Triggers and Risk Factors for Parkinson Disease Psychosis
Triggers and Risk Factors for Parkinson Disease Psychosis
Among 417 patients with PD, 86 patients were excluded because of use of antipsychotic drugs at study enrollment; therefore, 331 patients were evaluated.
One-hundred and ninety-two patients were followed until the occurrence of serious psychosis or the end of the study, and all prescriptions were collected longitudinally. During the study period 139 patients were censored because they were transferred to local hospitals (n = 128) or underwent surgery (n = 11). Fifty-two patients required anti-psychotic medications against psychosis (33 with visual hallucinations, 3 with auditory hallucinations, 1 with somatosensory hallucinations, and 15 with delusions). The assumption that anti-psychotic drugs were prescribed when psychotic symptoms were not improved by other treatments and the patients suffered was confirmed. No patients were prescribed anti-psychotic drugs for delirium. Two-hundred and seventy-nine patients (including the 139 that were censored) did not require them during the 2-year period (Figure 1). The incidence of psychosis was 116 (95% confidence interval [CI], 85–148) per 1,000 person-years.
(Enlarge Image)
Figure 1.
Flow diagram of patients included in and excluded from the study. The numbers of patients included and excluded in the analyses. The study involved Analysis I (patient-related factor in a survival time analysis and Analysis II (intra-subject comparison of medication). According to the purpose of analyses, all participants were included in Analysis I, and participants without psychosis were excluded from Analysis II.
Characteristics of patients with and without psychosis are summarized in Table 1. Sex, mH–Y stage, MMSE score and duration of PD were significantly different between patients with psychosis and those without psychosis. Kaplan–Meier curves for the cumulative incidence of psychosis according to mH–Y stage (1–3 or 4–5), and MMSE score (≤24 or >24) are shown in Figure 2. Within the 2 years, >35% of the patients with an MMSE score ≤24 experienced psychosis, whereas only ~10% of the patients with an MMSE score >24 succumbed. More than 30% of the patients with a mH–Y stage of 4–5 developed psychosis during the study period, whereas only ~10% of the patients with a mH–Y stage of 1–3 developed psychosis. Cox hazard models demonstrated that duration of PD, mH–Y stage and MMSE score were statistically significant risk factors with age-and sex-adjusted hazard ratios (HRs) of 1.43 (95% CI, 1.17–1.73) (per 5 years), 3.39 (95% CI, 1.88–6.11) (4–5 versus 1–3), and 3.60 (95% CI, 1.90–6.84) (≤ 24 versus > 24), respectively. The multivariable-adjusted HRs were 1.25 (95% CI, 1.00–1.55), 2.22 (95% CI, 1.11–4.40) and 2.66 (95% CI, 1.37–5.16), respectively.
(Enlarge Image)
Figure 2.
Kaplan-Meier curves of psychosis by the severity of PD and by cognitive function. mH–Y; modified Hoehn–Yahr stage, MMSE; Mini-Mental State Examination. A. Patients with MMSE scores >24 (n = 162) and those with MMSE scores ≤24 (n = 116) were compared after exclusion of 56 patients because of missing values. B. Patients with a mH-Y stage of 1–3 (n = 201) and those with a mH-Y stage of 4–5 (n = 120) were compared after exclusion of 13 patients because of missing values.
The correlation of medication doses among data collection time points depends on the length of the intervals (Additional file 2: Figure S2). Therefore a generalized estimating equation was adopted with an autoregressive matrix. Incorporating medications as predictable variables, the risk owing to drugs was analyzed in patients with psychosis adjusting for age, sex, duration of PD, mH–Y stage, and MMSE score. Because the interactions between anticholinergic drugs and duration of PD (HR 0.76, 95% CI, 0.64–0.89, p = 0.001), were highly significant, the interaction was also incorporated into the equation. There were no other significant interactions between factors. The HR for the use of anticholinergic drugs was 19.7 (95% CI, 2.39–163; p = 0.006). Donepezil was negatively associated with psychosis (HR 0.48; 95% CI, 0.27–0.85; p = 0.012) (Table 2). In the subgroup analysis of patients aged ≥70 years, use of anticholinergic drugs was positively associated with psychosis (HR 188, 95% CI, 13.9–2551, P < 0.001). Donepezil was again negatively associated with psychosis (HR 0.28, 95% CI, 0.08–1.01), and the association was statistically marginal but not significant (p = 0.051). In addition the dose of dopamine agonists was significantly associated with psychosis (HR 1.65, 95% CI, 1.02–2.66, p = 0.035). The interaction between anti-cholinergic drugs and PD duration was also significant (HR 0.62, 95% CI 0.50–0.77, p < 0.001). In contrast there were no significant associations between medications and psychosis in the subgroup of patients aged <70 years.
Results
Among 417 patients with PD, 86 patients were excluded because of use of antipsychotic drugs at study enrollment; therefore, 331 patients were evaluated.
Analysis I
One-hundred and ninety-two patients were followed until the occurrence of serious psychosis or the end of the study, and all prescriptions were collected longitudinally. During the study period 139 patients were censored because they were transferred to local hospitals (n = 128) or underwent surgery (n = 11). Fifty-two patients required anti-psychotic medications against psychosis (33 with visual hallucinations, 3 with auditory hallucinations, 1 with somatosensory hallucinations, and 15 with delusions). The assumption that anti-psychotic drugs were prescribed when psychotic symptoms were not improved by other treatments and the patients suffered was confirmed. No patients were prescribed anti-psychotic drugs for delirium. Two-hundred and seventy-nine patients (including the 139 that were censored) did not require them during the 2-year period (Figure 1). The incidence of psychosis was 116 (95% confidence interval [CI], 85–148) per 1,000 person-years.
(Enlarge Image)
Figure 1.
Flow diagram of patients included in and excluded from the study. The numbers of patients included and excluded in the analyses. The study involved Analysis I (patient-related factor in a survival time analysis and Analysis II (intra-subject comparison of medication). According to the purpose of analyses, all participants were included in Analysis I, and participants without psychosis were excluded from Analysis II.
Characteristics of patients with and without psychosis are summarized in Table 1. Sex, mH–Y stage, MMSE score and duration of PD were significantly different between patients with psychosis and those without psychosis. Kaplan–Meier curves for the cumulative incidence of psychosis according to mH–Y stage (1–3 or 4–5), and MMSE score (≤24 or >24) are shown in Figure 2. Within the 2 years, >35% of the patients with an MMSE score ≤24 experienced psychosis, whereas only ~10% of the patients with an MMSE score >24 succumbed. More than 30% of the patients with a mH–Y stage of 4–5 developed psychosis during the study period, whereas only ~10% of the patients with a mH–Y stage of 1–3 developed psychosis. Cox hazard models demonstrated that duration of PD, mH–Y stage and MMSE score were statistically significant risk factors with age-and sex-adjusted hazard ratios (HRs) of 1.43 (95% CI, 1.17–1.73) (per 5 years), 3.39 (95% CI, 1.88–6.11) (4–5 versus 1–3), and 3.60 (95% CI, 1.90–6.84) (≤ 24 versus > 24), respectively. The multivariable-adjusted HRs were 1.25 (95% CI, 1.00–1.55), 2.22 (95% CI, 1.11–4.40) and 2.66 (95% CI, 1.37–5.16), respectively.
(Enlarge Image)
Figure 2.
Kaplan-Meier curves of psychosis by the severity of PD and by cognitive function. mH–Y; modified Hoehn–Yahr stage, MMSE; Mini-Mental State Examination. A. Patients with MMSE scores >24 (n = 162) and those with MMSE scores ≤24 (n = 116) were compared after exclusion of 56 patients because of missing values. B. Patients with a mH-Y stage of 1–3 (n = 201) and those with a mH-Y stage of 4–5 (n = 120) were compared after exclusion of 13 patients because of missing values.
Analysis II
The correlation of medication doses among data collection time points depends on the length of the intervals (Additional file 2: Figure S2). Therefore a generalized estimating equation was adopted with an autoregressive matrix. Incorporating medications as predictable variables, the risk owing to drugs was analyzed in patients with psychosis adjusting for age, sex, duration of PD, mH–Y stage, and MMSE score. Because the interactions between anticholinergic drugs and duration of PD (HR 0.76, 95% CI, 0.64–0.89, p = 0.001), were highly significant, the interaction was also incorporated into the equation. There were no other significant interactions between factors. The HR for the use of anticholinergic drugs was 19.7 (95% CI, 2.39–163; p = 0.006). Donepezil was negatively associated with psychosis (HR 0.48; 95% CI, 0.27–0.85; p = 0.012) (Table 2). In the subgroup analysis of patients aged ≥70 years, use of anticholinergic drugs was positively associated with psychosis (HR 188, 95% CI, 13.9–2551, P < 0.001). Donepezil was again negatively associated with psychosis (HR 0.28, 95% CI, 0.08–1.01), and the association was statistically marginal but not significant (p = 0.051). In addition the dose of dopamine agonists was significantly associated with psychosis (HR 1.65, 95% CI, 1.02–2.66, p = 0.035). The interaction between anti-cholinergic drugs and PD duration was also significant (HR 0.62, 95% CI 0.50–0.77, p < 0.001). In contrast there were no significant associations between medications and psychosis in the subgroup of patients aged <70 years.
Source...