Mood Differences among Parkinson’s Disease Patients with Mild Cognitive Impairment

Abstract

Objective

Studies with healthy elderly adults suggest apathy, depression and anxiety are more common among individuals with mild cognitive impairment (MCI). We examined differences in mood/amotivational symptoms among Parkinson’s patients with and without MCI.

Methods

Parkinson patients (N=214) underwent neurocognitive evaluations including assessment of apathy (Apathy Scale; AS), depression (Beck Depression Inventory-II; BDI-II) and trait anxiety (State-Trait Anxiety Inventory-Trait scale; STAI-T).

Results

Trait anxiety and depression were more severe in PDs with MCI. Delineation of MCI into amnestic vs non-amnestic subtypes revealed greater depression, apathy and anxiety among amnestic MCI relative to cognitively-intact PD patients.

Conclusion

Parkinson patients with MCI report greater mood symptoms compared to Parkinson patients who are cognitively intact.

Introduction

Parkinson’s disease (PD) is a neurodegenerative disorder characterized by motor impairments, as well as mood and cognitive disturbances. The wide-range of symptoms is due in part to dopamine depletion in the substantia nigra and disruption of subcortical-thalamo-cortical circuits {¹}.

Cognitive impairments in PD primarily consist of slowed processing speed, multi-tasking/working memory difficulties and increased forgetfulness; however the pattern of cognitive impairment, as well as the course/trajectory of cognitive decline, can vary across patients {²}. For almost a decade, the concept of identifying individuals with mild cognitive impairment (MCI), who are at-risk for future dementia, has been investigated as a possible meaningful concept among PD {³}. Indeed, a MCI stage has been viewed as an “intermediate” stage between cognitively intact PD patients and those with Parkinson’s disease dementia (PDD) {^2,4}. Amnestic and non-amnestic MCI subtypes may also be useful in learning whether certain cognitive profiles represent increased risk for PDD. A non-amnestic, executive dysfunction profile has been primarily associated with frontal-striatal dysfunction involving the dorsolateral prefrontal cortex (DLPFC), while an amnestic profile may reflect a “posterior” profile with additional dysfunction in cholinergic systems and/or the beginning stages of co-morbid Alzheimer’s pathology {^5–6}. Evidence has been mixed regarding which MCI-subtype is indicative of future PDD {^4,6,7}.

The clinical and research utility of MCI relates to its potential to identify individuals at risk for dementia, ideally leading to early interventions that can delay the onset of dementia. However, MCI status has also been shown to relate to other meaningful outcomes. Within PD, MCI status has been linked to worse motor symptoms and stage severity, postural instability, and depression {^4,8}. In normal elderly, not only is depression linked to MCI status, but also apathy and anxiety {⁹}. Identification of mood symptoms in cognitively impaired individuals is important because mood symptoms are strong predictors of quality of life that may be responsive to interventions {^9,10}. Furthermore, due to the negative influence of mood symptoms on cognition, treatments targeting depression, apathy and anxiety may lead to better cognitive functioning {^11,12}.

Apathy (disorder of motivation/goal-directed behavior), depression, and anxiety are common, but dissociable, mood symptoms in PD {¹³}. The high occurrence of mood symptoms in PD may be partly due to the reaction of being diagnosed with a serious medical condition; however there is a direct relationship between PD pathology and emotional dysfunction (particularly apathy). Striatal dopamine depletion leads to a disruption of frontal-subcortical circuits. These circuits include cortical prefrontal regions important for emotional processing, such as the anterior cingulate cortex (ACC) and orbital-frontal cortex (OFC) {¹}. In addition to dopamine depletion, high rates of mood symptoms might also be linked to dysregulation of separate neurotransmitters such as serotonin, norepinephrine and possibly acetylcholine {¹⁴}.

Both mood and cognitive symptoms are important predictors of patient-centered outcomes, however the relationship between mood disturbances and MCI status in PD has received little attention {^11,15}. The aim of the current study was to examine, in Parkinson patients, the relationship between MCI and common mood and motivation dysfunctions. The overall hypothesis is that due similar neural circuitry (frontal-striatal circuits) underlying both emotional and cognitive processes, mood symptoms will be greater among Parkinson patients with MCI, with apathy being particularly more severe among PD patients with MCI.

Methods

Design & Participants

This cross-sectional design included a convenience sample of 214 patients with idiopathic PD according to UK Brain Bank criteria. The Institutional Review Board at the University of Florida (UF) approved the study and consent was attained prior to patient participation. All patients underwent a detailed neuropsychological assessment, including completion of mood questionnaires as part of their routine clinical care through the UF Center for Movement Disorders and Neurorestoration. Patients were excluded based on the following criteria: 1) previous brain surgery such as deep brain stimulation, 2) severe psychiatric disturbance (e.g. schizophrenia), or 3) severe cognitive impairment defined as a score on the Dementia Rating Scale-II (DRS-II) below the 5^th percentile.

Mood and Clinical Measures

Patients completed standard questionnaires assessing apathy, depression and trait anxiety. Apathy was measured with the Starkstein Apathy Scale (AS) {¹⁶}. The Beck Depression Inventory-II (BDI-II) measured depressive symptoms {¹⁷}. The Trait scale on the State-Trait Anxiety Inventory assessed long-standing anxiety {¹⁸}. For all measures, higher scores represent higher levels of severity for their respective construct. Data pertaining to demographics (age and education) and PD-related severity (duration of motor symptoms and ratings of motor symptoms) were collected by a movement disorder neurologist or trained fellow. Motor symptom severity was determined by the Unified Parkinson Disease Rating Scale Part III (UPDRS motor score) and was taken while patients were “on” medication {¹⁹}.

Mild Cognitive Impairment Classifications

All patients completed neuropsychological measures of executive functioning, delayed verbal memory, attention/working memory, processing speed, language and visuospatial abilities as part of a neuropsychological examination. The attention/working memory tests consisted of the Forward Span and Backward Span scores of the Digit Span subtest from the Wechsler Adult Intelligence Scale-III. Verbal memory measures included the 20 minute delay recall score from the Hopkins Verbal Learning Test-R and the 30 minute delayed recall of the Logical Memory Stories-II from the Wechsler Memory Scale-III. The Boston Naming Test (total correct without cues) and an Animal Fluency test comprised the language domain. Tests of executive functioning included the Trail Making Test-Part B, the color-word interference trial of the Stroop Color-Word Test, and the letter fluency test of the Controlled Oral Word Association test. Visuospatial tests included the Benton Facial Recognition Test and the Judgment of Line Orientation test. Lastly, the processing speed domain included the Trail Making Test-Part A, and the Word reading trial of the Stroop Color-Word task. All tests were normed for age, education and gender based on test-specific manuals or previously published norms and then converted into a Z-score metric.

Mild cognitive impairment was defined according to the Movement Disorder Society (MDS) criteria. The MDS Level II criteria were used as it is appropriate for assessments that include at least two tests within each cognitive domain: attention, executive functioning, processing speed, language, memory and visuospatial {¹²}. Individuals were classified as MCI if they performed 1.5 standard deviations (SD) below the mean on at least two tests. Of note, tests can be within the same domain or in separate domains. Individuals with MCI were further classified into an Amnestic or non-amnestic MCI subtype. Individuals were classified into the amnestic group if either of the memory tests (HVLT-R delayed recall or Stories delayed recall) scored less than 1.5 SD below the normed average (i.e. a Z score less than −1.5).

Statistical Analyses

A series of independent t-tests were computed to assess group differences. Initial analyses separated groups dichotomously (i.e. cognitively intact vs. MCI) and examined differences in apathy, anxiety and depression. Subsequent analyses separated individuals into three groups: cognitively intact, amnestic-MCI and non-amnestic MCI. Indicators of normality were appropriate for mood variables (skewness and kurtosis values < 1).

Multiple hierarchical regressions were additionally computed in order to covary for potential confounder variables. Mood variables (depression, apathy and anxiety) were entered as the dependent variable for each regression. Predictor variables included MCI status, age, education and motor symptoms severity (on medication UPDRS motor subscale). Mild cognitive impairment status was forced entered into the model and additional covariates (age, education, motor symptom severity) were entered in step-wise fashion so only predictors that significantly contributed to the model were included. Additional regression analyses were repeated to examine MCI subtypes (i.e. amnestic versus non-amnestic). Dummy coded MCI status variables were forced entered into each model with additional covariates entered in stepwise fashion.

Results

Sample Characteristics

Table 1 shows the sample characteristics of the 214 Parkinson patients. Briefly, the sample had a mean age of 64 years, was primarily Caucasian (95.8%) and male (70.6%). Tremor was the dominant symptom in 80.8% of patients, and average duration of symptoms was 10 years. Average UPDRS motor scores while patients were “on medication” was 26.1. Compared to cognitively intact PD patients, those with MCI had longer duration of motor symptoms (t (212) = 2.27, p = .024) and worse UPDRS motor scores (t (212) = 4.70, p < .001). The two groups did not differ in terms of age, education or levodopa equivalency dose (LED; all p values > .05).

Table 1.

Sample Characteristics

N=214 Parkinson Patients
151 Males (70.6%); 63 Females (29.4%)	Total	Intact Cognition	MCI
		N=113	N=101
	Mean (SD)	Mean (SD)	Mean (SD)
Percent Caucasian	95.8	98.2	93.1
Age in years	64.3 (9.5)	63.7 (9.5)	65.1 (9.5)
Years of Education	14.9 (2.9)	15.0 (2.8)	14.9 (3.0)
Years with Symptoms	10.0 (5.8)	9.1 (4.6)	10.9 (6.6)
UPDRS Motor Score, on medication	26.1 (10.8)	23.0 (9.5)	29.3 (9.6)
Percent Tremor Predominant	80.8	80.5	81.2
Levodopa Equivalency Dose	829.1 (564.3)	816.4 (590.0)	843.3 (536.7)
Percent on Antidepressant Medication	38.6	37.8	39.6
Percent on Anti-anxiety Medication	19.7	21.6	17.5
Dementia Rating Scale-Total	137.25 (4.9)	139.3 (3.4)	135.0 (5.2)
Beck Depression Scale-II	9.4 (6.3)	8.5 (5.9)	10.4 (6.5)
Apathy Scale	10.7 (6.3)	9.9 (5.9)	11.6 (6.5)
State-Trait Anxiety Inventory-Trait Scale	35.4 (9.8)	34.0 (9.6)	37.0 (9.8)

UPDRS= Unified Parkinson’s Disease Rating Scale-Part III; SD= standard deviation; MCI= mild cognitive impairment

Almost half (47.2%) of our sample met criteria for MCI. Of the individuals with MCI, 59% were categorized as amnestic MCI and 41% met criteria for non-amnestic MCI. The amnestic and non-amnestic groups did not differ in terms of age, education, UPDRS motor scores, duration of motor symptoms or LED (all p values > .05).

Mood Differences among Parkinson Patients with and without MCI

Independent samples t-tests were conducted to examine differences in mood scales between MCI and cognitively intact PD patients. Results revealed that PD patients with MCI reported more severe depression (t(212)=2.23, p=0.027) and trait anxiety (t(212)=2.27, p=0.024) compared to cognitively intact PD patients. Apathy was not significantly greater in MCI than non-MCI patients (p=.053).

In order to control for possible confounding variables, three follow-up hierarchical regressions were conducted controlling for age, education and severity of motor symptoms. Results revealed that the overall regression models significantly predicted depression (F(2,211) = 12.80, p = <.001, r²=.08), apathy (F(2,211) = 5.60, p = .004, r²=.05) and trait anxiety scores (F(2,211) = 5.46, p = .020, r²=.02). Depressive symptoms were more severe among PD patients with MCI (β = .134, p = .016) and individuals with less education (β = −.237, p < .001). Apathy symptoms were more severe among PD patients with MCI (β = .129, p = .048) and those with less education (β = −.172, p = .009). Trait anxiety symptoms were more common among PD patients with MCI (β = .134, p = .020), and were not related to age, education or motor symptom severity.

Mood Differences Among MCI Subtypes

Additional analyses examined potential mood differences among amnestic, non-amnestic, and cognitively normal subtypes (Figure 1). Results revealed the amnestic MCI group reported significantly more severe depression, apathy and trait anxiety compared to the cognitively-intact group. The non-amnestic group did not significantly differ from either the cognitively-intact nor the amnestic MCI group in any mood variable (all p values >.05).

Figure 1.

Raw means scores on measures of apathy, depression and trait anxiety are depicted for each group with higher scores representing greater severity; MCI= mild cognitive impairment; error bars depict the 95% confidence interval; * p≤.05.

In order to control for possible confounding variables, follow-up hierarchical regressions were conducted controlling for age, education and severity of motor symptoms. The overall models significantly predicted depression (F(2,211) = 9.36, p = <.001, r²=.09), apathy (F(2,211) = 4. 07, p = .008, r²=.04) and trait anxiety (F(2,211) = 3.03, p = .050, r²=.02). Examination of individual predictors revealed that individuals with amnestic MCI (relative to controls; β = .165, p = .009) and less education (β = −.229, p < .001) had more severe symptoms of depression. Trait anxiety symptoms were more severe among PD patients with amnestic MCI (β = .145, p = .016) and were not significantly related to other variables. Apathy symptoms were more severe in the amnestic MCI group (β = .152, p = .027) and patients with less education (β = −.167, p = .011). The non-amnestic MCI group did not significantly differ than controls on any mood measure (apathy, depression or trait anxiety)

Correlations Between Mood and Neuropsychological Tests

Partial correlations were conducted to examine the relationship between mood (depression, apathy and trait anxiety) and performance on neuropsychological measures. This was carried out on the entire sample controlling for age and education (Table 2). To minimize the number of comparisons and reduce family-wise error rate, we computed composite scores for each neurocognitive domain: attention, verbal memory, language, visuospatial, executive function, processing speed. We did this by converting all the normed scores obtained for each test into Z-scores, and then averaging the Z scores of all tests within a single domain. Indicators of normality were appropriate for mood variables and neuropsychological composite scores (skewness and kurtosis values < 1). Partial correlations revealed that higher scores on the Apathy Scale were related to worse language scores. Higher trait anxiety scores were related to worse executive function, verbal memory, and language performance. Scores on the BDI-II were not significantly related to any neurocognitive domain.

Table 2.

Partial Correlations Between Mood Symptoms and Neurocognitive Domains

		Executive Function	Verbal Memory	Attention	Language	Processing Speed	Visuospatial
Depression	Partial Correlation	−.073	−.120	−.034	−.136	−.091	−.053
	p	.292	.083	.206	.051	.192	.448
Apathy	Partial Correlation	−.123	−.130	−.002	−.167	−.125	−.082
	p	.076	.062	.977	.016	.072	.241
Trait Anxiety	Partial Correlation	−.179	−.211	−.053	−.231	−.123	−.112
	p	.009	.002	.444	.001	.076	.107

Controlling for education and gender; significant correlations appear in bold.

Discussion

In this study we found that approximately 47% of our PD sample met criteria for MCI. Those with MCI had more severe symptoms of depression and trait anxiety relative to cognitively intact PD patients. The relationship between mood symptoms and MCI was primarily driven by the amnestic MCI group and was not accounted for by age, education or motor symptoms severity. In brief, symptoms of apathy, anxiety, and depression were greater in the amnestic MCI group relative to cognitively normal PD patients.

These findings suggest that Parkinson patients with MCI may be experiencing greater mood symptoms than cognitively normal PD patients. Identifying mood symptoms among individuals with cognitive impairment is important for at least two reasons. First, self-reported quality of life has a stronger relationship with mood symptoms (particularly depression) than cognitive symptoms {^{9, 10}}. Secondly, due to the relationship between mood symptoms and cognition, interventions targeting mood symptoms may have the potential to delay the progression of further cognitive impairment {^{2, 11}}. These two points are complemented by the fact that mood symptoms may be responsive to interventions, while efficacious interventions for cognitive impairment are still needed {²⁰}.

To date, at least two previous studies have examined the occurrence of mood symptoms in PD patients with and without MCI {^21,22}. Both of these studies gave the Neuropsychiatric Inventory (NPI), a short 12-item screener that assesses apathy, depression and anxiety using a single item each. The smaller of these studies followed the MDS MCI classification system and found the MCI group (N=48) differed from the cognitively intact PD group (N=54) on the apathy item, but not on the other domains or the total NPI score {²²}. In a second study with a larger sample of PD patients (N=410), Monestero and colleagues {²¹} defined MCI subtypes (amnestic vs non-amnestic). The major finding was greater apathy, depression and overall neuropsychiatric symptom severity in the amnestic MCI PD group. Critically, they did not fully follow the Level II criteria recommended by the MDS for defining MCI. In their study, MCI criteria was based on only 1 test being impaired within any cognitive domain (versus a two test impairment criteria recommended by the MDS criteria). This possibly led to a miscategorization of MCI. Taken together, discrepancies between our findings and the two previous studies likely reflect differences in how MCI was defined and operationalized and differences in how mood symptoms were assessed (i.e., single item measure versus standardized questionnaire for each mood domain).

The mechanisms underlying cognitive and mood symptoms are complicated but may overlap. Dysregulation of similar neural mechanisms may explain why mood symptoms are more common among individuals with cognitive impairment. One proposed mechanism involves the temporal-spatial pattern of dopamine degeneration in the striatum {¹⁴}. Pathological studies have shown a dorsal-ventral gradient, in which striatal dopamine degeneration begins in the dorsal-lateral head of the caudate and then spreads to the ventral-medial portions {²³}. Indeed, the ventral regions of the caudate are relatively spared in the early stages of PD (compared to the dorsal regions) {²⁴}. Furthermore, while the dorsal region of the caudate projects to areas associated with cognitive functions (particularly executive functions) such as the dorsal-lateral prefrontal cortex (DLPFC), the ventral regions of the caudate project to areas associated with motivational and behavioral changes such as the anterior cingulate cortex (ACC) and the orbital-frontal cortex (OFC) {¹}. As such, mood symptoms (particularly a loss of motivation as seen in apathy and depression) may reflect the later stage of striatal-dopamine depletion in which DLPFC functioning has already been disrupted. While the current study did not find a strong relationship between mood and executive functions (traditionally associated with DLPFC), previous studies in PD have reported relationships between executive functioning apathy and depression {¹¹}.

Apathy may be particularly related to striatal dopamine degeneration, however the mechanism for depression in PD likely includes the disruption of additional neurotransmitter systems. Indeed, past studies have shown that depression, unlike apathy, does not follow the same trajectory as motor symptoms in PD; suggesting that multiple neurotransmitters such as serotonin, or norepinephrine may be involved {^25,26}. Future studies are needed to elucidate the mechanism of depression, independent of apathy, in PD and to clarify the contribution of depression to cognitive profiles.

The relationship between anxiety and cognitive functioning has received scant attention in PD. Past studies of neuropsychiatric symptoms found no differences in anxiety among PD patients who were cognitively intact, MCI or PDD {^21,22}. These discrepancies may be partially due to the fact that these two studies assessed anxiety with a single item of the NPI, rather than a full measure. Among the normal elderly, some studies have shown trait anxiety to be associated with late life cognitive impairment, particularly executive functioning and episodic memory, but results have been inconsistent {²⁷}. Possible mechanisms underlying the relationship between anxiety and cognitive impairment in PD include amygdalar dysfunction secondary to striatal-dopamine depletion {²⁸}.

Limitations are present in the current study. First, the difference in mood symptoms among patients with and without MCI was relatively small in terms of effect sizes. Future longitudinal studies are needed to examine the directionality and temporal relationship between the development of mood symptoms and MCI, as well as the clinical utility (i.e. do mood symptoms moderate the conversion from MCI subtypes to dementia). This sample consisted of Parkinson patients referred for a neuropsychological evaluation from an outpatient movement disorder center. This may potentially limit the generalizability of findings. Additionally, neuropsychological measures did not include non-verbal memory tests, however a memory impairment biased toward verbal or non-verbal information is not a common cognitive feature of PD. This study relied on self-report measures of mood, and future studies may benefit from informant/caregiver reports. Even though MCI is a potentially useful diagnostic entity, issues exist regarding what tests to use, what cut-off score to use, and how many impaired tests per domain are required in defining MCI. The current study utilized the MDS criteria for MCI {²}. The MDS criteria are generally consistent with DSM-V criteria for mild neurocognitive disorder, as both criteria require objective (both recommend 1–2 SD below the mean) and subjective evidence of cognitive decline, and relatively intact activities of daily living {²⁹}. However, while DSM-5 criteria do not explicitly specify the number of tests impaired, the MDS criteria specifies impairments on at least 2 neuropsychological tests. In the current study 47.2% of the sample met criteria for MCI, a number that is generally consistent with previous studies in PD that have found 19%–53% of PD participants met criteria for MCI {²}. A less conservative criteria (i.e. defining participants with cognitive impairment on only 1 test, as opposed to 2 tests) would have increased the likelihood of false positive misclassification of participants.

Conclusion

This study found depression and trait anxiety to be more common among MCI-PD patients. In addition to identifying PD patients at risk for dementia, MCI may be useful for identifying individuals experiencing significant mood symptoms.