13 - 26: Depression, Apathy and Anxiety Disorders

Depression, apathy and anxiety disorders are among the most common comorbid psychiatric conditions in Parkinson’s disease (PD). Recent studies have consistently demonstrated that depression has a negative impact on patients’ quality of life as well as on the motor and cognitive symptoms of the illness. Apathy is being increasingly diagnosed among elderly PD patients, primarily those with dementia. There is less information about the clinical relevance of anxiety in PD, but recent studies have demonstrated a high frequency of both typical and atypical anxiety disorders in PD. The present chapter will review the nosology and diagnostic methodology for depression, apathy and anxiety in PD; we will summarize the epidemiology of these psychiatric disorders and discuss their clinical correlates and putative mechanisms. We will finish the chapter by discussing the most effective treatment modalities for these conditions.

There is no general consensus about the most valid methods to assess and diagnose depression in PD. One of the most important nosological limitations is the overlap between symptoms of depression and symptoms of PD (e.g. motor retardation vs. bradykinesia, poor concentration and bradyphrenia; loss of energy in both depression and PD). A workgroup established by the National Institutes of Neurological Disorders and Stroke, and the National Institutes of Mental Health proposed provisional criteria for depression in PD [1]. The workgroup stressed the need to validate the DSM-IV categories of major depression, minor depression and dysthymia, as well as the concept of subsyndromal depression in patients with PD. They recommended the use of the inclusive approach to symptom assessment (i.e. to consider all symptoms as related to depression, regardless of the overlap with parkinsonism), to distinguish loss of interest/anhedonia from apathy, to consider whether putative depressive symptoms may be the expression of motor fluctuations, to assess depression at consistent times and during the ‘on’ and ‘off’ states, and to obtain additional information about mood changes from a next of kin or caregiver. Other investigators reported a relatively low frequency of guilt, self-blame and worthlessness in depressed patients with PD [2, 3].

Starkstein et al. [4] have recently examined the validity, sensitivity and specificity of depressive symptoms for the diagnosis of dysthymia, sub-syndromal depression, and major and minor depression in a series of 173 patients with PD. The main finding was that all DSM-IV clinical criteria for major depression and dysthymia were significantly associated with sad mood. Moreover, there was no significant difference in frequency and severity of depressive symptoms, when PD patients with sad mood were compared with sad mood controls of similar age but without PD. Depression diagnosis based on loss of interest without sad mood was significantly more common in minor than in major depression, suggesting that minor depression in PD may be closer to apathy.

Schrag et al. [5] have recently examined rating scales for depression for use in PD. They concluded that the Hamilton Depression Rating Scale (HAM-D), the Beck Depression Inventory (BDI), the Hospital Anxiety and Depression Scale (HADS), the Montgomery-Åsberg Depression Rating Scale (MADRS) and the Geriatric Depression Scale (GDS) are all useful to screen for depression in PD. To measure severity of depressive symptoms, they recommended the HAM-D, the MADRS, the BDI and the GDS. They further suggested that depression should not be diagnosed based on a cut-off score on a depression scale but based on information provided by semi-structured psychiatric interviews and the use of standardized diagnostic criteria. They further suggested to specify the timing of assessment regarding motor fluctuations and to include collateral information.

The frequency of depression in PD has been reported to range widely from less than 10% to greater than 80%. This variability may be explained by sampling bias (e.g. patients recruited from the community or attending movement disorders clinics, different assessment methods, and demographic differences, e.g. differences in the proportion of women, elderly patients, and severity of illness). Using the BDI, the Global PD Survey reported significant depression in 50%, although only 1% of the sample reported depressive symptoms to the clinician [6]. Based on GDS cut-off scores, Holroyd et al. [7] diagnosed depression in 15% of 100 consecutive patients with PD. In a series of 1,449 randomly selected outpatients with PD, Riedel et al. [8] reported 25% of depression based on MADRS cut-off scores. Using structured psychiatric interviews in a series of 173 patients attending a movement disorders clinic, Starkstein et al. [4] reported 30% of major depression, 20% of dysthymia and 10% of minor depression. A recent meta-analysis reported a prevalence of 31% of major depression in PD [9]. Other studies reported frequencies of 17 and 21% for major and minor depression, respectively [10].

Starkstein et al. [11] reported a significant association between major depression, more severe parkinsonism and specific neuropsychological deficits. In two longitudinal studies, these authors demonstrated that depression in patients with PD is associated with faster motor, cognitive and functional decline [12, 13]. Cross-sectional studies demonstrated a significant association between depression and worse quality of life [14], functional capacity and caregivers’ quality of life [15, 16], increased mortality [17], increased burden for caregivers [14, 18], motor-related disability [18, 19], lower cognition [7], sleep disturbances and fatigue [20]. Depression was also reported to be a predictor of social and physical functioning in men with PD [21].

Several studies examined biological correlates for depression in PD. Remy et al. [22] used [¹¹C] RTI (methyl (1R-2-exo-3-exo)-8- methyl-3-(4-methylphenyl)-8-azabicyclo[3.2.1]octane-2-carboxylate)-32 positron emission tomography (PET) to examine in vivo markers of dopamine and norepinephrine terminals in 8 PD patients with major depression and 12 PD patients without depression. Depressed patients showed reduced RTI binding (reflecting a loss of catecholaminergic innervations) in the locus coeruleus, anterior cingulate cortex, thalamus, amygdala and ventral striatum as compared to non-depressed patients. The authors concluded that decreased catecholaminergic innervation in the amygdala and the anterior cingulate may be related to both depression and anxiety in PD, whilst similar decrease in the left ventral striatum may be related to increased apathy. In a recent study, Politis et al. [23] assessed 10 antidepressant-naïve PD patients with major depression, 24 PD patients without depression, and 10 healthy controls using ¹¹C-DASB PET (a compound which selectively binds to the 5-HT transporter). The main finding was that PD patients with the highest depression scores showed increased ¹¹C-DASB binding in the amygdala, hypothalamus, raphe nuclei, and posterior cingulate cortex compared to patients without depression, suggesting that abnormal serotonergic transmission may play an important role in the mechanism of depression in PD. Further support for the association between serotonergic dysfunction and depression in PD was provided by the finding of a significant association between depression and the S allele of the 5-HT transporter [24], although discrepant findings were recently reported [25].

Palhagen et al. [26] used HMPAO SPECT to examine brain perfusion changes before and after 12 weeks of citalopram treatment in 11 PD patients with major depression and 12 individuals with major depression but no PD. After the 12-week treatment period, there was a significant reduction in regional cerebral blood flow (rCBF) in the left fronto-dorsolateral region among PD patients with major depression, whilst patients with major depression and no PD showed a heterogeneous increase in the right hemisphere rCBF.

Depression is usually undertreated in PD. Among 34% patients with PD meeting diagnostic criteria for depression, about two thirds were untreated for the mood disorder [27]. In the context of the Parkinson Study Group which included 27,410 patients with PD, 26% of the sample was on antidepressants, 51% on selective serotonergic reuptake inhibitors (SSRIs), 41% on tricyclic antidepressants (TCAs) and 8% on other compounds [28].

Four randomized controlled trials (RCTs) have been carried out to examine the efficacy of psychoactive drugs to treat depression in PD and have produced discrepant findings. The first study showed that the TCA desipramine and the SSRI citalopram were more effective than placebo [29], while the second study showed that the TCA nortriptyline but not the SSRI paroxetine was more effective than placebo [30]. A third study that compared the efficacy of the selective norepinephrine reuptake inhibitor atomoxetine with placebo showed no significant difference [31], while an open-label study showed no differences between citalopram and placebo [32]. Finally, a recent RCT showed that the dopamine agonist pramipexole improved depressive symptoms in PD [31]. However, the difference on the BDI between the pramipexole and placebo groups was of only 1.9 points, which is of dubious clinical significance. In conclusion, while TCAs may be useful to treat depression in PD, their side effects and contraindications (especially among individuals with a neurodegenerative condition) may severely restrict their use to a small group of patients, and some antidepressants may exacerbate the physical symptoms of PD [33].

Systematic effectiveness studies of psychotherapy are still lacking. Farabaugh et al. [34] enrolled 8 patients with depression in a 12-week trial of individual cognitive-behavioural therapy (CBT). The authors found a linear decrease in HAM-D scores with remission in 4 of 7 patients. Feeney et al. [35] evaluated CBT outcome for major depressive disorder in 3 people with PD. Using an A-B single case experimental design in which each person acts as their own treatment control, 2 out of the 3 participants made significant treatment gains that were maintained at 6 months’ follow-up. In this pilot study, CBT outcome was enhanced when a bereavement model was added to explain the individual's emotional adjustment to living with PD. A recent small-scale study showed similar positive results. This uncontrolled 15-patient study explored the feasibility of using CBT to treat depression in PD and found that patients experienced significant reductions in depressed mood and negative cognitions over the course of 10-14 weeks of treatment, providing preliminary evidence as to the effectiveness of this approach [36]. Patients improved their ability to negotiate physical limitations, addressed barriers to medication adherence, and learned to pace daily activities appropriately, set more realistic goals, identify coping skills, and modify maladaptive cognitive and behavioural responses to physical symptoms [37, 38]. Modified individual CBT may be an effective treatment of depression, especially for patients with medication intolerance. An RCT is now needed to evaluate the efficacy of CBT in PD.

Apathy is defined as a syndrome characterized by deficits in goal-directed behaviour and the simultaneous diminution of the cognitive and emotional concomitants of goal-directed behaviour [39]. The construct of apathy was standardized in a set of criteria by Starkstein and Leentjens (table 1).

Partial validation to this set of diagnostic criteria has been provided in a recent publication [40]. The Movement Disorders Society Task Force on Rating Scales for PD constituted an ad-hoc committee to assess psychometric attributes of existing apathy and anhedonia rating scales for use in PD. One of the main limitations to rate symptoms of apathy is the overlap with symptoms of depression and parkinsonism (e.g. loss of interest, energy and pleasure are all prominent in depression and may also result from the motor problems of PD). The committee recommended the use of the Apathy Scale (AS) [41], which was specifically developed and validated in PD patients. The Lille Apathy Rating Scale (LARS) was considered to be well designed for PD and of potential usefulness, whereas there is little information regarding the properties of the Apathy Evaluation Scale (AES) and the Apathy Inventory (AI). One of the main limitations to assess the attributes of apathy scales in PD is the lack of validated diagnostic criteria. Item 4 of the Unified PD Rating Scale (UPDRS) which focuses on ‘motivation and initiative’ may be used as a screening guide for apathy, given that it has adequate sensitivity and specificity for the clinical diagnosis of apathy in PD [42]. On the other hand, the overlap with both depression and cognitive impairment should be considered when developing new instruments to assess apathy in PD.

The frequency of apathy varied depending on the instruments used for assessment. Based on cut-off points of severity rating scales, the frequency of apathy was reported to range from 17 to 70% [43-47]. Given the overlap between apathy and depression, the frequency of apathy is significantly lower in the absence of comorbid depression [40, 43]. In one of the first studies to examine the frequency of apathy in PD, Starkstein et al. [43] diagnosed apathy in 42% of a series of 50 patients attending a movement disorders unit. Two thirds of the patients with apathy were also depressed. Kirsch-Darrow et al. [47] compared the frequency of apathy and depression among patients with PD and patients with primary dystonia. The main finding was that apathy without depression was present in 28% of the PD sample but in none of the patients with dystonia. In a recent study, Pedersen et al. [48] examined the prevalence of apathy in a community-based sample that included 232 patients with PD. Using the motivation and initiative item of the UPDRS, the authors diagnosed apathy on 35% of the sample. Similar frequencies were reported by Sockel et al. [49] using the LARS. In a study that included 175 PD patients assessed with the Neuropsychiatric Inventory, Aarsland et al. [50] reported that during a 22-month period the incidence of apathy was 27%, suggesting that most patients with PD will develop apathy at some stage during the progression of the illness.

Few studies have examined the longitudinal evolution of apathy in PD. Dujardin et al. [51] assessed 40 PD patients with neither dementia nor depression, 20 of whom had apathy. Patients with apathy showed more severe cognitive deficits than those without apathy. After a median follow-up of 18 months, 8 of the 20 patients with apathy converted to dementia as compared to 1 of 20 patients in the no apathy group. Cognitive decline in the group with apathy was most severe on the memory and executive functions domain. Pedersen et al. [52] carried out a 4-year follow-up study of 79 PD patients from a population-based study. At baseline, 14% of the sample had apathy (as diagnosed with the Neuropsychiatric Inventory), and all of them remained apathetic at the 4-year follow-up. About 50% of the patients with no apathy at baseline became apathetic at the follow-up assessment. Compared to patients with no apathy, those with incident apathy had a higher frequency of dementia and depression at follow-up as well as a faster increase in UPDRS motor scores. Butterfield et al. [53] assessed 68 PD patients using the AES and tests of executive functions and memory. Apathy (but not depression) was significantly associated with executive dysfunction and retrograde amnesia.

In a recent study, Starkstein et al. [43] assessed a series of 164 patients attending a Movement Disorders clinic using the AS, and 32% of the patients met standardized diagnostic criteria for apathy. Patients with apathy were older, had more cognitive deficits, higher depression scores and more severe parkinsonism than PD patients without apathy. On a multiple regression analysis, both the severity of depression and cognitive deficits were significantly associated with more severe apathy. Among patients with neither depression nor dementia, apathy was diagnosed in 13% of the PD group. Dementia is another relevant clinical correlate of apathy in PD. Starkstein et al. [43] reported apathy in 47% of patients with dementia vs. 23% of PD patients without dementia. Similar findings were reported by Dujardin et al. [54] (56% vs. 9%, respectively). Taken together, these findings suggest that apathy identifies a subgroup of PD patients with more severe depression and cognitive deficits, and greater functional impairment.

Few studies examined neuroanatomical correlates of apathy in PD. Reijnders et al. [55] carried out a 3-tesla volumetric MRI study on 55 patients with PD. Using a cut-off on the LARS to diagnose depression, 16% of the patients were diagnosed with apathy. There was no association between apathy scores and severity of motor symptoms or disease duration, but more severe apathy was significantly related with more severe depression. Neuroimaging showed a significant association between higher apathy scores and lower grey matter densities in the bilateral pre-central gyrus, bilateral inferior parietal gyrus, bilateral insula, right posterior cingulate gyrus and right precuneus. The authors suggested that low grey matter density in the premotor cortex may be related to lower motor activation, whilst low grey matter density in the insula may be related to the blunted affect as part of apathy.

Several studies have used PET to examine the metabolic correlates of apathy in PD. Le Jeune et al. [56] showed a significant association between higher apathy scores and decreased glucose metabolism in the bilateral posterior cingulate. Remy et al. [22] showed that increased apathy was associated with decreased ¹¹C-RTI-32 binding (a marker of dopamine and norepinephrine terminals) in the ventral striatum. Drapier et al. [57, 58] reported that patients with apathy had more severe mesolimbic dopaminergic denervation.

Thobois et al. [59]. reported that about half of a sample of patients with PD treated with subthalamic nucleus-deep brain stimulation (STN-DBS) who tolerated a reduction of 80% in anti-parkinsonian drugs developed apathy within the first 6 months after surgery. A subgroup of these patients was assessed with [¹¹C]-raclopride PET, and those with apathy showed greater denervation in the orbitofrontal cortex, dorsolateral prefrontal cortex, anterior cingulate cortex, left thalamus, bilateral globus pallidus and right temporal cortex.

Early studies on the impact of testosterone levels in PD demonstrated a significant association between low testosterone levels and increased apathy scores. However, recent studies using more adequate assessments for apathy were unable to replicate those preliminary findings [60].

Apathy has been reported as a relatively frequent complication among patients undergoing STN-DBS. Le Jeune et al. [56] examined 12 PD patients 3 months before and after STN-DBS using the AES and ¹⁸fluorodeoxy-glucose (FDG-PET). In spite of a significant motor improvement with DBS, apathy (but not depression) scores increased significantly. This increment on apathy scores was significantly related to increased metabolism in the right frontal medial gyrus and right inferior frontal gyrus, and decreased metabolism in the right posterior cingulate gyrus and the left frontal medial cortex. The authors speculated that the STN may have a regulatory function in cortico-subcortical connections. Based on these findings, Levy and Czernecki [61] suggested that apathy in PD may result from deficits in basal ganglia-prefrontal cortex interaction coupled with changes in dopaminergic tone.

There are no RCTs for apathy in PD, and the only therapeutic evidence is based on open label studies. Czernecki et al. [62] reported that dopamine agonists improved motivation in the ‘off’ state. Ropinirole (1-18 mg/day) was reported to improve apathy in a case series of 8 patients treated during 6 months, and no significant adverse events were reported [63]. Finally, an RCT of atomoxetine (with depression as the primary outcome measure) showed no improvements on apathy scores [31].

Whilst anxiety symptoms are common in PD, their frequency has been reported to vary widely. Similar to depression and apathy, this variation may be related to different assessment methods. The MDS Task Force on Rating Scales for PD created an ad-hoc committee to analyze the instruments used to assess anxiety in PD and to determine their psychometric attributes [64]. The main finding was that all the anxiety rating scales analyzed included items that overlap with depression scales and with motor symptoms of PD.

One diagnostic dilemma is that anxiety was found to be related to the motor fluctuations of PD [64]. Therefore, it may be difficult to ascribe anxiety-like symptoms to a DSM-IV anxiety disorder or to the motor fluctuations of PD, and the most frequently used anxiety scales are unable to capture the phenomenology of atypical anxiety disorders in PD. A critique of anxiety rating scales used in PD (e.g. the Beck Anxiety Inventory, the HADS, the Zung Self-Rating Anxiety Scale, the Spielberger State-Trait Anxiety Inventory, and the Hamilton Anxiety Rating Scale) concluded that none of them is suitable for use in PD. Another suggestion was that patients should be assessed in the ‘on’ condition given that patients with motor fluctuations may perceive and report anxiety symptoms differently in ‘on’ vs. ‘off’ states.

Several recent studies examined the frequency of anxiety disorders in PD. Negre-Pages et al. [65] assessed 422 ambulatory patients with PD and 98 age- and gender-comparable non-PD individuals using the HADS. The authors reported clinically relevant anxiety symptoms in 50% of non-demented PD patients, while the frequency in the control group was 29%. Of note, there was a strong comorbidity between anxiety and depression.

Pontone et al. [66] assessed 127 patients with the SCID and found that 50% had a lifetime anxiety disorder, with a diagnosis of anxiety disorders not otherwise specified being the most frequent diagnosis. About 40% of this group had situational anxiety related to parkinsonism (e.g. fear of falling and fear of freezing). Other patients had anxiety during the ‘wearing off’ of anti-parkinsonian drugs, whilst still others had panic-like episodes. Patients with lifetime anxiety have a higher frequency of a positive familial psychiatric history, a lifetime personal history of depression, and lower quality of life as compared to PD patients without lifetime anxiety. The comorbidity between anxiety and depression was high (65%), and the current frequency of anxiety disorders was 43%. Thirty percent of patients met DSM-IV criteria for anxiety disorders not otherwise specified, further suggesting that the psychiatric criteria used to diagnose anxiety in PD may not adequately apply to this population. Nineteen percent had a specific phobia, followed by panic disorder (10%) and social phobia (9%). Partial validation to the anxiety disorders was provided by the finding that quality of life was significantly worse for patients with anxiety disorders as compared to those without.

A recent study [67] assessed 79 PD patients using the Mini International Neuropsychiatric Interview and the Spielberger State-Trait Anxiety Inventory. Twenty-five percent of the sample met DSM-IV diagnostic criteria for a current anxiety disorder, 8% met criteria for panic disorder, 13% for social phobia, and 3% for generalized anxiety disorder. There was a significant association between the presence of anxiety disorders and motor complications (e.g. ‘on-off’ fluctuations and ‘freezing’), worse quality of life, and older age. Finally, the comorbidity between anxiety and depression was 14%.

Leentjens et al. [68] have recently completed a one-year cross-sectional multicentre study that included a consecutive series of 342 patients with PD assessed with the Mini International Neuropsychiatric Interview and several anxiety rating scales. The main finding was that 34% of patients met DSM-IV diagnostic criteria for at least 1 anxiety disorder, whilst 12% met criteria for 2 or more anxiety disorders. Generalized anxiety disorder was the most common diagnosis, followed by agoraphobia and social phobia. A logistic regression analysis showed that female gender, motor fluctuations, and a lifetime history of anxiety disorders were significantly related to the presence of anxiety. Given the psychometric limitations of the anxiety scales for use in PD, the authors suggested developing an anxiety rating scale specific to PD.

In conclusion, whilst anxiety disorders have been consistently reported to be present in 20-40% of PD patients, the type of anxiety disorder was reported to have a wide variation, from initial studies showing a relatively high frequency of panic disorder, to more recent studies showing non-episodic anxiety disorders to be the most frequent in PD. These discrepancies may be related to the use of different psychiatric instruments and diagnostic criteria for anxiety in PD, full or partial screening for DSM-IV anxiety disorders. The high frequency of anxiety disorders not otherwise specified found in recent studies suggests that the use of DSM-IV criteria may not provide an adequate gold-standard for diagnosing anxiety disorders in PD. Future criteria may consider the fact that both motor and autonomic symptoms of PD may be related to the heterogeneity of anxiety disorders in PD.

There is a paucity of treatment studies of anxiety in PD. Menza et al. [69] carried out an 8-week open-label study that included 10 PD patients using citalopram. They reported a significant decrease in HAM-A scores at study completion. The medication was well tolerated and there were no serious side effects.

Depression is one of the most common non-motor disorders in PD. Cross-sectional studies reported that about 20-30% of patients suffer major depression, whilst longitudinal studies suggest an incidence of about 20%. Therefore, most patients with PD will suffer depression at some stage during the illness. Depression is a marker of a ‘malignant’ type of PD as it is associated with faster cognitive, motor and functional decline and worse quality of life. The mechanism of depression remains unknown, but recent studies suggest dysfunction in specific frontal regions. Recent RCTs demonstrated the efficacy of escitalopram, nortriptyline and pramipexole. Future studies should demonstrate the usefulness of psychotherapy. Apathy is another frequent non-motor problem in PD and is significantly associated with depression and cognitive deficits. Its mechanism remains unknown, and no treatment has demonstrated adequate efficacy. Whilst anxiety disorders are common in PD, adequate instruments for diagnosis and measure have yet to be developed. Specific psychotherapeutic techniques should also be developed to treat this condition.

This study was supported by grants from the National Health and Medical Research Council of Australia, the Fremantle Hospital Research Foundation, the University of Western Australia, and the Michael J. Fox Foundation.