Psychiatric comorbidities in patients with Parkinson disease and psychosis

Patients and methods.

Results.

Discussion.

Previous studies of comorbid psychiatric disturbances in PD patients with psychosis focused primarily on correlations between visual hallucinations and depressive symptom rating scales, which are nonspecific for psychiatric pathology and do not yield prevalence rates for syndromic disturbances. The present analyses were based on comprehensive assessments of PD patients to determine current and past syndromic psychiatric diagnoses, careful characterization of hallucinatory and delusional phenomena, and quantitative measures of motor, cognitive, and psychiatric status. We found that PD patients with psychosis have a high rate of comorbid psychiatric disorders, especially affective syndromes, and that this additional psychopathology is associated with greater cognitive impairment and caregiver distress. Whereas psychosis directly and significantly impacted overall caregiver burden, depression added to functional impairment. Since affective disorders are treatable, specific attention should be paid to the detection and treatment of mood disorders in psychotic PD patients to reduce this added morbidity and caregiver distress. The association of affective disturbances with psychosis also supports a role for non-dopaminergic influences, especially serotonergic dysfunction.24

These data indicate that affective disturbances, including depression, anxiety, and apathetic syndromes, are common in PD patients with psychosis. We did not observe psychosis with comorbid hypersexuality or mania, as previously described25 and, in most cases, psychotic phenomena were independent of the affective disorder, i.e., the patients did not have mood-congruent psychotic features. Our findings support studies showing high rates of depressive symptoms in PD-related psychosis,1-3,7,26,27⇓⇓⇓⇓⇓ but low rates of psychotic affective syndromes.2,7⇓

Two other investigations established psychiatric diagnoses, but only determined rates for comorbid depressive disorders and characterization of psychotic phenomena and cognitive deficits was limited. In the first study, a population-based sample of 245 PD patients from Norway,2 16% had psychotic symptoms and 3.4% of the total sample had psychosis plus comorbid major depression. An earlier analysis of their sample showed that psychosis (based on UPDRS Thought Disorder ratings) was a significant risk factor for presence of major depressive disorder.8 The other study, based on 172 patients from a university-based movement disorders clinic in Israel, reported a 30% prevalence rate for psychosis, with 78% (23% of the total sample) having a DSM-IV depressive disorder as well as psychosis.7 Our rates of psychiatric comorbidities are intermediate between the Norwegian and Israeli samples, with sampling differences the most likely explanation for different prevalence rates. For example, age and disease severity are risk factors for psychosis2 and these may have been greater in patients in the Israeli sample who attended a tertiary care clinic, had at least a 5-year history of PD, and were only regarded as psychotic if symptoms were disruptive. The smaller sample size and recruitment of some participants from psychiatry clinics may have influenced our findings.

In our study, onset of affective disturbances preceded or coincided with onset of psychosis. Since this study was cross-sectional and did not include nonpsychotic patients with affective disorders, we cannot test specifically whether affective disorders per se increase the risk of psychosis. However, supportive data are provided by the Israeli sample, mentioned above,7 in which depression, especially in early onset PD, predicted later onset of psychosis. In their study, psychosis frequently accompanied comorbid depressive disorders and dementia, with higher rates of psychosis among patients with depression compared to those with dementia. Fénelon et al.1 also found higher depressive symptom ratings in patients with minor hallucinations than in non-hallucinators. One other study reported significantly higher rates of a history of depression in patients with visual hallucinations (43%) as compared to non-hallucinators (17%),28 but details regarding current affective disturbances were not provided. By contrast, another study3 failed to detect an association between visual hallucinations and prior history of a psychiatric disorder, although the role of past depressive disorders was not specified. Others26 found that PD patients with hallucinations and PD duration over 5 years had lower depressive symptom scores compared to non-hallucinators. However, the authors suggested that cognitive impairment may have contributed to underreporting of depressive symptoms. Prospective longitudinal studies are necessary to better understand these relationships among affective disturbances, psychosis (including hallucinations and delusions), and other non-motor aspects of PD.

The 13% overall prevalence rate for auditory hallucinations was unexpected. However, this is not substantially higher than reported rates of 9.7%1 and 8%29 for auditory hallucinations in recent studies. Even higher rates (35.5%) were associated with levodopa in the past.30 In most patients, auditory hallucinations were accompanied by visual hallucinations, as in other studies. Although auditory hallucinations may be associated with mood disorders, we did not find group differences with respect to psychiatric comorbidities (see table 2). It is possible that clinicians underappreciate auditory hallucinations in PD patients; our use of structured interviews (the SCID and NPI) may have facilitated ascertainment.

Greater cognitive impairment, one of the most consistently reported correlates of psychosis,1-3,7,26,28⇓⇓⇓⇓⇓ was also evident in our psychotic group and especially prominent in patients with psychiatric comorbidities. Compared to PDN, the PSY+ group was significantly more impaired on global cognitive measures (MMSE and DRS) as well as on the DRS memory and conceptualization subscores. However, both psychotic groups showed relative deficits on an index of frontal lobe functioning, the DRS Initiation-Perseveration scale. This combination of global and enhanced frontal cognitive deficits adds to the complexities of disease management.

Other than the association with dementia, previous studies have paid little attention to patterns of cognitive deficits in PD patients with psychosis. Most PD patients show selective cognitive deficits, especially on tests of executive function.31 As the disease progresses, up to 75% of patients develop more severe and widespread deficits that correspond to the diagnosis of dementia, but the average prevalence of dementia ranges from 15 to 30%.32 One study showed greater deficits on letter fluency in addition to greater global cognitive impairment in hallucinators,26 but the study design may have precluded demonstration of selective memory and executive function deficits as the more impaired psychotic patients were unable to complete the more demanding tasks. This was also the case for some patients in our study. Although the relationship between early cognitive deficits and later dementia is unclear, greater executive dysfunction is associated with a transition to dementia over a 1-year follow-up period33 and may thus represent an earlier and more specific risk factor or marker for development of psychosis.

Although the pathophysiology of psychosis in PD is not known, our findings suggest that affective disturbances contribute to development of psychosis. Dopaminergic antiparkinsonian medications and secondary dopamine receptor hypersensitivity and overstimulation in mesocortical and mesolimbic regions are most commonly implicated.26 However, development of psychosis is not directly related to the dose of dopaminomimetic medications2 or levodopa plasma level34 and non-dopaminergic neurotransmitter systems are also involved.24 A few reports describe development of psychosis in PD patients taking antidepressants that inhibit dopaminergic (e.g., bupropion)35 and serotonin reuptake,36 which may enhance dopamine release by stimulation of 5HT-3 receptors. Although antidepressant therapy typically improves affective psychoses, longitudinal analyses are needed to evaluate the possibility that antidepressants over time, when used with antiparkinsonian medications, exert a priming effect that contributes to development of psychosis.30

Depression in PD further involves changes in the serotonin and dopamine systems that could drive emergence of psychosis in the setting of neurochemical or structural changes associated with dopaminergic therapy, aging, disease progression, and cognitive impairment. For example, sleep-wake disturbances, which are common in depression and anxiety, could predispose to development of psychosis.37 Imbalance between dopaminergic and serotonergic systems24 or overstimulation of serotonergic receptors by dopaminergic medications38 may be relevant when there are comorbid affective disorders, which are known to be associated with serotonergic deficits.39 For example, higher depressive symptom scores in PD patients with psychosis were associated with an apomorphine-induced growth hormone response,40 an index of central dopaminergic receptor sensitivity. Thus, efficacy and tolerability of atypical antipsychotic agents in PD-related psychosis may be related to their potent antiserotonergic properties but reduced dopamine D2 receptor blockade.38

Since cognitive deficits in PD are associated with cholinergic deficits41 and the most consistent clinical correlate of psychosis is cognitive impairment, cholinergic deficits are also likely to play a role in the development of psychosis. Therapeutically, cholinesterase inhibitors in PD patients are reported to reduce psychosis42 and, in controlled trials, improve cognitive impairments in PD.43,44⇓

Although psychosis in general is associated with greater motor dysfunction and disability, we showed that comorbid psychopathology in psychotic patients has a substantial impact on PD at additional levels, including greater cognitive impairment and perceived caregiver burden. Whereas psychosis has a direct and significant impact on overall caregiver burden, depressive symptoms predicted greater motor impairments in ADLs and thus indirectly added to caregiver burden. Even though psychosis is reported as a main reason for nursing home admission of PD patients,5 nonpsychotic psychiatric disturbances, especially depression and anxiety, are an important source of additional physical morbidity and caregiver strain that may contribute to need for institutionalized care. Since clinically significant nonmotor symptoms occur in most PD patients and most patients have more than one comorbidity,45 identification of one psychiatric symptom suggests a need to evaluate for other psychopathologic symptoms and syndromes. Unfortunately, despite their clinical impact, comorbid psychiatric conditions and other non-motor symptoms in PD tend to be underrecognized and undertreated.46

Patients with advanced PD, who are already challenging to manage, often have multiple neuropsychiatric disturbances that add to treatment complexity and affect the well-being of caregivers. Early recognition of psychosis in addition to other psychiatric disturbances, patient and caregiver education, and thoughtful use of antiparkinsonian and psychiatric medications is integral to treatment. Since treatment of psychiatric comorbidities can reduce the burdens of the disease, clinicians should screen PD patients regularly for psychotic and affective symptoms. In particular, presence of psychosis should trigger assessment of affective pathology and, conversely, affective pathology should prompt an inquiry about psychotic symptoms. However, overlap among depressive symptoms, motor signs of PD, medication effects, and cognitive impairment can confound recognition of psychopathology.39 In our sample, the PSY had more severe psychotic symptoms, but overall cognitive and motor status was similar between the PSY and PSY+ groups. Accordingly, without specific inquiry, comorbid psychiatric states may not be detected and patients and caregivers will have unwarranted additional morbidity.