Progressive supranuclear palsy (PSP, also called Steele-Richardson-Olszewski syndrome) is the most common cause of an atypical parkinsonian syndrome with dementia. Albert et al.[1] characterized the dementia of PSP as a "subcortical dementia" in contrast to the "cortical dementia" of patients with Alzheimer's disease (AD). PSP patients usually exhibit slow information processing and executive dysfunction evidenced by difficulties with initiation and fluency, concept formation (e.g., interpretation of proverbs), and planning and ordering events in the correct sequence.[2-11] These cognitive features, suggestive of frontal system dysfunction, [2,7,11-13] are major manifestations of PSP and occasionally may be the presenting feature. [14]
Patients suffering from dementia frequently have noncognitive behavioral changes including alterations in mood and personality.[15,16] Behavioral changes influence the burden that caregivers experience, and are powerful predictors of institutionalization.[17-21] Neuropsychiatric disturbances may predict caregiver depression and burden.[22] Characterization of the neuropsychiatric aspects of PSP has been neglected. Identification, treatment, and understanding of the associated behavioral abnormalities may improve the quality of life for both PSP patients and their caregivers.
Most published reports on behavior in PSP patients are single case studies [23-25] and emphasize symptoms related to depression; rarely, psychotic or obsessive features are described.[26,27] The frequency and severity of depression in PSP patients is unknown. There are difficulties, moreover, in evaluating depression in the setting of dementia,[28] and it may be difficult at times to discriminate symptoms of sadness from those of apathy. [28-30]
We used the Neuropsychiatric Inventory (NPI), a tool with established validity and reliability,[31] to assess the behavioral abnormalities of PSP patients. The NPI evaluates both the frequency and severity of 10 behaviors that are commonly reported in patients with cognitive disturbances: delusions, hallucinations, agitation, dysphoria, anxiety, euphoria, apathy, disinhibition, irritability, and abnormal motor behavior (e.g., pacing). The dysphoria subscale of the NPI has items related to sad affect, depressed mood, and sad thoughts and does not include neurovegetative symptoms (changes in sleep, appetite, libido) that may occur in dementia without depression. To characterize better the behavioral aspects of PSP patients and to search for distinctive behavioral characteristics, we compared their performance on the NPI with that of patients with AD, who frequently exhibit behavioral abnormalities,[16,28,32,33] and with age-matched normal controls.
The PSP study subjects consisted of 22 consecutive outpatients, presenting to the National Institutes of Neurological Disorders and Stroke (NINDS) for evaluation and participation in research studies, who fulfilled the study criteria and consented to participate in this study. All PSP subjects met the criteria of Blin et al.[12] for probable PSP, which have excellent reliability and specificity.[34] The diagnosis of one PSP patient who died was neuropathologically confirmed using the NINDS neuropathologic criteria.[35,36]
The AD study subjects consisted of 50 consecutive outpatients presenting for dementia evaluation to the University of California, Los Angeles or West Los Angeles Veterans Affairs Medical Center dementia clinics who met the study criteria and whose caregivers agreed to be interviewed. All AD subjects met the Diagnostic and Statistical Manual of Mental Disorders criteria for dementia[37] and the National Institutes of Neurological and Communicative Disorders and Stroke/Alzheimer's Disease and Related Disorders Association (NINCDS/ADRDA) criteria for probable AD.[38]
Exclusion criteria for all subjects included history of alcohol or substance abuse, head trauma with loss of consciousness, and psychiatric disorder preceding the onset of current disease. All participating subjects gave institutional consent. Caregivers and participating subjects were interviewed with the NPI, as previously described.[31] Briefly, screening questions for each behavior were posed first, and if a positive response was obtained for one of the ten behavioral domains, then this aspect was further explored with scripted questions. The caregiver rated the affected behaviors using a 1 to 4 scale for the frequency (1 being occasionally, 2 often, 3 frequently, and 4 very frequently) and 1 to 3 score for the severity (1 being mild, 2 moderate, and 3 marked). The composite score for each behavioral domain was the product of the frequency and severity subscore for that particular behavior (maximum 12). The total score of the NPI is the sum of the subscales scores. The Mini-Mental State Examination [39] (MMSE) was also administered, usually, on the same day. The Mattis Dementia Rating Scale [40] was administered to 19 of the 22 PSP patients.
Statistical analysis included analysis of variance (ANOVA), Spearman correlations, and covariant and discriminant function analysis. For each behavior, subscores were compared among the two patient groups using ANOVA. Statistical significance was accepted at p < 0.05 level.
There were no significant between-group differences in education or duration of disease. Because age was significantly different between the two patient groups, we performed a covariant analysis using age as a covariate, even though a previous study found age not to be correlated with any of the 10 studied behaviors.[28] Since there were important differences in severity of dementia between patients with AD and those with PSP (MMSE), we also performed statistical analyses including only patients with mild dementia (MMSE 20 to 30) (mild dementia). Discriminant analysis was performed on the total patient data set.
Almost all the PSP patients (91%) suffered from apathy which was rated as continuously present. Apathy severity was rated to be in the moderate-to-severe range in 80%. Thirty-six percent of the PSP patients exhibited disinhibition, which was in the moderate-to-severe range in half. Dysphoria was infrequent; only 18% (four patients) were dysphoric, mostly mildly, none of them severely. Three of the four PSP patients with dysphoria also had apathy. Less frequently, PSP patients reported anxiety (18%), irritability (two patients, 9%), aberrant motor behavior (9%), or agitation (one patient, 5%). Hallucinations, delusions, and euphoria were never reported.
To explore the relationships among behaviors in PSP, Spearman correlation coefficients were calculated between all behaviors for all PSP patients. The anxiety composite score tended to be correlated with the agitation composite score (r = 0.41, p = 0.058); the frequency and severity of agitation was associated with the severity of anxiety (r = 0.46, p < 0.05); and the irritability composite score was correlated with the frequency of apathy (r = -0.45, p < 0.05).
We explored whether the PSP patients' apathy was associated with executive dysfunction by determining correlations of this variable with the Initiation and Perseveration subtest scores of the Mattis Dementia Rating Scale. The frequency of apathy was inversely correlated with the Initiation and Perseveration subtest score (r = -0.49, p < 0.05). The combination of the apathy and disinhibition composite scores and severity of dysphoria modestly predicted the Initiation and Perseveration subtest scores (0.48, p<0.001).
We found that the anxiety composite score was correlated with the Attention subtest of the Mattis Dementia Rating Scale (-0.58, p<0.001). Duration of symptoms was inversely correlated with the Supermarket Fluency subtest of the Mattis Dementia Rating Scale (-0.60, p<0.005) and with the anxiety composite score (0.48, p<0.05).
The behavior of the AD patients included in this study has been previously described[28]; 88% of the patients had measurable behavioral changes. The most frequent behavioral abnormalities of patients with AD (n = 50) were apathy (72%) and agitation (60%); also present were anxiety (48%), irritability (42%), dysphoria (38%), aberrant motor behavior (38%), disinhibition (36%), delusions (22%), and rarely hallucinations (10%).[29]
The behavior of the 40 control subjects also has been previously described. [31] Rarely did control subjects have changes in behavior (13%), and when they did, the changes were mild and limited to three behavioral domains (three subjects had dysphoria, one of them also had irritability, and two exhibited disinhibition). The NPI scores of control subjects were significantly lower than those of patients with AD and PSP for apathy, disinhibition, and anxiety. Moreover, dysphoria composite scores above 6, disinhibition composite scores above 4, and irritability composite scores above 2 were never present among normal subjects.
There was no significant difference between the total NPI score of the patients with PSP and patients with AD. The summary score of the NPI in PSP, however, was influenced heavily by high apathy scores. When apathy was eliminated from both the PSP and AD total NPI scores, the PSP NPI scores (4.6 +/- 1.2) were then significantly lower than those of the AD patients (10.9 +/- 1.4, p < 0.01), reflecting the greater degree of psychopathology among AD patients. Table 2 shows the NPI composite scores of all AD and PSP patients. The PSP patients had significantly higher apathy scores, while AD patients had significantly higher composite scores on delusions, agitation, anxiety, and irritability scales. Since the MMSE scores were significantly different in the two groups, we identified a subset of both groups with similar MMSE scores. Table 3 shows the NPI composite scores of the subjects matched for their degree of dementia (mild). PSP patients still had significantly higher apathy score and lower agitation and irritability scale scores than AD patients. Since duration of symptoms tended to be longer for the PSP group, we used duration as a covariate, and found significant remaining between-group differences for delusions (AD worse than PSP, p < 0.05), agitation (AD worse than PSP, p < 0.001), anxiety (AD worse than PSP, p < 0.01), apathy (PSP worse than AD, p < 0.001), and irritability (AD worse than PSP, p < 0.01). Age was also significantly different between the two groups, and we used it as a covariate and found significant between-group differences for agitation (AD worse than PSP, p < 0.01), anxiety (AD worse than PSP, p < 0.05), apathy (PSP worse than AD, p < 0.01), and irritability (AD worse than PSP, p < 0.05). When we analyzed the data with the MMSE as a covariate to determine which behaviors were independent of the cognitive impairment, the only significant between-group difference was for apathy (PSP worse than AD, p < 0.01).
A discriminant function analysis performed on the total data set revealed that any patient with high apathy scale scores and low agitation and anxiety composite scores would be placed correctly in the PSP group 85% of the time in this patient sample.
Apathy (91%) was the most frequent behavioral abnormality noted in PSP patients. Apathy in PSP was usually unrelated to dysphoria. Since apathy shares many behavioral features with depression, and is frequently mistaken for depressive disorder, [28-30] PSP patients are often treated, typically unsuccessfully, with antidepressants. Caregivers also misinterpret this type of behavior as depression even when changes in mood state are denied by the patient. The NPI may help physicians and family members distinguish apathy from depression. Although most of our PSP patients who had depression also experienced apathy, most who had apathy had little or no depression. Likewise, in other related disorders, such as in Parkinson's disease, in which depression is more common,[41-48] dysphoria is not always associated with apathy. [31]
Functional imaging studies of PSP patients show that they have deafferentation of the frontal lobes. [12,49,50] They have reduced perfusion despite limited local pathologic changes. The metabolic abnormalities are in areas disconnected by the prominent subcortical pathology of PSP.[51-53] Many of the neuropsychological disturbances of PSP are analogous to those described in patients with frontal lobe disorders. [1] Executive dysfunction is commonly encountered in patients with frontal lobe disorders and appeared in this and other studies of PSP. [1-4,7,8] Moreover, there were significant correlations between executive impairment as measured by the Dementia Rating Scale and the apathy scale of the NPI, suggesting that both are mediated by abnormalities of the frontal lobe or frontal-subcortical connections. PSP patients' executive dysfunction is classically manifested by decreased verbal fluency, concreteness in thinking, lack of insight, impaired reasoning, slowed information processing, poor information retrieval, impaired control over attention or execution of sequential actions, and problems in shifting back and forth between two different tasks. [2,3,6-8,11,13]
In our patients, frontal executive function was related to the duration of the disease, suggesting that frontal function worsens at later stages, as previously suggested.[8] The observation that apathy and disinhibition are not associated with disease duration implies that progressive dysfunction of the different subcortical circuits does not proceed in parallel and that neuropsychiatric disorders are independent of cognitive dysfunction. Five frontalsubcortical circuits unite regions of the frontal lobe with the striatum, globus pallidus, and thalamus in functional systems that mediate motor activities, eye movements, cognition, and behavior.[54] The five circuits originate in supplementary motor area, frontal eye fields, dorsolateral prefrontal cortex, orbitofrontal cortex, and anterior cingulate cortex, and they mediate volitional motor activity, saccadic eye movements, executive functions, social behavior, and motivation respectively.[55] In PSP, all five circuits are affected[12,49-51,56] and our study confirms that functional and behavioral abnormalities reflecting interruption of all five circuits are also present. Motor abnormalities are associated with dysfunction of the subcortical projections to the supplementary motor area-subcortical circuit; supranuclear palsy relates with disturbances of frontal eye-field-subcortical circuit; executive dysfunction impugns the function of dorsolateral prefrontal-subcortical circuitry; disinhibition is the behavioral correlate of orbitofrontal-subcortical involvement; and apathy correlates with dysfunction of the medial frontal-subcortical circuit. The current study is among the first to document the full repertoire of behavior disorders predicted in PSP on the basis of known pathologic involvement of structures participating in the frontal-subcortical circuits. Lesions in the pedunculopontine nucleus, caudate, and the nucleus basalis of Meynert (affected to different degrees in PSP)[51,56,57] may contribute to the observed cognitive and behavioral disturbances of these patients. The main fronto-subcortical neurotransmitters affected in PSP are acetylcholine and gamma-amino butyric acid (GABA)[58,59] (dopaminergic and serotoninergic pathways are relatively spared[51,58]); cholinergic or GABAergic agonists may be beneficial palliative therapies.
Depression is the most common behavioral change previously described in PSP. [23-25,60] However, studies as large as the current one are unusual and changes in other behaviors have rarely been investigated. Depression is difficult to assess in dementia because of overlapping symptoms of the two disorders, but the NPI facilitates the identification by eliminating all non-mood items from the dysphoria subscale. With this approach, we found that depression is not common in PSP (18%). This study indicates that apathy is the dominant behavioral change of PSP and that the apathetic state is only occasionally accompanied by a mood disturbance. Differentiation of apathy and depression may help focus therapeutic interventions in PSP.
Agitation was associated with anxiety, and anxiety was in turn related to impaired attention. Treating anxiety in PSP may improve both the attention and agitation, or conversely, treatment of attentional disturbances may improve anxiety and agitation.
Our study demonstrated that behavioral changes can discriminate - although not perfectly - between patients with PSP and those with AD. PSP patients had significantly more apathy and less agitation and anxiety than AD patients. Apathy was also the most frequent behavioral abnormality experienced by AD patients when they were compared with controls (in 72% of them). [28] However, overall, PSP patients exhibit apathy more frequently and more profoundly. Apathy differentiated PSP from AD patients, independent of their cognitive impairment. Further studies should investigate whether the apathy experienced by PSP and AD patients differ qualitatively.
AD patients exhibited a variety of behavioral abnormalities including agitation, irritation, delusions, and anxiety that were less evident or not present in PSP patients. Some of these behaviors, such as delusions, were related to disease duration and to a more severe degree of dementia. These behaviors likely have a different underlying pathophysiology.
Improved knowledge of the neurobehavioral abnormalities of our patients will allow us to diagnose and manage PSP better and address more directly the subjective complaints of patients and their caregivers.
We gratefully acknowledge the patients with PSP and AD, and their families, who generously devoted their time to participate in our study.