic-induced parkinsonism. ^[31] McKeith et al. ^[31] reported that while 80% of subjects with DLBD demonstrated signs of parkinsonism, these signs predated treatment with neuroleptic medications in only 45%. In this study, 21.4% with DLBD were treated with neuroleptic medications; all had demonstrated signs of parkinsonism prior to receiving neuroleptics or had demonstrated these signs after discontinuation of the neuroleptic medication. Second, there is a lack of uniformity in the definition of "parkinsonism." Some estimates are based on tremor, rigidity, or bradykinesia, ^[2] while others include gait or postural changes. ^[4] Our own estimates may be biased. If a sign was not commented on in a medical record, the data point was excluded rather than being scored as "absent." This may have resulted in elevated estimates. Conversely, our estimates may be low, as we evaluated nine (29%) of 31 with DLBD in the RAPD, a setting more likely to attract subjects with cognitive rather than motor complaints.

Rest tremor was more common in PD (85.0%) than DLBD (55.0%). Others have reported rest tremor in 47% ^[2] to 66% ^[4] with DLBD. The proportion of those with PD who have rest tremor is 90% in clinical series, ^[37] and 76 to 100% in pathologic series. ^[34,38]

Myoclonus was more common in DLBD (18.5%) than PD (0%). Burkhardt et al., ^[2] reporting on 38 DLBD cases (four new, 34 literature), noted that myoclonus was present in 15%. The above figures may be low, as stimulus-sensitive myoclonus was not assessed. ^[2,39] While myoclonus may be a side effect of levodopa, ^[40] none of our DLBD cases with myoclonus had received levodopa.

Dystonia was present in 6.9% with DLBD; one had received levodopa. Burkhardt et al. ^[2] reported that dystonia was present in 3% with DLBD. While dystonia was no more common in DLBD than PD, the distribution may have differed. Those with DLBD had torticollis or blepharospasm. The one PD patient with dystonia had typical early morning foot dystonia while taking levodopa. Kidron and Melamed ^[41] analyzed the types of dystonia in 207 subjects with PD who were being treated with levodopa. Dystonia most commonly affected the limbs, and rarely cranial or cervical structures. Dystonia involving cranial and cervical regions rarely has been reported in those with PD not being treated with levodopa. ^[42]

We found that in 58.3% of those with DLBD, there was no perceived need to treat with levodopa, and hence, a clinical trial of levodopa was never initiated. Perceived need for levodopa may have been a factor of the cognitive and psychiatric problems outweighing parkinsonism. There was a trend for levodopa-responsiveness to be more common in PD (100%) than DLBD (70.0%), although this did not reach significance. Doses of levodopa in our DLBD cases who did not have a response to medication are not known. Byrne et al. ^[4] found that 10 (91%) of 11 subjects with DLBD who were treated with levodopa had a definite response; one of 11 had a probable response. Conversely, Mark et al. ^[43] showed that transitional forms of Lewy body disease may be levodopa-unresponsive.

Cognitive deficits were more common in DLBD than PD. The majority of those with DLBD received clinical diagnoses of dementia compared with few of those with PD (see Table 1). In addition, psychiatric features, most notably delusions, were more common in DLBD than PD.

Those with DLBD were 10 times more likely than those with PD to have one of four clinical features (myoclonus, absence of rest tremor, no response to levodopa, or no perceived need to treat with levodopa). The sensitivity of these features was moderate (66.7%), and the specificity and positive predictive value were higher (85.7%). However, these estimates should be interpreted with caution; the relative frequencies of DLBD and PD in this autopsy population (nearly a 1:1) are probably not representative of a cohort derived from a movement disorders clinic. In addition, a separate issue is whether one may distinguish DLBD from parkinsonism plus syndromes (e.g., multiple system atrophy [MSA]) based of these features. Wenning et al. ^[44] examined clinical features in 100 cases of MSA. Their data suggest that rest tremor and response to levodopa may be less common in MSA than in DLBD because typical rest tremor was present in only 9% and only 29% had initial good or excellent response to levodopa. Dementia was not an integral feature of MSA. Conversely, neither the dysautonomia nor the ataxia, present in 97% and 52% of those with MSA, was present in any of our cases of DLBD. However, these potential differences must be approached with caution because the studies may have differed in terms of definitions and case ascertainment.

This study has limitations. First, DLBD is generally diagnosed pathologically, and hence, many data were obtained retrospectively. Despite this limitation, these data identify a number of clinical features that might distinguish PD from DLBD. This study also provides information that may be used to design data collection instruments for future prospective studies. Second, because of the rarity of DLBD, the sample was not population-based, and may not be representative of DLBD in the population.

In summary, we analyzed the extrapyramidal features of a large series of autopsy-proven cases of DLBD and PD seen at two medical centers. Our results indicate that the presence or absence of rest tremor, myoclonus, and perceived need for levodopa distinguish DLBD from PD. Additionally, the distribution of the dystonia may have differed. These features, in combination with reported differences in psychiatric manifestations, ^[33] may be used for diagnostic purposes, in distinguishing DLBD from PD, in prospective longitudinal cohort studies of DLBD.

Acknowledgments

We would like to thank Mrs. Margherita Frankel for her generous support of this research on diffuse Lewy body disease. We would like to thank Dr. James Goldman (Department of Pathology, Division of Neuropathology, Columbia University) and Dr. James Powers (Department of Pathology, Division of Neuropathology, University of Rochester) for reviewing this manuscript and supervising review of pathologic materials.