Psychiatric features in diffuse Lewy body disease

In 1961, Okazaki et al. ^[1] reported two cases of progressive dementia associated with large numbers of Lewy bodies in the cerebrum, brainstem, and spinal cord. Since then, there have been reports of at least 119 similar cases. ^[2-10] Perry et al. ^[2] described two forms of dementia accompanied by cortical Lewy bodies: diffuse Lewy body disease (DLBD), which tends to present at an earlier age and is usually associated with a parkinsonian syndrome, and senile dementia of the Lewy body type, which tends to present after age 70, with or without mild extrapyramidal features. These two forms are comparable with Kosaka's ^[3] two distinct forms of diffuse Lewy body disease: a "pure" form, which presents in younger patients as a parkinsonian syndrome, and a "common" form, tending to present as memory impairment in the elderly.

Although there is adequate documentation of the cognitive and some aspects of the extrapyramidal features associated with DLBD, ^[2-8,11] the psychiatric manifestations of this illness have received less attention. Some previously reported cases have included psychotic features, ^[2-6,8,9] leading to the inclusion of these features in the 1992 operational criteria for senile dementia of the Lewy body type proposed by McKeith et al. ^[10] These criteria require fluctuating cognitive impairment involving memory and higher cortical functions, a progressive course (often rapidly progressive), absence of other illnesses that may account for the fluctuating cognitive state, and no history of stroke. In addition, any one of the following must be present: visual or auditory hallucinations (usually accompanied by paranoid delusions), extrapyramidal features or neuroleptic sensitivity, repeated falls, or transient impairment of consciousness. ^[10] Although hallucinations and paranoid delusions are mentioned, they are not required for the diagnosis using these criteria. A limitation of this study is that depression and specific types of delusions other than paranoia were not discussed. Psychotic features are also included in the 1996 criteria for the clinical diagnosis of dementia with Lewy bodies (DLB). ^[12] In addition to progressive cognitive decline with impaired social or occupational functioning, these criteria require two of the following features for a diagnosis of probable DLB: fluctuating cognition with pronounced variations in attention and alertness; recurrent visual hallucinations, typically well formed and detailed; and spontaneous motor features of parkinsonism. For a diagnosis of possible DLB, one of three features must be present. Other features supportive of the diagnosis of DLB include repeated falls, syncope, transient loss of consciousness, neuroleptic sensitivity, systematized delusions, and hallucinations in other modalities. ^[12]

Previous studies have been limited by small sample sizes ^[6-9] or lack of control groups. ^[3-5] Some studies did not differentiate between different psychiatric phenomena. ^[3,9] The purpose of the present study was to determine the frequencies of specific psychiatric symptoms, including depression, hallucinations, and delusions in a large, multicenter cohort of demented subjects who were diagnosed with DLBD at autopsy, to compare this to the frequencies in two comparison groups.

Methods.

All cases were selected from either the Rochester Alzheimer's Disease Project (RADP) or the Columbia-Presbyterian Medical Center (CPMC) Brain Bank. The RADP is a multispecialty project at the University of Rochester Medical Center (URMC) through which AD patients and elderly control subjects are enrolled into clinical studies. Prospective subjects are recruited through affiliated clinics, hospitals, nursing homes, and community service organizations. The importance of a postmortem examination is discussed with all subjects and caregivers at the time of enrollment. Subjects may also be enrolled retrospectively at time of death. All CPMC DLBD and Parkinson's disease (PD) subjects came from the Brain Bank. Some of these had been followed by neurologists associated with the Brain Bank. Others were referred as outside cases.

We screened the medical records of all 170 subjects enrolled in the RADP who came to autopsy between 1984 and 1994 to identify those subjects with a pathologic diagnosis of DLBD or Alzheimer's disease (AD). Subjects with concurrent neuropathologic diagnoses such as stroke were excluded. Subjects were selected from the CPMC Brain Bank based on a diagnosis of DLBD or PD. A total of 124 subjects from the RADP and CPMC Brain Bank were selected. Twelve cases (4 with DLBD, 2 with AD, and 6 with PD) were excluded because of incomplete records. The remaining 112 subjects were divided into three diagnostic groups. The DLBD group consisted of 28 patients who met pathologic criteria for DLBD and included 18 subjects from the RADP and 10 from the CPMC Brain Bank. These subjects were compared with a group of 58 demented patients who were found to have AD at autopsy, without cortical Lewy bodies. The PD comparison group was drawn from the CPMC Brain Bank and included 26 subjects with an autopsy diagnosis of PD, also without cortical Lewy bodies. Hospital records, consulting physician's notes, and standardized data forms (in the case of prospectively enrolled subjects) were reviewed to assess the psychiatric features in all cases. The following clinical features were categorically encoded as present, absent, or not commented on: depression, hallucinations, and delusions. The category of present was used if the feature was specifically mentioned in the record on one or more occasions. Depression was counted as present if the diagnosis was documented by a primary or consulting health provider. Adequate documentation was not present in most cases to retrospectively apply DSM-IV ^[13] criteria for the diagnosis. Hallucinations were defined as sensory experiences that occur without external stimulation of the relevant sensory organ. ^[13] They were classified as auditory, visual, both auditory and visual, or unknown type. Delusions were defined as false beliefs, firmly held despite evidence to the contrary, and not ordinarily accepted by other members of the person's culture or subculture. ^[13] They were classified as paranoia, misidentification (identifying a familiar person incorrectly), phantom boarder (the thought that other people are living in the patient's home), theft, and other fixed delusions. The following cognitive symptoms were coded as present, absent, or not commented on: disorientation, language impairment, visuospatial impairment, impaired judgement, and memory impairment. A diagnosis of dementia was made retrospectively if the patient either displayed impairment in three or more cognitive areas ^[14] or if a diagnosis of dementia was indicated in the medical record.

Detailed pathologic methods used on all autopsy cases enrolled in the RADP and CPMC Brain Bank have been previously reported and are not described in detail here. ^[15,16] A pathologic diagnosis of DLBD was based on Kosaka's criteria and required the presence of numerous Lewy bodies in sections of cerebral cortex. ^[17] All DLBD and AD cases were immunohistochemically stained with antiubiquitin antibody. Twenty-four of 26 PD cases had been stained with antiubiquitin antibody; the remaining 2 cases had idiopathic PD with no atypical extrapyramidal or cognitive features.

Results.

Results are summarized in Table 1. Men out-numbered women by approximately 2:1 in the DLBD and PD groups. The gender distribution in the DLBD group was significantly different from the AD group, which was predominantly female (chi squared = 14.36, p = 0.0002). The mean age at time of death did not differ significantly between the DLBD and AD groups (T = -1.31, p = 0.19) or between the DLBD and PD groups (T = 1.29, p = 0.20). Mean duration of illness was longer in the PD group than in the DLBD and AD groups (F = 6.91, p = 0.0016). There was no difference between the DLBD and AD groups with regard to the duration of illness (T = 0.36, p = 0.072).

Depression was present in 50.0% of DLBD subjects. Of these 14 cases, 11 (79%) were treated with either antidepressant medication or electroconvulsive therapy; many of these cases were refractory to treatment, even after several medications were tried. In an additional case, the diagnosis was made by a psychiatrist, who recommended a trial of antidepressant medication. Depression was significantly more common in this group than in the AD group, in which 13.8% of subjects were depressed (chi squared = 13.00, p = 0.0003). The frequency of depression in the DLBD group did not differ significantly from the PD group, in which 57.7% of subjects were depressed at some time during the course of the illness (chi squared = 0.32, p = 0.57).

Hallucinations were present in 60.7% of DLBD cases. In four cases (23.5%), the hallucinations were visual only; in one (5.8%), auditory only; in four (23.5%), both visual and auditory; and undefined in eight (47.0%). There were no other types of hallucinations reported in the DLBD subjects. The frequency of hallucinations was significantly greater in this group than in the AD group (chi squared = 5.30, p = 0.021), in which 34.5% subjects had hallucinations. Of the 20 AD patients with hallucinations, they were visual in 10 (50%), both visual and auditory in 6 (30.0%), and of unknown type in 3 (15.0%). One subject had olfactory and gustatory hallucinations, thought to be related to seizure activity. Hallucinations were present in 53.8% of the PD group. They were visual in two (14.3%) and undefined in 12 (85.7%). There was no difference in the frequencies of hallucinations in the DLBD and PD groups (chi squared = 0.26, p = 0.61).

In 57.1% of DLBD cases, delusions were present. Paranoid delusions were the most common type and were present in 11 cases (68.7%); misidentification syndrome was present in 6 cases (37.5%), delusions of theft in 3 (18.8%), and phantom boarder in 2 (12.5%). In one case, the type of delusion was not defined. In five cases (31.3%), another fixed delusion was present; for example, one patient was convinced that his wife was talking in a special code to tell others to come take him away. In nine cases, more than one type of delusion was present. The frequency of delusions in the AD group did not differ from the DLBD group (chi squared = 0.10, p = 0.75). Of the 31 AD cases with delusions, misidentification and paranoia both occurred in 15 (48.4%). Nine (29%) had delusions of theft, five (16.1%) phantom boarder, and three (9.68%) other delusions. Delusions were reported in four PD subjects (15.4%). They were paranoid in one patient (25%) and undefined in the remaining three (75%). Delusions were less common in the PD subjects than in the DLBD cases (chi squared = 10.08, p = 0.0015).

Seven of 28 DLBD subjects had taken antiparkinsonian medications. Of these, all seven (100%) had reported hallucinations, compared with 10 of 21 DLBD subjects (47.6%) who were not treated (chi squared = 6.04, p = 0.014). Delusions were reported in five treated patients (71.4%) and 10 untreated patients (47.6%), a difference that did not reach statistical significance (chi squared = 0.78, p = 0.38). There was no difference in the frequency of hallucinations in the 17 PD subjects who were treated with these medications (58.8%) compared with the nine subjects who were not (44.4%) (chi squared = 0.49, p = 0.48) or in the frequency of delusions in PD subjects who were treated (17.6%) and those who were not (11.1%) (chi squared = 0.19, p = 0.66). No AD subject had taken antiparkinsonian medications.

All but one (96.4%) of the DLBD patients exhibited at least one psychiatric feature, in contrast to the AD group, in which 65.5% had at least one of the three symptoms (chi squared = 9.67, p = 0.0018), and the PD group, with 76.9% of subjects developing at least one of the psychiatric features during the course of the illness (chi squared = 4.55, p = 0.033).

Because there was a gender difference between the DLBD and AD groups and these groups also displayed very different frequencies of depression, we calculated the frequency of depression in both men and women with DLBD. No difference in the frequency of depression was found between gender groups, suggesting that gender does not play a significant role in the development of depression in DLBD.

All 28 DLBD and all 58 AD subjects were demented. Of the 26 PD subjects, 8 had dementia; in an additional 10, one or two cognitive symptoms were noted, such as forgetfulness or disorientation. In the remaining eight subjects, cognitive dysfunction was specifically noted to be absent.

Discussion.

Psychiatric symptoms are often cited as features of patients with Lewy body disease. Burkhardt et al. ^[4] reviewed 34 cases of DLBD, 30 of which had been previously reported in the literature. Psychotic features were present in 10 subjects (29%), with depression in 3 (9%), hallucinations in 5 (15%), and paranoia in 4 (12%). Two subjects had previous diagnoses of schizophrenia. ^[4] Kosaka ^[3] reviewed 37 previously reported Japanese cases, some of which were discussed in the review by Burkhardt et al. He also found that psychiatric symptoms as the presenting complaint were frequent: 17.9% of the 28 "common" cases and 22.2% of the remaining nine "pure" cases. ^[3] Crystal et al. ^[9] found "delusions, agitation, depression, anxiety, and hallucinations" in five of six DLBD subjects in the mild to moderate stages of the disease, in contrast to three AD subjects in whom no psychiatric abnormalities were seen until a severe dementia was present. In a review of 15 DLBD cases, Byrne et al. ^[5] found auditory and/or visual hallucinations and paranoid ideation in five and depression in an additional three subjects. These studies were limited in that no control groups were used for comparison.

In controlled studies, results have been conflicting. Hansen et al. ^[6] found no difference in the frequency of depression in nine subjects with diffuse neocortical Lewy bodies matched for severity of dementia with nine pure AD subjects (2 cases in each group). No subjects in either group had reported delusions or met the DSM-III criteria for psychosis. Forstl et al. ^[7] compared eight demented subjects with cortical Lewy bodies with eight age- and sexmatched AD subjects and found no difference in age of onset, duration of illness, or number of subjects with hallucinations (2 in each group). However, in a review of 14 subjects with senile dementia of the Lewy body type (SDLT), Perry et al. ^[2] found that nine (64%) presented with and continued to have visual hallucinations. In a comparison group of 14 AD subjects, none presented with hallucinations, and only 3 (21%) developed hallucinations (mainly auditory) later in the course. They also found a shorter mean duration of illness (3 +/- 2 years) in the SDLT group compared with 6 +/- 3 years in the AD group. ^[2] McKeith et al. ^[10] compared the frequency of visual and auditory hallucinations and delusions in a cohort of 20 SDLD subjects with a randomly selected group of AD subjects. They found that in the SDLT group, the frequency of visual hallucinations was 80%, auditory hallucinations 45%, and delusions 80%. In the AD subjects, the frequency of visual hallucinations was 19.1%, auditory hallucinations 0%, and delusions 19.1%. ^[10]

The frequency of psychiatric symptoms in our AD and PD comparison groups is similar to that reported in the literature. Depression is reported to occur in 10 to 30% of AD patients. ^[18] Wragg and Jeste ^[19] reviewed 30 studies reporting the frequencies of affective and psychotic symptoms in AD patients. The reported frequencies of depressive disorders ranged from 0 to 86%, with a median frequency of 19%. Most frequencies were in the 10 to 20% range. The median frequency of hallucinations was 28% (range 21 to 49%), with visual hallucinations (22%) more common than auditory hallucinations (13%). Most studies reported delusions occurring in the 30 to 38% range, with a median frequency of 33.5%. ^[19] Cummings ^[20] reviewed 26 studies that investigated the frequency of depression in PD and found a mean frequency of 40% (range 4-70%). When early studies were eliminated from the analysis, the frequency was higher; a mean frequency of 43% was seen in the studies reported from 1983 to 1990. ^[20] In a retrospective study by Mayeux, ^[21] 53% of 339 patients with PD who were seen during an 18-month period had a history of depression at some point during the illness.

Hallucinations and delusions in PD are often attributed to the use of antiparkinsonian medications, although they are also reported in PD even before the use of these agents. ^[22,23] In a review of the literature, Fischer et al. ^[24] found frequencies of psychotic symptoms ranging from 28 to 62%; in their longitudinal study of 25 patients with idiopathic PD who were treated with combination therapy, 80% had had at least one pharmacotoxic psychosis, with psychotic symptoms being more frequent in demented PD patients. Antiparkinsonian medications may have triggered hallucinations and delusions in some of our subjects, although they were also frequent in those who were not treated. We found that only hallucinations were more frequent in the DLBD subjects who were treated with antiparkinsonian medications than in those who were not; there was no difference in the frequencies of delusions in DLBD subjects or in either hallucinations or delusions in PD subjects with or without the use of these drugs.

Our observation that paranoid delusions are common in DLBD was in agreement with previous reports. However, we also found other specific types of delusions such as misidentification and theft, both in DLBD and AD. Although these types of delusions are commonly reported in AD, ^[18] their exact frequency has not been consistently documented in DLBD.

Whether DLBD is a variant of AD, ^[6] PD, ^[25] or both is controversial. Some have argued that DLBD lies on a clinicopathologic spectrum between AD and idiopathic PD. ^[26] Others describe DLBD as a distinct clinical entity. ^[2,27] Our findings could be consistent with either viewpoint. Only future investigations will allow a definitive conclusion.

Our study had limitations. First, it was a retrospective study. Subjects were not matched for the severity of dementia or for the severity of the psychiatric symptoms. It is possible that in different types of dementia there are differences in the characteristics of these symptoms that would not be detected by the methods used in this study. We could not code all symptoms, in all cases, as definitely present or absent. We dealt with this by excluding those cases in which clinical information was very limited. In the remaining 112 cases, we counted as absent those symptoms that were not commented on, as well as those that were noted to be absent were counted as absent. This may have led to our observed frequencies being lower than the true values.

Ours is the first multicenter study specifically focusing on psychiatric features affecting DLBD patients. Our study also represents the largest cohort of DLBD cases to date. Finally, we compared DLBD subjects with two comparison groups. Our results indicate that psychiatric findings are common in DLBD. In fact, all but one of our DLBD subjects displayed at least one psychiatric feature. Because absence of any psychiatric features in these cases was rare, we suggest that psychiatric symptoms be given more emphasis in the diagnostic criteria for DLBD.