Correlation between depressive symptoms and nocturnal disturbances in Japanese patients with Parkinson's disease

1. Introduction 

Non-motor symptoms such as cognitive dysfunction, psychiatric symptoms, and sleep disorders have attracted attention recently, in addition to motor symptoms, in Parkinson's disease (PD). In a community-based study, two-thirds of the patients with PD reported sleep disorders [1]; however, the etiology is still controversial. About 40% of PD patients have depression [2], which also involves sleep disorders [3] and is associated with impairment of activities of daily living [4]. Chaudhuri and Martinez-Martin [5] found a significant correlation between sleep disturbances and depression using the Parkinson's disease sleep scale (PDSS) [6], useful for multifactorial nocturnal problems. Using PDSS, other studies reported that sleep disturbances were associated with disease severity, daytime sleepiness [7], and impairment of activities of daily living [8]. Dhawan et al. [9] reported that untreated PD had many nocturnal problems, such as nocturia, nighttime cramps, dystonia, tremor, and daytime sleepiness in PDSS sub-items, suggesting that sleep disorders in PD are more likely to be related to the underlying dopaminergic deficit rather than the effect of dopaminergic treatment. On the other hand, severer changes in monoamines, such as dopamine, serotonin and noradrenalin are reported in the brain of depressed patients than non-depressed patients in PD [2], [10], and it has been reported that depression develops prior to the motor symptoms in PD patients [11]. Therefore, sleep disruption caused by depression may reflect the pathological course of the disease itself in PD.

Although we reported previously that the depressive state is a significant determinant of sleep disorders in PD [12], the specific nocturnal disturbances related to depressive symptoms remain elusive.

To determine the true status of nocturnal disturbances associated with depressive symptoms and the frequency of depressive state in Japanese patients with PD, we conducted the present survey using the PDSS [6] and Zung Self-Rating Depression Scale (SDS) [13] at multiple facilities. This study was a part of an epidemiological study on non-motor symptoms in PD [12].

3. Statistical analysis 

The Mann–Whitney U test was used for continuous data and the chi-square test was used for categorical variables. Kruskal Wallis test was used to test differences in sub-items of PDSS and PDSS total score between PD (depressed and non-depressed patients) and controls. The Spearman's correlation was calculated to compare the SDS and the PDSS total score in PD patients. A stepwise regression model was used to identify the determinants of SDS in PD patients that included age, gender, disease duration, PDSS, H&Y stage, UPDRS I (mental state), UPDRS II (activities of daily living), UPDRS III (motor performance), UPDRS IV (complications of treatment), use of any DA, and use of levodopa/DCI. Subsequently, stepwise regression analysis was applied to determine the important PDSS sub-items for SDS. Significance of differences was defined as two-tailed p<0.05. SPSS II Windows Ver 11.0 (SPSS Japan Inc.) was used for statistical analyses. All data are expressed as mean±standard deviation.

4. Results 

The SDS score was higher in PD patients (43.4±9.6) than in controls (35.4±8.2) (p<0.001). Of the 188 PD patients, 122 (64.9%) had depressive symptoms. Table 1 shows the clinical characteristics of depressed and non-depressed patients with PD. There were no differences in age, gender differences, and disease duration. Depressed patients had significantly higher scores of H&Y stages and UPDRS I–IV, and higher dose of levodopa/DCI compared with non-depressed patients. Significant differences in PDSS total scores and PDSS sub-items except PDSS items 2 and 11 were seen among depressed patients, non-depressed patients, and controls (Table 2, Fig. 1). However, there was no difference in PDSS total scores and PDSS sub-items except PDSS item 13 between controls and non-depressed patients.

Table 1. Clinical characteristics of depressed PD patients, non-depressed PD patients and control subjects

		Non-depressed PD	Depressed PD	Total PD	p Value
	Number	66	122	188
	Men/women	36/30	49/73	85/103	0.059a
	Age	66.8±9.7	66.2±8.2	66.4±8.7	0.473
	Disease duration	6.0±5.1	7.4±5.3	6.9±5.3	0.056
	H&Y stage	2.3±0.7	2.6±0.8	2.5±0.8	0.014
	UPDRS total score	27.0±14.7	36.0±19.1	32.9±18.1	0.001
	UPDRS I	0.6±0.9	1.4±1.9	1.1±1.7	<0.001
	UPDRS II	7.8±5.3	10.3±6.6	9.4±6.3	0.010
	UPDRS III	17.3±9.7	21.8±11.7	20.2±11.2	0.010
	UPDRS IV	1.3±1.6	2.5±3.4	2.1±3.0	0.045
	Levodopa/DCI (mg/day)	240.9±193.5	314.8±207.9	366.9±157.7	0.019
	DA (mg/day)	224.6±140.7	248.8±171.0	240.5±161.1	0.595

Depressed patients with PD were defined as SDS≥40. Data are mean±SD. PD: Parkinson's disease, H&Y stage: Hoehn and Yahr stage, UPDRS: Unified Parkinson's Disease Rating Scale, PDSS: Parkinson's disease sleep scale, Levodopa/DCI: levodopa with decarboxylase inhibitor, DA: dopamine agonists, SDS: Zung Self-Rating Depression Scale.

a Chi-square test.

Table 2. Total scores and sub-items of PDSS in Parkinson's disease and controls

		Controls	Non-depressed PD	Depressed PD	Total PD	p Value
	Total	126.6±17.8	123.2±17.9	107.2±27.1	112.8±25.4	<0.001
	Item 1	7.4±3.1	7.2±3.1	6.0±3.5	6.4±3.4	0.001
	Item 2	7.9±2.8	7.6±3.4	7.2±3.5	7.4±3.4	0.316
	Item 3	6.5±3.2	6.5±3.4	5.2±3.8	5.7±3.7	0.011
	Item 4	9.2±1.8	9.2±1.9	8.2±2.8	8.6±2.6	<0.001
	Item 5	9.1±1.8	9.0±2.1	8.1±2.9	8.4±2.7	<0.001
	Item 6	9.0±1.9	8.7±2.1	7.4±3.2	7.9±2.9	<0.001
	Item 7	9.6±1.3	9.2±2.1	8.1±3.0	8.5±2.8	<0.001
	Item 8	5.3±3.8	4.4±3.9	4.1±3.7	4.2±3.8	0.045
	Item 9	9.7±1.4	9.1±2.1	8.5±2.6	8.7±2.4	<0.001
	Item 10	9.5±1.6	9.4±1.5	7.8±3.3	8.4±2.9	<0.001
	Item 11	8.5±2.3	9.0±2.0	8.3±2.5	8.5±2.4	0.131
	Item 12	9.5±1.3	9.4±1.3	8.0±3.1	8.5±2.7	<0.001
	Item 13	9.8±1.1	8.4±2.7	6.9±3.6	7.4±3.4	<0.001
	Item 14	7.6±2.9	7.8±3.0	6.6±3.4	7.0±3.3	0.012
	Item 15	8.0±2.6	8.3±2.6	6.6±3.4	7.2±3.3	0.001

View Large Image Download to PowerPoint

Fig. 1 Profiles of mean PDSS scores of sub-items in depressed PD, non-depressed PD and controls. Significant differences in PDSS sub-items except items 2 and 11 were seen among depressed patients, non-depressed patients, and controls.

The PDSS total score correlated significantly with the SDS in PD patients (r=−0.41, p<0.001, Fig. 2). In PD patients, stepwise regression model with SDS as the dependent variable and age, gender, disease duration, PDSS, H&Y stage, UPDRS I, UPDRS II, UPDRS III, UPDRS IV, use of any DA, and use of levodopa/DCI as the independent variables identified PDSS (p<0.001) and UPDRS I (p=0.002) as significant determinants of SDS (Table 3). Another model, in which the SDS was used as the dependent variable, while sub-items of PDSS (items 1–15), age, gender, disease duration, H&Y stage, use of any DA, and use of levodopa/DCI were the independent variables, identified item 15 (daytime sleepiness) (p=0.002), item 13 (early morning tremor) (p=0.008), item 12 (nocturnal dystonia) (p=0.015), and item 3 (sleep maintenance insomnia) (p=0.026) as significant determinants of the SDS in PD patients (Table 4).

View Large Image Download to PowerPoint

Fig. 2 Spearman's correlation analysis of SDS and PDSS in PD patients. Correlation coefficient=−0.41, p<0.001.

Table 3. Stepwise regression analysis of SDS in Parkinson's disease

	Parameter	Clinical variable	R2
	SDS	PDSS	0.200
		PDSS+UPDRS I	0.301

SDS was used as the dependent variable, while age, gender, disease duration, PDSS, H&Y stage, UPDRS I, UPDRS II, UPDRS III, UPDRS IV, use of levodopa/DCI, and use of any DA were independent variables. For abbreviations, see Table 1.

Table 4. Stepwise regression analysis of SDS in Parkinson's disease (including sub-items of PDSS)

	Parameter	Clinical variable	R2
	SDS	Item 15	0.108
		Item 15+item 13	0.168
		Item 15+item 13+item 12	0.196
		Item 15+item 13+item 12+item 3	0.218

SDS was used as the dependent variable, while age, gender, disease duration, PDSS (items 1–15), H&Y stage, use of any DA, and use of levodopa/DCI were independent variables. Item 15: daytime sleepiness, item 13: early morning tremor, item 12: nocturnal dystonia, item 3: sleep maintenance insomnia.

5. Discussion 

In this study, depressive symptoms were observed in 64.9% of patients with PD. The higher prevalence rates for depressive symptoms compared with other study may be attributed to screening instrument or patient population. Although a number of studies have used various evaluation methods such as SDS, the Beck Depression Inventory, and the diagnostic criteria of Diagnostic and Statistical Manual of Mental Disorders-IV, the PD-specific evaluation method for depression is still lacking [20]. In SDS, sub-items such as fatigue, sleep disorders, and constipation may be based on the symptoms of PD itself, while items related to daily life and work may be affected by impairments of motor function. In addition, because the SDS is an evaluation tool that relies on the subject's own self-assessment, there is a tendency for overestimating depression in PD.

In neurobiological findings, postmortem examination of the brain of PD patients demonstrated 50–85% cell loss in the substantia nigra and locus ceruleus, 0–43% in the dorsal raphe nucleus and 32–87% cell loss in the basal nucleus of Meynert [21]. Thus, the depletion of endogenous neurotransmitters in the brain varies among patients with PD, suggesting that this variation may have an effect on varied prevalence rates for depression in PD. As for the relationship between excessive daytime sleepiness, it has been demonstrated that bilateral destruction of ventral tegmental area (VTA) causes excessive daytime sleepiness in animal experiments [22] and PD patients with depression show greater dopamine loss in the mesolimbic dopaminergic pathway from VTA, as well as serotonin depletion in the dorsal raphe nucleus, noradrenalin depletion in the locus ceruleus, and acetylcholine depletion in the pedunculopontine nucleus and nucleus basalis of Meynert compared with patients without depression [23]. Moreover, it is believed that degeneration of cholinergic neurons and depletion of noradrenalin can cause disorders of rapid eye movement sleep, and serotonin depletion can reduce the amount of slow-wave sleep [24]. Based on these findings, PD patients with depression may show exacerbation of daytime sleepiness due to change in neurotransmitters.

In our study, Spearman's correlation analysis showed a significant correlation between depressive symptoms and nocturnal disturbances. Although depressed PD showed higher disease severity and severer motor dysfunction compared to non-depressed PD, motor function (UPDRS III) did not enter the final models as a significant variable in stepwise regression analysis with SDS. The results of stepwise regression analysis with SDS as the dependent variable identified PDSS and UPDRS I as significant determinants of the SDS. These findings support the previous reports that sleep disorders are exacerbated by depression in PD patients [3], [25] and that cognitive dysfunction often complicates major depression [26]. We believe that depression, sleep disorders, and cognitive dysfunction interact with one another.

Happe et al. [3] reported significant correlations between SDS and the narcolepsy score and periodic limb movement disorder score (including items regarding restless legs syndrome) in their study of 56 patients with PD and 59 control subjects. On the other hand, in our study, daytime sleepiness and sleep maintenance insomnia and the causes thereof, including early morning tremors and nocturnal paroxysmal dystonia, correlated with the SDS (depressive symptoms) in PD. It is widely believed that sleep maintenance insomnia is more common than difficulty in falling asleep in patients with PD [1]. Because sleep maintenance insomnia was closely associated with depressive symptoms in this study, we consider depression to be associated with nocturnal symptoms, or it may potentially be due to abnormality of the structures involved in sleep regulation caused by the disease process itself.

Dhawan et al. [9] compared nocturnal disturbances among 59 cases of PD (25 cases were untreated and 34 cases were treated) and 131 control subjects using PDSS. The results showed that nocturia, nighttime cramps, dystonia, tremors, and daytime somnolence were important factors for untreated PD. However, these symptoms were similar to the nocturnal symptoms that are important to PD with severe depressive symptoms in our study, they did not include a description of depressive symptoms. It has been reported that dystonia and tremors in PD increase in the “off” state, and depression exacerbates nocturnal off-period-related motor symptoms [2]. Furthermore, depression is reported to trigger the “wearing off”, and a hypodopaminergic state and a psychological “off” have been considered as causes for this phenomenon [27].

With regard to the relationship between dystonia and nocturnal symptoms, Lees et al. [28] found dystonia in 34% of 215 cases of PD, and Starkstein et al. [25] reported that depression was the most important factor associated with sleep disorders and pain. Goetz et al. [29] discussed the possibility that depression changes the interpretation of pain and the possibility that depression exacerbates pain, because pain was frequently observed in patients with severe depression among 95 cases of PD. Animal experiments have shown that tremors are due to depletion of the neostriatal content of serotonin and dopamine [30], while clinical studies indicated that bradykinesia and rigidity are more severe than tremors as characteristics of motor function in depressed PD [2]. Conversely, early morning tremors were closely associated with the depressive state in our study, and this is believed to be due to nocturnal off-period-related symptoms rather than daytime motor function.

In conclusion, there was a significant correlation between patients with depressive symptoms and nocturnal disturbances. We demonstrate that nocturnal disturbances and cognitive dysfunction were significant determinants of depressive symptoms and that daytime sleepiness, dystonia, tremor and sleep fragmentation seem to be the important nocturnal disturbances in depressed patients with PD.