Because of marked motor and autonomic abnormalities, the cognitive, behavioral, and neuropsychiatric symptoms (NPSs) in MSA were not the main clinical focus for many years. MSA was initially assessed in small series of patients and mainly identified by bedside impression. However, with the introduction of validated questionnaires and tests, it became clear that NPSs such as depression, anxiety, and apathy commonly coexisted within the MSA phenotype (
3), with marked effects on quality of life and disability (
4,
5). Moreover, the presence of depression in MSA appears to be independent of the severity of motor impairment, suggesting that it may be a part of the clinical spectrum of the disease rather than secondary to the severity of motor dysfunction (
6).
Therefore, the aims of this study were to determine the pattern of NPSs in our cohort of patients with MSA-P; to longitudinally follow up these patients over a 1-year period; to establish the possible relationship of the pattern of NPSs with motor, autonomic, and cognitive symptoms; and to compare rates of NPSs reported by caregivers and those reported by the patients themselves.
Results
Neuropsychiatric Inventory Scores
In a cross-sectional analysis of 47 MSA-P patients (
Table 1), at least one NPS was present (NPI score >0) in 94% of investigated patients (
Table 2). The most prevalent symptoms were depression, changes in sleep/nighttime behaviors, apathy, and anxiety (92%, 85%, 78%, and 76%, respectively). In order to avoid clinically irrelevant NPI scores, a separate analysis involved only those patients with a score ≥4 on at least one of the NPI items; again, depression and sleep disturbances were the most common, occurring in approximately two-thirds of patients, while apathy and anxiety rates decreased to 43% and 53%, respectively. The average severity score was also highest in the same NPI domains.
The individual symptom in the agitation subsyndrome with the highest prevalence rate was irritability (45%), followed by agitation (25%) and disinhibition and aberrant motor behavior (9% each). Hallucinations were recorded for 21% of patients, but only 2% had a clinically relevant score. Other psychotic symptoms were infrequent. Although the change of appetite/eating domain was not as commonly reported, when it occurred, the severity of such symptoms was scored as clinically relevant.
Severity and Frequency of NPSs in Terms of Age, Disease Duration, Motor Severity, and Frontal and Global Cognition Status
Results of the severity and frequency of NPSs are summarized in
Table 3. Disinhibition behavior was observed only in elderly patients (≥60 years old), but only three out of 15 patients had clinically relevant symptoms. Apathy and changes in appetite/eating were more profound in patients with a longer disease duration (>3 years). Patients with more severe motor symptoms expressed more agitation, depression, and apathy and generally showed higher NPI total scores and higher NPI distress scores. In terms of frontal and global cognition status, no differences were found in neuropsychiatric profile, except that patients with global cognitive decline (DRS score ≤125) were found to have fewer anxiety symptoms (
Table 3).
Disease-Related Factors Associated With NPI Total Score
In multiple regression analysis (
Table 4), the baseline independent variables that we identified as significant in predicting increasing NPI total scores were the MMSE score and the UMSARS part IV score (
Table 4). No association was found between NPSs on one hand and autonomic and cognitive scales measuring global and frontal cognitive decline on the other.
Measuring Depression, Apathy, and Anxiety by NPI Compared With Psychiatric Tools Based on Patient Assessment
The overall proportions for symptoms of depression, apathy, and anxiety based on caregiver reports (NPI) on the one hand and patients’ reports on other hand (HAM-D, BDI-II, HAM-A, AS) differed significantly.
Twenty-eight patients had clinically relevant NPI depression scores, and all of these were also assigned as having depression based on their HAM-D cut-off score. Seven patients were classified as nondepressed as assessed with the NPI, but they had verified mild depression as assessed with the HAM-D (prevalence based on the NPI versus prevalence based on the HAM-D: Fisher’s exact test, p<0.001). The BDI-II registered depression in 26 out of 28 patients, while six patients who had clinically irrelevant scores on the NPI were considered to have mild depression on the BDI-II (prevalence based on the NPI versus prevalence based on the BDI-II: Fisher’s exact test, p<0.001).
The NPI apathy score was marked as clinically relevant in 20 patients, while the AS confirmed this finding in only 16 of these patients. Eleven patients who had no apathy based on the NPI had scores on the AS above the proposed cut-off (prevalence based on the NPI versus prevalence based on the AS: Fisher’s exact test, p=0.009).
Twenty-five patients who had a clinically relevant NPI anxiety score were also identified as having anxiety based on the HAM-A cut-off score. Three patients were identified as having a mild degree of anxiety on the HAM-A but did not reach a clinically relevant score on the NPI anxiety domain (prevalence based on the NPI versus prevalence based on the HAM-A; Fisher’s exact test, p<0.001).
Evolution of Clinical Variables Over 1-Year Follow-Up
The baseline characteristics of 25 patients with MSA-P who were subsequently followed up at 1 year are presented in
Table 1. We found no differences in age, age at onset, sex, education, disease duration, or severity of illness as assessed using the Hoehn and Yahr stage and UMSARS parts I, II, and IV between patients who completed the follow-up (N=25) and those who dropped out of the study (N=22) (t test, Mann-Whitney U test, chi-square test, p>0.05).
Estimation of changes in clinical, motor, cognitive, and neuropsychiatric variables in 25 MSA-P patients followed for 1 year (
Table 5) revealed the large effect size (effect size >0.80) in motor (the UMSARS subscales, UPDRS total score, and Hoehn and Yahr stage), autonomic (SCOPA-AUT), and cognitive (DRS) domains. The NPI total score did not show significant change over the 1-year follow up. Additionally, none of the 12 NPI composite scores changed (Wilcoxon signed-rank test, p>0.05).
None of the additional neuropsychiatric measures (HAM-A, HAM-D, BDI-II, AS) showed significant changes in their mean scores over the follow-up period (
Table 5). In addition, the overall prevalence of patients who were assigned as having depression, apathy, or anxiety based on cut-off scores on the HAM-D, BDI-II, HAM-A, and AS were not changed from baseline in the follow-up visit (chi-square test or Fisher’s exact test, p>0.05).
In the final model of regression analysis applied to longitudinal data, after elimination of the nonsignificant variables during the backward stepwise process, the only independent predictor of change of the NPI total score over the 1-year follow-up period was disease duration (unstandardized beta coefficient=2.688, SE=1.116, p=0.025).
Discussion
The main findings of this study were that the prevalence of NPSs in patients with MSA-P was very high (at least some level of NPS expression was present in 94% of patients), with depression, sleep disturbances, apathy, and anxiety being the most frequently occurring features; that the evolution of NPSs was found to be independent of motor, autonomic, and cognitive symptoms; that none of the scales measuring NPSs, including the NPI, were capable of detecting changes over the 1-year follow-up period; and that although the overall prevalence of depression, apathy, and anxiety obtained from caregivers and the patients themselves was similar, reports from these two sources cannot be considered interchangeable.
The general pattern observed by the NPI in our cohort of patients with MSA-P was comparable to that of previous studies, which also underscored depression as the key and most consistent neuropsychiatric feature of MSA (
6,
24). However, direct comparison is difficult, because these studies used different methodological approaches, including different instruments for the assessment of NPSs.
The prevalence of a clinically relevant rate of depression of approximately 60% as measured with the NPI was in accordance with the data reported in previous studies with larger numbers of MSA patients (
5,
24–
29). The proportion of patients with depression as measured by the HAM-D and BDI-II was higher (74% and 68%, respectively) compared with the proportion reported through a caregiver assessment. MSA-P patients themselves appeared to report more NPSs than their caregivers, which is consistent with a previous finding in a study of patients with Parkinson’s disease (
30).
However, we found a higher rate of anxiety in caregivers’ reports (53% of patients had a clinically relevant score on the NPI anxiety scale) compared with patient reports (47% of patients had scores above the cut-off on the HAM-A). These estimates were higher than those reported in previous studies (
25,
27,
29,
31), while the rate of apathy was similar to that reported in previous studies (
24,
29,
32). Interestingly, apathy was a symptom primarily observed by caregivers (the rate of apathy was remarkably different between caregiver and patient reports; 57% and 42%, respectively), suggesting that poor insight into the symptoms may be an integral part of apathy expression (
33).
Similar to previous studies, changes in personality with a clinically relevant irritability score were observed in 17% of patients in our study (
25,
29). Hallucinations were reported for 21% of patients in the cohort, but only 2% of patients with MSA-P had clinically relevant scores. This is consistent with including hallucinations on the list of nonsupportive features as defined by current diagnostic criteria (
2). Agitation, disinhibition, and aberrant motor behavior were not frequently occurring and if present did not reach the clinically relevant score, as observed in another study that used the NPI as a diagnostic tool (
29).
Another important issue was our investigation of the possible association between NPSs and motor, autonomic, and cognitive symptoms in MSA-P. In our results, the severity of motor impairment, assessed with the UMSARS part IV, was found to be the strongest independent predictor of the higher NPI total score among patients with MSA-P. For each increase of one point on the UMSARS part IV, which actually reflects the transition to the next stage of global disability, there is an increase of 6.644 in the NPI total score. Similarly, Cao et al. (
22) reported that higher scores on the Frontal Behavioral Inventory were associated with increased severity of disease, as assessed with the UMSARS. Patients with a higher burden of motor symptoms and longer disease duration expressed more severe depression and apathy. In addition, lower scores on the MMSE, which is considered a rough screening tool for cognitive decline, have also been identified as an independent predictor of an increase in the overall NPI score. However, the association of NPI scores with scores on the DRS and FAB, which are detailed and specific scales measuring global and frontal cognitive status, has not been confirmed. Therefore, based on these facts, we still cannot argue about the potential association of cognitive decline with the higher rate and severity of NPSs in MSA.
Although significant deteriorations of motor, autonomic, and cognitive symptoms in our longitudinal cohort were captured by the appropriate scales over the 1-year follow-up period, this was not the case with NPSs. None of the scales measuring NPSs, including the NPI, were able to detect changes in MSA-P patients over the follow-up period. Furthermore, the overall prevalence of depression, apathy, and anxiety did not change significantly over the 1-year follow-up. Thus, one may conclude that the evolution of NPSs does not follow the progression of motor, autonomic, and cognitive problems in MSA-P, suggesting that depression and other NPSs are an integral part of the clinical spectrum of the disease rather than secondary due to the severity of motor dysfunction (
6). Interestingly, behavioral and mood disturbances in neurodegenerative disorders do not necessarily progress uniformly over time (
34). For example, in MSA, as well as in Parkinson’s disease and progressive supranuclear palsy, apathy and depression showed a different pattern of deterioration over time: apathy showed linear worsening over time, while depression seemed to remain consistent or followed a fluctuating course (
24,
35,
36). Therefore, heterogeneity of presentation and inconsistency in the progression of NPSs may represent a particular challenge in longitudinal studies, including the selection of appropriate scales.
In addition to the relatively small number of pathologically unproven MSA-P patients, among the limitations of our study was the fact that the most common NPSs, such as depression, apathy, and anxiety, were not diagnosed according to clinically validated criteria. Furthermore, the dropout rate of more than 47% in our longitudinal study, together with the constraints of a small sample size, may have biased our results and affected the analytic approach. However, no differences between the patients who were followed and those who dropped out were found in the baseline demographic variables and severity of motor symptoms, which made us more confident in drawing conclusions.