SEARCH

SEARCH BY CITATION

Keywords:

  • longitudinal;
  • nonmotor symptoms;
  • Parkinson's disease;
  • predictors;
  • quality of life

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Objectives:  Few longitudinal studies have evaluated health-related quality of life (HRQoL) in Parkinson's disease (PD) and these studies have not evaluated all potentially relevant domains of PD. Our objective was to identify domains at baseline that predict change in HRQoL, and to evaluate the relation between change in HRQoL and change in particular domains of PD.

Methods:  A total of 336 patients who participated in the longitudinal SCOPA-PROPARK cohort study and had data from the first and third annual evaluation were included in this study. The Scales for Outcomes in Parkinson's disease (SCOPA) evaluation was used to assess impairments and disabilities. HRQoL was assessed with the EuroQol-5D visual analogue scale. Multiple linear regression analysis with “change in HRQoL” as dependent variable was used to identify factors that influence the change in HRQoL.

Results:  Health-related quality of life as well as most impairment and disability domains decreased significantly from baseline to follow-up. The two regression models of “change in HRQoL,” adjusted for baseline HRQoL, included 1) the baseline domains autonomic dysfunction, nighttime sleep problems, and cognitive function, and 2) “change in psychosocial problems,”“change in depressive symptoms,” and “change in cognitive function.”

Conclusions:  Patients who have autonomic dysfunction, nighttime sleep problems, and cognitive dysfunction are at risk for deterioration in HRQoL. Deterioration in HRQoL over 2-year time was associated with worsening in psychosocial well-being, mood, and cognitive function. Interventions aiming to improve these domains are important and would likely contribute to improvement in HRQoL, although more research is necessary.


Introduction

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Parkinson's disease (PD) is a neurodegenerative disorder that affects almost 2% of the population over 65 years of age [1]. PD is characterized by the motor symptoms braykinesia, rigidity, tremor, and postural instability, as well as a significant response to dopaminergic treatment. Nevertheless, increasingly, it is recognized that the clinical spectrum of PD is much broader including many nonmotor symptoms [2], and motor and psychiatric complications of therapy. Because many areas may be affected in PD, the disease is associated with a decrease in health-related quality of life (HRQoL). HRQoL is defined as those aspects of self-perceived well-being that are related to or affected by the presence of disease or its treatment [3].

The various factors that impact upon HRQoL in PD have mainly been identified in cross-sectional studies [4–6]. Nevertheless, to describe the pattern of change and to describe the magnitude of causal relationships between variables, a longitudinal study design is required. Knowledge on the longitudinal course of HRQoL in PD is of great importance for estimating and prioritizing care requirements of patients with PD. We found only four longitudinal studies that evaluated change in HRQoL in relation to baseline or changes in particular domains of PD. In one study on 111 PD patients, no clinical or demographic factors at baseline predicted deterioration of HRQoL over a 4-year period [4]. Another study evaluated 301 early PD patients who participated 4 years in a randomized, double-blind, placebo-controlled trail, and aimed to determine the impact of motor complications on QoL. It concluded that motor complications did not have a significant negative effect on HRQoL [7]. The same study population was used in another study to evaluate the differences between treatment with either initial levodopa or pramipexole on HRQoL [8]. Compared to levodopa, the initial use of pramipexole resulted in a somewhat larger gain in HRQoL and was explained by pramipexole's influence on nonmotor mechanisms. Recently, predictors of decline in HRQoL in PD were evaluated in 362 PD patients enrolled in the DATATOP study, with a mean follow-up of 1.7 years [9]. Baseline depression and self-rated cognitive function were associated with decline in the 36-item short-form health survey (SF-36) Physical Component Summary score, whereas older age and worse ADL at baseline were associated with a stronger decline in the SF-36 Mental Component Summary score. The only features that deteriorated concurrently with HRQoL were postural instability and gait disorder. These longitudinal studies, however, either reflected only patients with early PD, or were conducted in a trial setting, but above all did not evaluate all potentially relevant impairment domains of PD that are included in the clinical spectrum.

In 1999, the Scales for Outcomes in Parkinson's disease (SCOPA) project was initiated with the aim to set the stage for research on the disease course of PD by developing methodological sound rating scales that cover all domains of this disorder. The SCOPA framework follows a clear pathway that links impairments to disability, and incorporates the patient's perspective on health status. Using the SCOPA framework in a longitudinal design, the current study set out to 1) identify domains at baseline that predict change in HRQoL; and 2) to evaluate the relation between change in HRQoL and change in particular domains of PD.

Methods

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Design

The study is part of the “PROfiling PARKinson's disease” (PROPARK) study, a longitudinal cohort study of patients with PD, who are profiled on phenotype, genotype, disability, and global outcomes of health (http://www.scopa-propark.eu). Reliable and valid measurement instruments for the different domains of PD were derived from the SCOPA project. Within the SCOPA project, assessment instruments were evaluated for reliability (test–retest reliability, interrater reliability, and internal consistency), and validity (content validity, factorial validity, criterion validity, and discriminant validity). The scales were refined if they exhibited psychometric shortcomings, or were developed if no instrument was available. Data were obtained from patients who had their baseline assessment between April 2003 and August 2005 and their follow-up assessment 2 years later, between May 2005 and October 2007.

Participants

All patients fulfilled the United Kingdom Parkinson's Disease Society Brain Bank criteria for idiopathic PD [10]. Age at onset and disease duration are important determinants of the disease course in PD and are related to various manifestations of the disease [11,12]. To obtain an adequate distribution of these characteristics across the cohort, we constructed four strata, based on age at onset (onset of the first symptoms as perceived by the patient [≤/>50 years]) and disease duration (≤/>10 years). Recruitment stopped when approximately 100 patients per stratum were included. This number of patients could not be achieved in the principal center (Leiden University Medical Centre). Therefore, nearby university and regional hospitals were requested to participate in the recruitment of patients. No other selection criteria were applied. The study was approved by the medical ethical committee of the Leiden University Medical Center (LUMC) and all participants gave informed consent.

Outcome Measures

Information was obtained on clinical and sociodemographic variables and included age at onset, disease duration, disease severity measured with the Hoehn and Yahr (H&Y) scale [13], medication, falls in the last year, age, marital status, educational level, and employment status. Levodopa equivalents (LDE) were calculated for the dose of levodopa (LDE-Dopa) and dopamine agonists (LDE-DA), and a total LDE was calculated by adding up the equivalents [14]. The following domains of the disablement process were assessed:

Impairments.  Motor symptoms and motor complications (SPES/SCOPA, sections motor symptoms [MS] and motor complications [MC]) [15], cognitive function (SCOPA-COG) [16], psychiatric complications (SCOPA-PC) [17], depressive symptoms (Beck Depression Inventory [BDI]) [18], nighttime sleep problems (NS) and daytime sleepiness (DS) (SCOPA-SLEEP sections NS and DS) [19], autonomic dysfunction (SCOPA-AUT) [20], and pain (visual analogue scale [VAS] for pain). Patients were asked to rate their average pain in the last month on a line ranging from 0 (no pain) to 100 (worst imaginable pain).

Disability.  Activities of daily living (SPES/SCOPA ADL) [15], and psychosocial problems (SCOPA-PS) [21].

Quality of life.  Health-related quality of life was measured using the VAS from the EQ-5D (EQ-VAS), a generic HRQoL instrument [22]. Patients were asked to rate their health status on a line ranging from 0 (death) to 100 (best imaginable health state) while taking physical, psychological, and social aspects into account. Except for the SCOPA-COG and the EQ-VAS, higher scores reflect more severe problems. Data were collected by means of self-report questionnaires (SCOPA-SLEEP NS and DS, SCOPA-AUT, BDI, SCOPA-PS, EQ-VAS, VAS-PAIN). Items in all questionnaires refer to the past month. Patients received the questionnaires 2 weeks before their assessment and were asked to complete the questionnaires at home, at their convenience. A trained researcher administered the SPES/SCOPA sections MS, MC, and ADL, SCOPA-COG, and the SCOPA-PC. A partner, relative, or caregiver was requested to be present during the assessment. The majority of the patients were assessed at the LUMC. To avoid bias toward recruiting less severely affected patients, patients who were unable to come to the hospital were assessed at home.

Analysis

Patients with a missing value on the EQ-VAS were excluded from the analysis. Means and standard deviations for all impairment and disability domains, and HRQoL were calculated for baseline, follow-up, and change scores. Change scores were calculated as the difference between follow-up and baseline (deteriorations in HRQoL thus with a positive figure, deteriorations in the impairment domains with a negative figure). The effect size of the EQ-VAS was calculated by the Standardized Response Mean (SRM): the mean change in score, divided by the standard deviation of the change in scores [23]. A larger SRM indicates greater responsiveness. Paired samples t-tests were used to compare baseline and follow-up scores of the domains. Independent samples t-test were used to compare different groups of patients. Pearson correlations were calculated to assess the relation between change in HRQoL and 1) domain scores at baseline; 2) change in domain scores. Correlation coefficients were defined as very weak (r = 0–0.19), weak (r = 0.20–0.39), moderate (r = 0.40–0.59), strong (r = 0.60–0.79), or very strong (r = 0.80–1.00) [24]. Multiple linear regressions were performed with “change in HRQoL” as dependent variable. To adjust for baseline HRQoL, a first block consisted of baseline HRQoL (method enter) as independent variable and a second block consisted of domain scores at baseline or change in domain scores as independent variables (method stepwise). All analyses were performed with Statistical Package for the Social Sciences 14.0 software (SPSS 14.0).

Results

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Participants

The SCOPA-PROPARK cohort included 420 patients at baseline, of which 336 patients were available for longitudinal evaluation of HRQoL (Fig. 1). The mean (SD) age at baseline was 59.7 (11.0) years and the mean (SD) disease duration was 10.5 (6.5) years (Table 1). Patients who were not included in the longitudinal evaluation did not differ from the patients who were with respect to disease duration, sex, years of education, or having a partner. They were, however, significantly older, older at onset of disease, more often retired, and had a higher H&Y stage and lower baseline HRQoL.

image

Figure 1. Flow chart of patients. EQ-VAS, EuroQol-visual analogue scale; PD, Parkinson's disease.

Download figure to PowerPoint

Table 1.  Patient characteristics at baseline
N336
  1. H&Y, Hoehn and Yahr.

Male/female (% male)218/118 (65)
Age, years (SD)59.7 (11.0)
Education, years (SD)12.0 (4.1)
Patient with partner, N (%)270 (80)
Employment, N (%)96 (29)
Age at symptom onset, years (SD)49.3 (11.4)
Age at diagnosis, years (SD)52.2 (10.7)
Disease duration since symptom onset, years (SD)10.5 (6.5)
H&Y 1/2/3/4/5/missing13/169/92/47/7/8

Changes from Baseline to Follow-up

Health-related quality of life decreased significantly from 68.2 (14.6) at baseline to 63.6 (14.8) at follow-up (mean difference [95% CI]−4.6 [−3.0 to 6.2]) (Table 2). The SRM of HRQoL was 0.3 (4.6/14.8). Based on Cohen's criteria, this is considered a small effect [25]. Most impairment and disability domains deteriorated significantly as well, except for cognitive functioning, depressive symptoms, nighttime sleep problems, and daytime sleepiness (Table 2). Both the total LDE and the LDE-Dopa increased significantly, but the LDE-DA did not.

Table 2.  Domain scores and doses of dopaminergic medication at baseline and follow-up (N = 336)
Domains (range)BaselineFollow-upP-value*
  • *

    P-value of Student's t-test.

  • AUT, autonomic dysfunction; BDI, Beck Depression Inventory; COG, cognitive function; DA, dopamine agonists; Dopa, dose of levodopa; EQ-VAS, EuroQol-visual analogue scale; LDE, levodopa dosage equivalent; SCOPA, Scales for Outcomes in Parkinson's disease; SCOPA-PC, SCOPA-psychiatric complications; SCOPA-PS, SCOPA-psychosocial functioning; SCOPA-SLEEP DS, SCOPA-SLEEP daytime sleepiness; SCOPA-SLEEP NS, SCOPA-SLEEP nighttime sleep; SPES/SCOPA ADL, SPES/SCOPA activities of daily living; SPES/SCOPA MC, SPES/SCOPA motor complications; SPES/SCOPA MS, SPES/SCOPA motor symptoms.

EQ-VAS (0–100)68.2 (14.6)63.6 (14.8)0.000
SPES/SCOPA MS (0–42)13.1 (4.8)15.5 (6.0)0.000
SPES/SCOPA MC (0–12)1.7 (2.3)2.4 (2.7)0.000
SCOPA-COG (0–43)26.4 (5.9)26.5 (6.9)0.843
SCOPA-PC (0–18)1.9 (1.8)2.4 (2.0)0.000
VAS-PAIN (0–100)26.6 (26.4)33.9 (26.0)0.011
SCOPA-AUT (0–69)17.3 (8.1)18.4 (8.2)0.000
SCOPA-SLEEP NS (0–15)4.5 (3.8)4.7 (3.6)0.175
SCOPA-SLEEP DS (0–18)4.9 (3.8)5.2 (3.8)0.100
BDI (0–63)9.9 (6.3)10.2 (6.4)0.330
SPES/SCOPA ADL (0–21)8.5 (3.4)10.5 (3.7)0.000
SCOPA-PS (0–33)8.3 (5.0)9.5 (5.6)0.000
Total LDE, mg/day601.3 (451.7)700.2 (486.0)0.000
LDE-Dopa, mg/day357.7 (359.1)465.9 (414.9)0.000
LDE-DA, mg/day246.7 (226.1)236.3 (224.6)0.290

Baseline Predictors of Change in HRQoL

There was a moderate correlation between “change in HRQoL” and “HRQoL at baseline” (Table 3), i.e., a higher baseline HRQoL is followed by a stronger decline. Furthermore, “change in HRQoL” correlated only very weakly with two baseline domains: depressive symptoms and psychosocial problems. In the linear regression model of “change in HRQoL,” adjusted for baseline HRQoL, variables that remained in the model were the baseline domains autonomic dysfunction, nighttime sleep problems, and cognitive function. Together, these variables explained 30% of the variance (Table 4). To evaluate whether the absence of depressive symptoms and psychosocial problems in the regression model could be attributed to the adjustment for baseline HRQoL, we calculated the correlation between HRQoL and depressive symptoms at baseline (r = −0.58) and between HRQoL and psychosocial problems at baseline (r = −0.63).

Table 3.  Pearson correlation between change (follow-up–baseline) in HRQoL and impairment and disability domains at baseline, and change in domains (N = 336)
DomainsBaseline domains–Change in EQ-VASChange in domains–Change in EQ-VAS
  • *

    Correlation is significant at the 0.05 level.

  • Correlation is significant at the 0.01 level.

  • This correlation between EQ-VAS at baseline and change in EQ-VAS implies that a high score at baseline is associated with more decline during follow-up.

  • AUT, autonomic dysfunction; BDI, Beck Depression Inventory; COG, cognitive function; EQ-VAS, EuroQol-visual analogue scale; HRQoL, health-related quality of life; SCOPA, Scales for Outcomes in Parkinson's disease; SCOPA-PC, SCOPA-psychiatric complications; SCOPA-PS, SCOPA-psychosocial functioning; SCOPA-SLEEP DS, SCOPA-SLEEP daytime sleepiness; SCOPA-SLEEP NS, SCOPA-SLEEP nighttime sleep; SPES/SCOPA ADL, SPES/SCOPA activities of daily living; SPES/SCOPA MC, SPES/SCOPA motor complications; SPES/SCOPA MS, SPES/SCOPA motor symptoms.

SPES/SCOPA MS0.063−0.235*
SPES/SCOPA MC0.023−0.053
SCOPA-COG0.0850.067
SCOPA-PC−0.057−0.080
VAS-PAIN0.064−0.091
SCOPA-AUT−0.020−0.094
SCOPA-SLEEP NS−0.005−0.156
SCOPA-SLEEP DS−0.024−0.061
BDI0.145−0.369
SCOPA-PS0.165−0.466
SPES/SCOPA ADL0.042−0.180
EQ-VAS−0.4941
Table 4.  Linear regression model of change in health-related quality of life (HRQoL) based on baseline domains (N = 336)
Dependent variableIndependent variables*BetaR
  • *

    Variables are ordered in the table as they appeared in the model.

  • Standardized beta.

  • Independent variables: EQ-VAS at baseline was entered in a first block. A second block, method stepwise included baseline values of: Motor symptoms, motor complications, cognitive function, psychiatric complications, pain, autonomic function, nighttime sleep problems, daytime sleepiness, depressive symptoms, psychosocial problems, and activities of daily living.

  • AUT, autonomic dysfunction; COG, cognitive function; EQ-VAS, EuroQol-visual analogue scale; SCOPA, Scales for Outcomes in Parkinson's disease; SCOPA-SLEEP NS, SCOPA-SLEEP nighttime sleep.

Change in HRQoLBaseline EQ-VAS0.3523
Baseline SCOPA-AUT−0.185
Baseline SCOPA-SLEEP NS−0.141
Baseline SCOPA-COG0.111
Total30

Relations between Change in Domains and Change in HRQoL

“Change in HRQoL” correlated very weakly but significantly with changes in nighttime sleep and ADL, weakly with changes in motor and depressive symptoms, and moderately with change in psychosocial problems (Table 3). After adjusting for baseline HRQoL, the regression model included “change in psychosocial problems,”“change in depressive symptoms,” and “change in cognitive function”; the full model explained 48% of the variance in “change in HRQoL” (Table 5).

Table 5.  Linear regression model of change in health-related quality of life (HRQoL) based on change in domains (N = 336)
Dependent variableIndependent variables*BetaR
  • *

    Variables are ordered in the table as they appeared in the model.

  • Standardized beta.

  • Independent variables: HRQoL at baseline was entered in a first block. A second block, method stepwise included change in: Motor symptoms, motor complications, cognitive function, psychiatric complications, pain, autonomic function, nighttime sleep problems, daytime sleepiness, depressive symptoms, psychosocial problems, and activities of daily living.

  • BDI, Beck Depression Inventory; COG, cognitive function; EQ-VAS, EuroQol-visual analogue scale; SCOPA, Scales for Outcomes in Parkinson's disease; SCOPA-PS, SCOPA-psychosocial functioning.

Change in HRQoLBaseline EQ-VAS−0.4525
Change in SCOPA-PS−0.3519
Change in BDI−0.173
Change in SCOPA-COG0.101
Total48

Discussion

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Even in a relatively short follow-up period of 2-year time, patients with PD experienced a significant decrease in HRQoL. Although this change is small for the total group (−4.6), if patients keep deteriorating at this rate, they decline 25% of the scale in 10-year time. Furthermore, some patients deteriorated considerably, whereas other patients showed large improvements. In this period, most domains except cognitive function, depressive symptoms, nighttime sleep problems, and daytime sleepiness significantly deteriorated as well. We first analyzed which baseline factors predicted change in HRQoL, because such information has its utility for guiding management decisions. At baseline, the degree of autonomic dysfunction, nighttime sleep disturbances, and cognitive dysfunction were associated with deterioration in HRQoL. These findings thus highlight a need for appropriate recognition and clinical management of these nonmotor features in PD. Based on the univariate analyses, depressive symptoms and psychosocial problems at baseline had a larger influence on “change in HRQoL” than cognitive dysfunction and nighttime sleep disturbances at baseline. In the regression model, however, depressive symptoms and psychosocial problems were excluded probably by the adjustment for “HRQoL at baseline,” at which level both domains are strong predictors of HRQoL. In the DATATOP cohort, baseline depression and self-rated cognitive function were associated with decline in the SF-36 Physical Component Summary score (mean decline: 3.3) [9]. Although depression remained in their regression model, the variance was mainly attributed to baseline HRQoL (17%) and depression and cognition together only explained 2%. The small influence of depression is likely also a consequence of the adjustment for baseline HRQoL. Cognitive function was a predictor in both our study and the study by Marras, even though different aspects were evaluated, self-rated cognitive function versus a cognitive test battery.

We also evaluated which disease-related domains worsened and if this exerted an influence on the deterioration in HRQoL. Such information is essential for making future policy and management programs for PD. An increase in psychosocial problems, depressive symptoms, and cognitive dysfunction was associated with a deterioration in HRQoL. Psychosocial problems and depressive symptoms were also major contributors to HRQoL in our cross-sectional study [26]. By confirming the importance of these variables in a longitudinal study, this underscores that psychosocial problems and depressive symptoms are key factors that should receive priority in the management of PD.

The absence of a clear association between HRQoL and the conventional PD domains MS, ADL, and MC is striking. These are the main focus in PD management, and despite the increase in dopaminergic medication, both HRQoL and the three main PD domains deteriorated. Change in these domains did not, or only weakly, correlate with change in HRQoL, and none of these variables were in the final regression model.

Several disease-specific HRQoL measures for PD have been developed to measure aspects of health that are of particular concern to patients with PD, for instance the Parkinson's Disease Questionnaire (PDQ-39) [27]. Nevertheless, because such measures encompass items that address specific aspects associated with PD, this would increase the association between HRQoL and these clinical domains. In view of the objectives of this study, we therefore used a generic instrument, the EQ-VAS. The EQ-VAS has been extensively validated, is easy to complete, and has been found useful in a PD population [28]. To ensure that the multidimensionality of HRQoL is captured, patients were explicitly asked to take the physical, psychological, and social aspect of their health state into account while completing the EQ-VAS. Although generic instruments may be less responsive than disease-specific measures, we were able to measure a significant decrease in HRQoL.

This study included 336 PD patients with a follow-up of 2 years, who were extensively assessed for all relevant PD domains and HRQoL. Some bias may have occurred because patients who were lost to follow up were older, older at disease onset, were more often retired, had a more advanced disease stage, and had a worse HRQoL. As a consequence, the influence of problems at baseline on change in HRQoL may have been underestimated. Despite these dropouts, a large group of patients who covered the broad spectrum of PD features from mild to severe remained available for the longitudinal analyses.

To conclude, HRQoL in PD decreases already over a relatively short time period. Patients who display autonomic dysfunction, nighttime sleep problems, and cognitive dysfunction are at risk for deterioration in HRQoL, and should be carefully monitored. Deterioration in HRQoL over 2-year time was associated with worsening in psychosocial well-being, mood, and cognitive function. Interventions aiming to improve these domains are important and would likely contribute to improvement in HRQoL, although more research is warranted.

Source of financial support: Fonds NutsOhra and the van Alkemade-Keuls Foundation.

References

  1. Top of page
  2. ABSTRACT
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References