Abstract
Abbreviations
Addenbrooke’s Cognitive Examination
Area under curve
Diagnostic and Statistical Manual of Mental Disorders 5th edition
Hoehn-Yahr Stage
Lille Apathy Rating Scale
Levodopa Equivalent Dosage
likelihood-ratio
Montgomery-Asberg Depression Rating Scale
Mattis Dementia Rating Scale
Movement Disorders Society
Movement Disorders Society Sponsored version of Unified Parkinson’s Disease Rating Scale
Motor Complications (Part 4 of MDS-UPDRS)
Motor Examination (Part 3 of MDS-UPDRS)
Motor Experiences of Daily Living (Part 2 of MDS-UPDRS)
Non-motor Experiences of Daily Living (Part 1 of MDS-UPDRS)
Mini Mental State Examination
Montreal Cognitive Assessment
Neurocognitive disorder
Parkinson’s disease
Receiver Operating Characteristic
standard deviation
INTRODUCTION
Parkinson’s disease (PD) is the second most common neurodegenerative disorder. In addition to the motor symptoms (bradykinesia, rigidity, tremor, and postural instability), PD is associated with numerous non-motor features [1]. Among the non-motor symptoms, neuropsychological and neurocognitive dysfunctions are the most problematic. From the early (even premotor) stages of the disease, depression, apathy and fatigue may be present [2, 3] and they may feature significant impact on health-related quality of life [4]. Neuropsychological problems may include a variable combination of executive dysfunction, visuo-spatial and constructional dysfunction, explicit and implicit memory deficits [5, 6]. Notably, in the progressive stages of the disease, neurocognitive disturbances may worsen [5] and the majority of patients may eventually develop mild cognitive impairment or Parkinson’s disease dementia [7–9].
Apathy, as a distinct entity, was described by Marin in 1991 [10]. With accordance to Marin, the primary feature of apathy syndrome is the lack of motivation which may be associated with other cognitive, affective and behavioral changes. The apathy syndrome can be described by the presence of following symptoms: Diminished interest in everyday activities Diminished social activities Diminished initiative with executive dysfunction Blunting of emotional responses Diminished responses to positive or negative emotional stimuli
Apathy is classified as a feature associated with depression by both the 4th and 5th revisions of Diagnostic and Statistical Manual of Mental Disorders (DSM-IV and DSM-5) [11]. However, recent studies demonstrate apathy may be present without evidence of depression or neurocognitive disorders [12] and therefore it is often considered as a distinct entity. Although both apathy and depression in PD may have synonymous features (e.g., indecisiveness, loss of interest), only depression is associated with dysphoria, or the feeling of sadness, guilt, pessimism, thoughts and the potential intention of suicide while harboring feelings of hopelessness [13].
However, the research on apathy has been hampered by a lack of a widely accepted diagnostic criteria. In 2008, a consensus meeting was held by the European Psychiatric Association, the European Alzheimer’s Disease Consortium and the Association Française de Psychiatrie Biologique, whose goal was to propose consensus diagnostic criteria for apathy as a syndrome [14]. These diagnostic criteria have subsequently been validated in 306 patients demonstrating good acceptability, inter-rater reliability and known-groups validity [15]. Subsequently, these diagnostic criteria for apathy syndrome have also been validated in PD [13].
Available apathy scales
The prevalence of apathy was reported between 17 and 70% in PD [16, 17]. The immense variance has not only contributed to the differences in the definitions of apathy but also to the heterogeneous scales applied.
As of 2015, there are three apathy scales which may be used in the PD population based on the recommendations of the Movement Disorders Society Task Force [16]: the apathy scale proposed by Starkstein in 1992 (Apathy Scale, AS) [18], the motivation/initiation item of the Unified Parkinson’s Disease Rating Scale (UPDRS) and the Lille Apathy Rating Scale (LARS) developed by Sockeel et al. in 2006 [19].
Although the AS is a recommended screening measure for apathy which can be administered to a wide spectrum of patients in a relatively short period of time, it does not provide information about the different clinical manifestations (i.e., cognitive, behavioral and affective features) of apathy.
The motivation/initiation item of UPDRS is a single object; therefore, it also provides limited information on the clinical symptoms of apathy. Although the Movement Disorders Society-based MDS-UPDRS was designed as the successor of UPDRS and published in 2008, its ‘Apathy’ item has not yet been validated against clinical criteria.
In contrast, the LARS is a semi-structured interview designed to screen and measure the severity of apathy with accordance to the definition of Marin [10]. It contains 33 questions, in which 30 responses are coded on a three point Likert-type scale (yes, unknown and no) and the additional 3 items with a five point Likert-type scale. The items are classified into 9 domains, including reduction of everyday productivity, lack of interest, lack of initiative, extinction of novelty seeking, lack of motivation, blunting of emotional responses, lack of concern, poor social life and decreased level of self-awareness. The scale determines an apathy score between –36 and +36, higher values signifying more severe apathy. Furthermore, there are four subscales (with the possible values between –4 and +4) measuring the four cardinal features of apathy: intellectual curiosity, action initiation, emotion and self-awareness.
The LARS has been translated into many languages and also validated by independent groups [20, 21]. It is considered proven, sensitive and capable of distinguishing between apathy and depression. The sensitivity, specificity, validity and reliability of the LARS was found to be high [19, 20]. Recently two additional formats of LARS have since been developed and validated, including a caregiver-based version [22] and a short-form [23]. Although LARS was also recommended by an independent review panel for the PD population specifically [24], it was also utilized inother neuropsychiatric conditions [25].
However, the specific cut-off scores for detecting apathy vary considerably among different populations and different study sites. The developers of the LARS concluded a global cut-off of –16 best differentiated apathetic and non-apathetic individuals withsensitivity and specificity of 0.89 and 0.92, respectively [19]. Sockeel et al. further proposed a system for classifying the severity of apathy based on LARS global scores, with scores between –36 and –21 suggesting the absence of apathy, between –21 and –16 suggesting mild apathy, between –16 and –9 indicating moderate apathy, and above –9 indicating severe apathy [19]. However, a recent study on Anglo-Saxon population revealed the cut-off value of -22 is more appropriate (sensitivity = 64% , specificity = 92% , positive predictive value = 88% , negative predictive value = 75%) [20]. On the Spanish version of LARS; however, the optimal cut-off point was significantly different (–14 in PD and –11 on healthy controls [21]).
Although LARS has been extensively utilized in PD for detecting the occurrence of apathy [4, 26] and testing the effectiveness of certain therapeutic interventions [12], the highly variable cut-off values substantiate the applicability of LARS as being more difficult yet a suitable screening tool. Hypothesizing the variability of threshold values cannot be fully attributable to the different cultural aspects among various nations, but may also be, at least partially, due to disregarding the neurocognitive and affective problems of the individuals, we initiated a cross-sectional study to determine the optimal cut-off values for PD. To achieve more reliable threshold values, different cut-off values were estimated based on the neuropsychiatric status of the patients (depressed vs. non-depressed and having or not having NCD). The second aim of our study was to formally validate the ‘Apathy’ item of MDS-UPDRS against the clinical criteria of apathy and determinate its usefulness in the process of screening.
MATERIALS AND METHODS
Patients
Notably, 584 PD patients examined and treated at the Department of Neurology, University of Pécs were enrolled into this study. Each patient fulfilled the clinical diagnostic criteria for idiopathic Parkinson’sdisease [27]. History of cerebrovascular disease, alcoholism or other conditions (systemic disease, endocrine disorders, tumors) capable of altering mental status served as exclusion criteria for participation. Each patient signed a written form of consent in accordance with the ethical approval of the Regional an Institutional Ethical Board of University of Pécs (4678/KK/2011).
Obtained rating scales
Global severity of PD-related symptoms was assessed by the Hungarian validated version of Movement Disorders Society version of UPDRS (MDS-UPDRS, the successor of UPDRS) [28] and the Hoehn & Yahr Scale (HYS) [29]. Neurocognitive status of patients was categorized by the Hungarian validated version of following instruments: Addenbrooke’s Cognitive Examination (ACE) [7], Mini Mental State Examination (MMSE) [7], Montreal Cognitive Assessment (MoCA version 7.2) [30], Mattis Dementia Rating Scale (MDRS) [7]. Presence of mild and major neurocognitive disorders (NCD) in PD was defined by the fulfillment of applicable DSM-5 diagnostic criteria [11] verified by the achieved scores on the above mentioned screening tests [30]. Presence and severity of depression was evaluated by using both the DSM-5 criteria for depression [11] and the Montgomery-Asberg Depression Rating Scale (MADRS, values >14 points) [7]. The presence of apathy was assessed by the definition for apathy syndrome proposed by Robert et al. [14] and validated by Drijgers and coworkers in PD [13]. Severity of apathy was graded by both LARS [19] and the ‘Apathy’ item of MDS-UPDRS (item 1.5) [28, 29].
Calculating threshold estimates
We applied sensitivity- and specificity-based methods to calculate the threshold allowing the optimum discrimination between groups of patients having and not having apathy. For example, the score producing the greatest sensitivity and specificity for discriminating patients with apathy from patients without apathy can be considered as the estimated threshold. Used in conjunction with threshold estimations, sensitivity is the proportion of the patients who have apathy based on the diagnostic criteria and whose LARS scores exceed the threshold value. Similarly, specificity is the proportion of subjects who do not have apathy and whose LARS scores are below the threshold value. In our study, we applied the receiver operating characteristic (ROC) curve technique to determine the most suitable threshold values. Assuming false- positive and false-negative identifications are equally undesirable, we determined the cut-off value with the most optimal balance between sensitivity and specificity estimated as the point on the ROC curve closest to the point of (0,1). It was calculated as the minimum value of the square root of ((1- sensitivity)2+(1-secificity)2)[31, 32]. For the most optimal cut-off values the positive (LR+) and negative (LR–) likelihood-ratios were also determined.
Statistical analysis
All statistical analyses were implemented using the IBM SPSS software package (version 22, IBM Inc., Armonk, NY). Interestingly, most of the data did not follow the normal distribution, and so, Mann-Whitney and Chi-square tests were applied. Multiple regression analysis was performed to detect the relationship between LARS total score (dependent variable) and other demographic, PD-related and neurocognitive variables (independent variables). Statistical significance level was set to 5% . Due to the fact, the SPSS Suite did not yet feature built-in functions for calculating positive and negative likelihood ratios, we utilized the syntax available on the IBM website (http://www-01.ibm.com/support/docview.wss?uid=swg21483380, assessed on Jan 15, 2013).
RESULTS
Sociodemographic and disease-specific data of the study population is presented in Table 1. Based on the proposed definitions of apathy, 477 patients did not have apathy (81.7%) and 107 had clinically pertinent apathy (18.3%). The sample contained 310 patients with NCD (53.1%) and 217 patients with depression (37.2%).
Multiple regression analysis
A multiple regression was initiated to predict the total score of LARS from various clinical variables including gender, age, age at disease onset, years of education, disease duration, handedness, HYS, MDS- UPDRS Parts I-IV, ACE, MDRS, MADRS, levodopa dosage (measured in levodopa equivalent dosage, LED), DA LED, and total LED. Due to collinearity, age of disease-onset, MDRS, MMSE scores were excluded from the regression analysis (tolerance <0.001). The data met the assumption of independent errors (Durbin-Watson value = 1.911). Using a stepwise method (Criteria: Probability of F to enter < = 0.050, Probability of F to remove > = 0.100) it was found that MADRS total score (βstandardized = 0.588, p < 0.001), ACE total score (βstandardized = –0.259, p < 0.001), gender (coded as 1 = males and 2 = females, βstandardized = –0.129, p = 0.001) and DA dosage in LED (βstandardized = –0.091, p = 0.019) explain the highest significant amount of variance in the value of the LARS score, F(4,331) = 89.371 (p < 0.001), R2 = 0.519, R2adjusted = 0.513. The other examined variables did not significantly contribute to the model. Based on the multiple regression model, the following equation could be formed:
Determining threshold values for LARS
Although the age, gender-distribution, disease duration, disease type and the total dosage of anti-Parkinsonian medication were comparable between the apathetic and non-apathetic patients, patients having apathy had significantly worse PD-related symptoms (measured by both MDS-UPDRS and HYS), worse neurocognitive performance and more severe depressive symptoms (Table 1). As expected, both the subscores and the total score of the LARS were significantly decreased in apathetic patients (Table 2).
We found a significant, positive correlation between the LARS score and the MDS-UPDRS item 1.5‘Apathy’ (Spearman’s rho = 0.516, p < 0.001, Fig. 1).
Subsequently, we performed a ROC analysis to assess the ideal cut-off values in discriminating apathetic patients from non-apathetic patients. Due to the multiple linear regression, analysis revealed both the neurocognitive status (measured by ACE) and the severity of depression (measured by MADRS) significantly contributed to apathy, and therefore, we included these two co-factors in determining the threshold values. Notably, different threshold values were calculated for patients with (1) both NCD and depression, (2) with NCD, (3) with depression, and (4) without NCD or depression. The most optimal thresholds and their associated accuracy values are shown in Table 3.
Apathy (1.5) item of MDS-UPDRS
We identified a moderate, yet significant correlation between the total score of the LARS and the ‘Apathy’ item of MDS-UPDRS (rho = 0.516, p < 0.001, Fig. 1). The best discriminating value for screening apathy was 0.5 in all conditions (specificity: 0.578–0.801, sensitivity: 0.425–0.514, +LR: 2.075–2.58 and –LR: 0.607–0.690).
DISCUSSION
In the present study we evaluated the usefulness of two distinct clinical tools for screening apathy syndrome among Parkinson’s disease patients. The presence of apathy was assessed by the definition for apathy syndrome proposed by Robert et al. [14] and validated by Drijgers and coworkers in PD [13].
Several studies demonstrated a strong correlation between affective status and apathy in PD, especially with executive dysfunction. Apathy may cause cognitive inertia, which restricts the incentive and pursuit of enjoying both former and new interests. This is based on executive networks dysfunction: difficulties to redirect attention to novel stimuli, manipulating complex information or generating plans for the future [38]. Furthermore, apathy may serve as a predictive factor for cognitive decline and Parkinson Dementia [33].
Depression in PD may differ from other forms unassociated with PD. Depressed PD patients experience higher frequency of dysphoria, irritability, sadness, pessimism about the future and suicidal idealization. On the other hand, there is less noticeable frequency of guilt, self-blame, feelings of failure and suicide (despite the higher frequency of idealization). These symptoms may shroud symptoms of apathy, due to a large overlap between apathy and depression, both featuring a series of common symptoms [34].
Although the neurobiology of apathy is not fully evaluated, recent evidence suggests that depression and apathy syndrome are distinct in PD. First, clinical studies demonstrated that apathy was present without evidence of depression or neurocognitive disorders [12] and therefore it was considered as a distinct entity. In addition, some clinimetric data also highlighted the distinct features between apathy and depression. Using confirmatory factor analysis on the Beck Depression Inventory and Apathy Scale, separate factors can be isolated for depression and anxiety: (1) symptoms of “pure” apathy, (2) symptoms of “pure” depression, (3) overlapping symptoms associated with loss of interest or pleasure and (4) somatic symptoms. Therefore, features such as indifference, motivation, lack of concerns and blunted emotions are highly characteristic for apathy and are thereby distinct from depression [35].
Although several studies suggested LARS may be an excellent tool in the screening and grading apathy in PD, the highly variable cut-off scores currently published make its use limited. Based on the recently published research on apathy, we assumed the heterogeneity of threshold values may be partially due to disregarding the neurocognitive and affective status of patients.
In the multivariate regression analysis, we strived to identify the most prominent factors contributing to apathy measured by LARS, we revealed not only the neurocognitive status (measured by ACE), the severity of depression (measured by MADRS), but also the gender and the applied dosage of dopamine agonists (measured in LED) determined the degree of apathy. Being male, having more severe depression (higher MADRS scores), having a more severe neurocognitive disorder (lower ACE scores) and receiving none or lesser amounts of dopamine agonists (in LED mg) are associated with higher LARS scores and consequently more severe apathetic symptoms.
Previous studies revealed the role of depression and neurocognitive impairment in the development of apathy [18, 36], yet the importance of gender and dopamine agonist treatment was less known [4]. Dopamine agonists are known to have antidepressive effects and in some cases can provoke manic symptoms due to their potential mesolimbic effects [37, 38]. However, their use is limited in the elderly or patients afflicted with dementia due to the potential hallucinations and irritability. Therefore, it is unknown whether their direct anti-apathetic effects or their limited use by age or neurocognitive status contribute to the degree of apathy in PD.
Although the age, gender and disease-duration were comparable in the apathetic and non-apathetic groups, symptoms of PD assessed in both MDS-UPDRS and HYS were most severe in the apathetic group. However, the multiple regression analysis did not identify the severity of PD as a predictor of LARS score. We hypothesize, the discrepancy in distribution of major neurocognitive disorders support the likelihood of this apparent contradiction. Whereas 42 patients had major NCD in the non-apathetic group (8.8%), in the apathetic group 68 had (63.6% , p < 0.001). Due to the poor neurocognitive status often associated in more severe PD symptoms in general, in particular, our case in which the MDS-UPDRS scores and the HYS did not prove to be an independent predictor of apathy (LARS score).
In the second part of the study, we calculated the most optimal threshold values detecting the presence of apathy (Table 3). Based on the presence of depression and NCD, we established four different threshold values. These cut-off values were in the range of previously determined Anglo-Saxon thresholds (between - 16 and –22 points) [19–21]. The specificity of the test is high, sensitivity is moderate, of which agrees with the results of Zahodne and colleagues [20].
Similar to the ‘Motivation’ item of UPDRS, the ‘Apathy’ item of MDS-UPDRS may serve as an easy- to-assess, but suitable screening tool in the detection of apathy among PD patients. It is not a surprising result, because the Non-motor Experiences of Daily Living items of MDS-UPDRS were designed as a good screening questionnaire aimed at identifying areas of disability. The scale itself registers presence and severity of items but has an official APPENDIX with recommended more specific scales to pursue areas in more detail [29, 39]. Despite of its disadvantage being a single item tool, we propose to screen apathy in the utilization of MDS-UPDRS. In general clinical practice it may be an efficient way towards evaluating a large pool of patients on a regular basis. However, its clinical use may be limited because it cannot evaluate the full spectrum of the apathy syndrome.
In cases of positive identification of apathy (having >0 score on the ‘Apathy’ item of MDS-UPDRS), we recommend the utilization of LARS with extensive neuropsychological testing form depression and neurocognitive disorders.
Our study clearly demonstrates a single universal threshold value for the LARS should not be utilized in screening for apathy due to its often and assumed association to depression and neurocognitive disorders. In general, we suggest the application of the threshold valid for normal cognition and normal affective status (–22.5 points) for screening. In cases of subjects exhibiting a LARS score higher than –22.5 points, neuropsychological examination is warranted towards evaluating the presence of depression and neurocognitive disorder. Subsequently, the presence of apathy should be determined with the appropriate cut-off value specific for the neuropsychiatric status of thepatient.
Limitations and strengths of the study
Although we endeavored to strategize this study with precision, the authors are aware of the potential limitations. In the role of a simultaneously primary and a tertiary center, the University of Pécs has both non-complicated PD patients from the surrounding areas and the advanced PD patients from the nationwide primary and secondary centers. Therefore, the pool of patients are likely different from those of typical primary or secondary centers.
However, in the role of a tertiary center, consider our advantage: Relatively a large portion of severe (HYS 4&5) patients were also included in our study. This may have contributed to a relatively high percentage of the population featuring advanced PD and or minor/major neurocognitive disorders.
CONCLUSIONS
In this study we revealed not only the neurocognitive status and the severity of depression, but also the gender and the applied dosage of dopamine agonists determined the degree of apathy. Based on the presence of depression and neurocognitive disorder, we determined four different threshold values; therefore, there is no single and universal threshold value yet capable of identifying apathy under all circumstances.
Footnotes
ACKNOWLEDGMENTS
Our study was supported by the Bolyai Scholarship of Hungarian Academy of Sciences, OTKA PD103964 and Hungarian Brain Research Program (KTIA_13_NAP-A-II/10) government-based funds. We would thank the help of Éva Balázs and Katalin Takács, our Parkinson’s disease nurses.
The present scientific contribution is also dedicated to the 650th anniversary of the foundation of theUniversity of Pécs, Hungary.