Use of the Neuropsychiatric Inventory to Characterize the Course of Neuropsychiatric Symptoms in Progressive Supranuclear Palsy

Due to the striking clinical features of the classical phenotype of progressive supranucelar palsy (PSP),¹ including early postural instability with backward falls, vertical supranuclear gaze palsy, and symmetrical, mainly proximal akinetic-rigid syndrome, this devastating disorder has been predominantly characterized as a motor disorder. However, several studies analyzed behavioral and neuropsychiatric symptoms of PSP and emphasized depression and/or apathy as the most prevalent.^2–7 Moreover, the clinicopathological studies of PSP cases revealed that in some of them, behavioral features could predominate the clinical presentation, with considerable overlap with a differential diagnosis of frontotemporal dementia.^8,9 Studies dealing with a quality of life in PSP also focused the neuropsychiatric symptoms as important determinants of patients’ life satisfaction.¹⁰

Therefore, in order to contribute to better characterization of psychiatric abnormalities in PSP, we aimed to determine the pattern of neuropsychiatric symptoms in our cohort of PSP patients and their dynamic changes over a follow-up period of 1 year, as well as to examine the effects of different variables on these changes.

Methods and Participants

In this prospective cohort study conducted in tertiary hospital settings, we initially recruited 59 consecutive patients diagnosed as probable PSP,¹ who were able to comply with the requirements of the study (Table 1). Based on the prominent symptoms in the first 2 years of the disease duration, 51 patients fulfilled criteria for classical phenotype of Richardson syndrome from the very beginning of the disease, while only eight patients were classified as parkinsonian variant of PSP.¹¹ At the moment of examination, after almost 6 years of the disease duration (mean±SD=5.9±1.6), all of the patients with parkinsonian phenotype of PSP had already developed the classical features of PSP, which is in line with previous reports.¹¹ Patients with a previous or current history of other neurological, psychiatric, and major medical disorders, as well as those with a history of substance abuse and neuroleptic therapy, were not included in the study.

The study was approved by the Ethics Committee of the Faculty of Medicine, University of Belgrade, and written, informed consent was obtained from each patient.

At the study entry (baseline), a detailed demographic and clinical interview was performed. The age at onset was defined as the age of the first appearance of symptom(s) attributable to PSP according to history and medical charts, supported by an interview with the patient and the patient’s caregivers (spouses in the majority of cases). Staging of PSP was obtained using the Hoehn and Yahr staging system.¹² Patients’ disability was assessed using the Unified Parkinsonʼs Disease Rating Scale (UPDRS),¹³ as well as the PSP Rating Scale (PSPRS), which was divided into the following six subscales: PSPRS history, PSPRS mentation, PSPRS bulbar, PSPRS ocular motor, PSPRS limb motor, and PSPRS gait.¹⁴ Higher values in each subscale and in the total PSPRS represented higher level of activity limitations and participation restrictions. Cognitive functions were evaluated using the revised form of the Addenbrooke’s Cognitive Examination-Revised (ACE-R),¹⁵ the Frontal Assessment Battery (FAB),¹⁶ and the Mattis Dementia Rating Scale (DRS).¹⁷

The presence of psychiatric symptoms was assessed using the Neuropsychiatric Inventory (NPI),¹⁸ which has been utilized to examine behavioral disturbances in prior studies of PSP.^2,3,5 The scoring of the NPI was based on the interview with the main caregivers: 38 spouses, 12 children, and nine siblings. The NPI assesses the frequency (4-point scale) and severity (3-point scale) of 10 neuropsychiatric disturbances (delusions, hallucinations, agitation, dysphoria, anxiety, euphoria, apathy, disinhibition, irritability, and aberrant motor behavior), and a score from 0 to 12 was obtained for each scale by multiplying frequency by severity. The total score of the NPI is the sum of the subscale scores. In addition, depressive and anxiety symptoms, as well as apathy, were evaluated using the Hamilton Depression Rating Scale (HAM-D),¹⁹ the Beck Depression Inventory (BDI),²⁰ the Hamilton Anxiety Rating Scale (HAM-A),²¹ and the Apathy Scale (AS),²² respectively.

In the longitudinal part of the study, 25 participants were subsequently followed-up for 1 year (mean follow-up: 12.1 months; range: 11.0–13.3 months). Of the 34 patients assessed at baseline, five died, 12 declined further participation due to the severity of PSP, 10 had caregivers who were present at baseline but were not accessible at the follow-up visit in order to perform the NPI, and the remaining seven withdrew consent or were lost to follow-up.

Statistical Analysis

Statistical analysis of baseline data included both correlation and multiple linear regression analysis. Since NPI scale scores were not distributed normally, nonparametric Spearman’s correlation coefficient was used. To determine the disease-related factors associated with the increased NPI total score, multiple linear regression analysis was performed. The dependent variable in this model was the NPI total score, while motor, cognitive, and neuropsychiatric variables were independent factors (UPDRS total, UPDRS-III, HAM-A, HAM-D, BDI, AS, Mini-Mental State Examination [MMSE], PSPRS total and six subscale scores, ACE-R total and five subscale scores, FAB total, DRS total and five subscale scores). The model was adjusted by age, gender, disease duration, education, and Hoehn and Yahr scale score as potential confounding variables.

The value of changes of selected motor, cognitive, and neuropsychiatric variables between baseline and the 1-year follow-up was quantified using the Wilcoxon signed-rank test. The level of these differences was calculated as an effect size ([mean follow-up-mean baseline/pooled SD]) with 95% confidence intervals. According to Cohen’s thresholds,²³ an effect size was categorized as follows: trivial (0–0.19), small (0.20–0.49), medium (0.50–0.79), and large (≥0.80). Backward stepwise linear regression analyses were used to examine how the changes of different clinical variables over the 1-year follow-up period (duration of disease, Hoehn and Yahr scale, UPDRS total, UPDRS-III, HAM-A, HAM-D, MMSE, BDI, AS, PSPRS total, ACE-R total, FAB total, and DRS total as independent variables) contributed to the change of the NPI total score over the follow-up period (dependent variable).

Probability value of p<0.05 was considered significant. The SPSS 17.0 statistical software package (SPSS Inc., Chicago) was used in the statistical analysis.

Results

NPI Scores

Demographic and clinical features of the 59 PSP patients included in the cross-sectional baseline cohort and the 25 PSP patients from the longitudinal cohort are presented in Table 1.

Only one among the 59 PSP patients had no reported symptom on the NPI. The most common symptoms, registered in more than two-thirds of PSP patients, were apathy, depression, and disinhibition (92%, 76%, and 64%, respectively). The highest mean scores were also found in the same NPI domains. These results persisted even when severity scores were calculated only for those who experienced particular symptoms. However, since low scores on the NPI items may be clinically trivial, the analysis also involved those with a score ≥4 on at least one of the NPI items: again, most frequent was apathy (78%), followed by depression (39%) and disinhibition (36%) (see Table 2).

Delusions, hallucinations, and euphoria were infrequent in our patients.

Relations Among NPI Subscores

The total NPI score in PSP patients was strongly influenced by all of the NPI subscores. The apathy subscore highly correlated with depression (ρ=0.263; p=0.045), disinhibition (ρ=0.463; p<0.001), and aberrant motor behavior (ρ=0.258; p=0.049). Additionally, depression highly correlated with anxiety (ρ=0.440; p<0.001). Disinhibition was significantly influenced by all other subscores on the NPI (delusions: ρ=0.516, p<0.001; hallucinations: ρ=0.362, p=0.005; agitation: ρ=0.745, p<0.001; anxiety: ρ=0.274, p=0.036; euphoria: ρ=0.612, p<0.001; apathy: ρ=0.463, p<0.001; irritability: ρ=0.708, p<0.001; aberrant motor behavior: ρ=0.703, p<0.001), except by the depression domain.

Disease-Related Factors Associated With NPI Total Score

Multiple linear regression analysis revealed that several motor, cognitive, and neuropsychiatric variables had significant influence on the NPI total score in the cross-sectional baseline cohort of PSP patients (Table 3). The independent determinants of increased NPI total score were anxiety, depression, apathy, PSPRS total score, some domains of ACE-R (attention/orientation, fluency, visuospatial) and total ACE-R, and lastly, DRS initiation/perseveration.

Changes in the Evaluated Clinical Variables Over the 1-Year Follow-Up Period

Changes in the clinical, motor, cognitive, and neuropsychiatric variables in 25 PSP patients who were followed for 1 year are presented in Table 4. The large effect size (>0.80) was observed in the motor sphere (PSPRS and subscales, UPDRS total and motor scores, Hoehn and Yahr scale), as well as in the cognitive sphere (MMSE, Mattis DRS, while the ACE-R and FAB were less sensitive to change). Among neuropsychiatric measures, only the NPI total score and the AS have shown large effect sizes.

In the final regression model, after elimination of the nonsignificant variables during the backward stepwise process, the independent predictors of change of the NPI total score over the 1-year follow-up period were found to be changes in the Hoehn and Yahr scale (0.739±0.619), the HAM-A (0.261±2.895), the MMSE (−1.521±1.648), and the ACE-R total score (−3.826±7.761) (see Table 5).

Discussion

In the cross-sectional part of this study, apathy was the core neuropsychiatric symptom in our cohort of PSP patients, followed by depression and disinhibition. The general pattern of psychiatric problems obtained by the NPI was comparable with previous studies, which underscored apathy as the key and most consistent neuropsychiatric feature of PSP.^{2,3,5–7,24,25}

Generally, apathy was recognized as a specific construct of behavioral, affective, and cognitive features, arising from alterations in the prefrontal cortex-basal ganglia circuits.^26,27 The possible explanation for a high rate of apathy in PSP lies in the prominent subcortical and brainstem pathological changes with the depletion of dopamine and other neurotransmitter levels, followed by the lack of stimulation of frontosubcortical circuits.²⁸ Several neuroimaging studies in PSP demonstrated these frontal lobe disconnections in circuits associated with apathy, in particular affection of the mesiofrontal circuits and the anterior cingulum.^29,30 However, present apathy scales used in PSP studies are constructed to measure unidimensional aspects of apathy, despite growing evidence of its multidimensional nature in neurodegenerative disorders.³¹ The new recently developed Dimensional Apathy Scale,³² specifically designed to measure three neurobehavioral apathy subtypes, might help us to unravel the different factors contributing to decreased goal-directed behaviors and neurobiological bases of apathy. Finally, defining different components of apathy and understanding their interplay could help us in establishing individual treatment approaches for such patients.

Both prevalence and the NPI depression score were above the values observed in previous studies,^2,3,5,25 although the rates varied among studies.^4,7 Yatabe et al.²⁵ previously noticed that one of the possible reasons for this discrepancy in the level of depressive symptoms lies in the type of the referral center for PSP patients.²⁵ Studies based on data from movement disorders clinics,³³ including the results from our study, showed more severe depression (higher NPI depression score) in a majority of patients than the studies conducted by dementia clinics.^2,5,25,28

Apathy and depression were statistically highly correlated, but different NPI subdomains also contributed to these two neuropsychiatric features. For example, disinhibition and aberrant motor behavior strongly influenced only the NPI apathy score, while anxiety was highly correlated with the depression score but not with the apathy score. Despite the fact that the NPI utilized caregiver ratings that might misinterpret apathy and depression to some extent, these data suggest that we were probably dealing with two different psychiatric features and that the NPI was a sensitive tool for their distinction. In line with this, when measured by the patient-oriented instruments (AS and HAM-D/BDI, respectively), apathy and depression were identified, among other variables, as independent determinants of increased NPI total score in multiple regression analysis.

The third most prevalent psychiatric symptom in our cohort was disinhibition, also associated with the frontal lobe damage. It might be present in up to a third of PSP patients,^2,3,5 which was in accordance with our observation that 36% of these patients had an NPI disinhibition score ≥4. Although depression and disinhibition were both assigned to the disruptions in orbitofrontal circuits,⁷ we were not able to find any correlation between the specific NPI scores of these two psychiatric problems.

Finally, positive psychiatric symptoms such as hallucinations and delusions were infrequent in our PSP patients. They were more likely to be a part of the clinical spectrum of Parkinson’s disease (PD) and other synucleinopathies, as a result of complex interplay between disease factors and dopaminergic therapy.^5,34

In our longitudinal study, the majority of clinical, motor, cognitive, and neuropsychiatric variables showed significant changes during the 1-year follow-up, reflecting rapid, progressive course of PSP (Table 4). Despite the fact that our longitudinal study was based on only 25 PSP patients from a single center, annual differences and effect sizes in the scores of PSPRS, designed to capture deficits in the different functional domains in PSP, were in agreement with a recent multicentric report on 144 patients.³⁵ The largest effect sizes were observed in the total PSPRS score and gait domain, while limb and mentation domains showed the smallest effect sizes.

Similar to previous studies, we demonstrated that cognition does not change appreciably over the course of a year.^35–37 In terms of scales addressing global cognitive status in our cohort of PSP patients, the largest effect sizes were observed in the DRS total score, together with DRS conceptualization and DRS construction subscores, while the ACE-R was shown to be less sensitive to reflect disease progression in PSP. In line with previous reports, the FAB scale showed no ability to capture decline in frontal cognitive status in our PSP patients, probably because (due to previous observation) this scale does not correlate with disease duration and severity.^35,36,38 Although these results cannot be safely recommended due to the relatively small number of patients, the DRS should be considered in future trials as a potential outcome measure. Consistent with this observation, the DRS demonstrated better specificity and sensitivity than that of the ACE-R and FAB in parkinsonian disorders.³⁹ Moreover, the Movement Disorder’s Task Force has recently put the DRS on the list of the suggested scales to assess cognitive features of PSP patients.⁴⁰

Among neuropsychiatric measures, only the NPI and the AS showed a large effect in our cohort of PSP patients, while depression and anxiety scales (HAM-A, HAM-D, BDI) were not sensitive enough to capture behavioral changes over the 1-year follow-up. Apathy was not only the most frequent psychiatric symptom in PSP patients, but compared with depression and anxiety, it deteriorated most profoundly over the follow-up period. In a longitudinal study of PD patients, Zahodne et al.⁴¹ found that apathy and depression showed a different pattern of deterioration over time: linear worsening was associated with apathy, while depression seemed to follow a fluctuating course over time.

In contrast to our data, Litvan et al.⁴² did not observe significant changes in features measured by the NPI in PSP patients. One possible explanation for this contrast is that our PSP patients were more impaired at baseline (i.e., UPDRS-III, UPDRS total, and PSP total scores).

Similar to several other reports, data from regression analyses indicated that the total NPI score was independently influenced by measures of executive dysfunction.^2,5,6 On the other hand, the NPI score was also influenced by measures of PSP severity, which is in contrast to previous finding of Gerstenecker et al.³ Furthermore, the progression to higher Hoehn and Yahr scale scores, together with cognitive deterioration, strongly influenced the worsening of neuropsychiatric symptoms in the follow-up. This implied some correlation in disruptions of the frontal-subcortical circuits mediating motor, cognitive, and behavioral abnormalities in the course of PSP,⁴³ at least in the advanced stages of the disease.

Several limitations of our study should be mentioned. First, the NPI reflected caregiver ratings, and the evaluated symptoms might be misrepresented to some extent. Second, depression, apathy, and anxiety were additionally evaluated using multiple-item questionnaires and were not diagnosed according to clinically validated criteria, which required caution in interpreting data. Third, the dropout rate of more than 50% in our longitudinal study might have affected our results, especially if we assume that this applied to the most severe cases, which made them underrepresented in our analysis. Fourth, we did not include a control group; however, since the main aim of our study was to determine the neuropsychiatric profile in PSP and its dynamic changes over time, we found that a control group was not obligatory. Finally, we used the 10-item NPI scale and did not analyze neurovegetative changes using the 12-item version, which have been reported be common in PSP.^3,33

In conclusion, our study is one of few longitudinal studies on PSP, to our knowledge, that provides some new insight into dynamic changes of the neuropsychiatric profile of PSP patients. Our findings suggest that apathy was a predominant feature of the behavior profile of PSP and the most prone to deteriorate over the follow-up period. Finally, the NPI seemed to be a sensitive measure of behavioral changes in PSP and could be included among potential outcome measures in future clinical trials in PSP.