Parkinson’s disease - physical activity - bradykinesia - fatigue
doença de Parkinson - atividade física - bradicinesia - fadiga
Parkinson’s disease (PD) is a chronic and degenerative disease of the central nervous
system characterized by a combination of motor and non-motor disorders[1]. PD may severely compromise the ability to perform tasks such as walking, writing,
turning, and moving in bed[1], mainly because of a combination of motor limitations, fatigue, and apathy[2]. The progression of PD is associated with progressive deficits, which might contribute
to sedentarism, and result in reduced physical ability in these patients[3].
A recent study indicated that individuals with PD are 29% less active than healthy
individuals[2]. However, the ability of patients in the initial phase of PD to perform physical
exercise may be comparable to that of healthy individuals when they maintain a certain
level of regular physical activity. Although the daily pattern of physical activity
in the initial phase of the disease appears similar to that of controls[4], this tends to change when PD progresses[4].
Recent studies have assessed factors that might play a role in the level of physical
activity of PD patients[2]
,
[5]. Factors related to the disease such as severity of disease, levodopa daily dose
and motor impairment level; factors related to the individual such as age and walking
performance were pointed out as determinants of the physical activity level[2]
,
[5]. Factors such as lower limb bradykinesia, clinical subtype and fatigue, that may
decrease the level of physical activity in individuals with PD, have not yet been
studied.
It is known that inactivity is considered a risk factor for several diseases and may
worsen various motor and non-motor symptoms that are affected in PD[2]. Additionally, previous studies have explained only up to 28% of the variance of
the level of physical activity of these individuals[2]
,
[5]. Knowledge of the factors that affect the level of physical activity of individuals
with PD may guide clinicians that work with PD rehabilitation, and thereby prevent
inactivity and its consequences. Therefore, the aim of the present study was to analyze
factors associated with the disease (progression time, severity, clinical subtype,
and lower limb bradykinesia) and the individual (age, ability to perform daily living
activities, and fatigue) that may predict or explain the variance in the level of
physical activity of individuals with PD.
METHOD
Participants
This was a cross-sectional study in which the participants were recruited by convenience
from the University’s Outpatient Clinic for Movement Disorders. Participants were
recruited by students and evaluated by a trained researcher. Individuals with idiopathic
PD, diagnosed by a neurologist specialized in movement disorders, according to the
UK Brain Bank criteria[6], classified as a score of 1–4 on the modified Hoehn and Yahr (HY) disability scale[7], using antiparkinsonian medication, and who did not show other neurological, musculoskeletal,
or cardiovascular disorders were included in the analysis. All participants were briefed
on the objectives of the research, and then they signed a free and informed consent
form. This study was approved by the Ethics in Research Committee of the university.
Measures
Identification and characterization
The study participants were subjected to an initial evaluation for data collection,
which included questions regarding identification, age, medication use, progression
of the disease, and associated pathologies.
Clinical evaluation and questionnaires
To evaluate the severity of the disease, activities of daily living (ADL) and motor
sections of the Unified Parkinson’s Disease Rating Scale (UPDRS) and the modified
HY scale were used[7]
,
[8]. The bradykinesia score from the lower limbs was calculated using the sum of the
scores of items 26, 27, 29, and 31 of the UPDRS. The higher scores of those items
indicated increased severity of bradykinesia[9]. The clinical subtypes of PD were classified as tremor-dominant (T), rigid-akinetic
(RA), and mixed (M) according to the classification proposed by Schiess et al.[10]. The perception of fatigue was assessed using the Fatigue Severity Scale (FSS),
and the level of physical activity was determined using the Human Activity Profile
(HAP). The HAP was adapted and validated to portuguese (Brazil)[11] and comprises 94 items of routine activities with different functional levels, and
can classify individuals as inactive, moderately active, or active[11]
,
[12].
Procedures
The initial evaluation for the collection of data was performed through an interview
and an identification form. The participants were informed about the aims of the study,
and the eligible individuals were subjected to a clinical functional evaluation using
the partial UPDRS and the modified HY scale. Next, the level of physical activity
of the participants was evaluated by a trained examiner using the HAP questionnaire.
Subsequently, subjects were classified into PD subtypes, and the lower limb bradykinesia
score was calculated. Lastly, the FSS was applied to classify the subjects’ fatigue
levels.
Statistical analysis
All analyses were performed using R version 2.13.0 software, and descriptive statistics
was carried out for all the investigated variables. First, simple linear regression
analysis was conducted to verify the relationships between the HAP as a dependent
variable and the various independent variables (age, PD progression time, UPDRS-motor
score, UPDRS-ADL score, modified HY stage, clinical subtypes, lower limb bradykinesia,
and fatigue). Next, multiple linear regression analyses were performed, using forward
selection for the variables. In the first step, it was included in the model the variable
which explained most the variance in the univariate regression, provided it was significant
for a significance level of 0.05. In the following step, an additional variable was
added, provided it resulted in a better fit (in terms of explained variance) compared
to the other still not included variables and that all variables in the model remain
significant. Similar steps followed until the point in which the inclusion of any
other variable would not result in a model with all independent variables statistically
significant.
RESULTS
A total of 46 individuals with PD participated in this study, including 29 men and
17 women, with a mean age of 65.9 ± 12.1 years (range, 42–88 years). The mean time
of disease progression was 11.1 ± 7.4 years (range, 2–38 years). One (2%), 4 (9%),
21 (46%), 6 (13%), 9 (19%), and 5 (11%) individuals were classified as stage 1, 1.5,
2, 2.5, 3, and 4 according to the modified HY scale, respectively. Regarding the clinical
subtypes of PD, twenty-four (52%) patients were classified as RA, 12 (26%) as T, and
10 (22%) as M. Furthermore, 56.5% of the participants showed significant fatigue.
According to the HAP score, 50% of the individuals were classified as active, 19.6%
as moderately active, and 30.4% as inactive.
The FSS, UPDRS-ADL, UPDRS-motor, and HAP scores are shown in [Table 1]. The results from the univariate linear regression analysis are shown in [Table 2]. Age, PD progression time, modified HY stage, UPDRS-ADL, UPDRS-motor, lower limb
bradykinesia, and fatigue all showed significant associations with the HAP score.
Table 1
Scores obtained in the evaluations performed in the study (n = 46).
|
Test
|
Mean (standard deviation)
|
Range
|
|
FSS
|
4 (2.1)
|
1–7
|
|
UPDRS-ADL
|
31.3 (17.4)
|
0–52
|
|
UPDRS-MOTOR
|
14.2 (8.8)
|
0–108
|
|
HAP
|
63.8 (24.2)
|
0–94
|
FSS: Fatigue Severity Scale; UPDRS: Unified Parkinson’s Disease Rating Scale; ADL:
activities of daily living; HAP: Human Activity Profile.
Table 2
Coefficients of univariate regression.
|
Variable
|
Regression coefficient
|
95% confidence interval
|
|
Age
|
-1.08
|
(-1.58, -0.59)*
|
|
PD time
|
-1,37
|
(-2.25, -0.48)*
|
|
HY
|
-23.24
|
(-30.04, -16.44)*
|
|
UPDRS-ADL
|
-2.01
|
(-2.57, -1.45)*
|
|
UPDRS-MOTOR
|
-0.89
|
(-1.21, -0.57)*
|
|
RA
|
-9.91
|
(-23.77, 3.94)
|
|
Tremor
|
4.35
|
(-11.70, 20.41)
|
|
Bradykinesia
|
-3.77
|
(-4.80, -2.74)*
|
|
Fatigue
|
-4.80
|
(-7.86,-1.73)*
|
PD time: Parkinson’s disease progression time; HY: Hoehn and Yahr disability stages;
UPDRS: Unified Parkinson’s Disease Rating Scale; ADL: activities of daily living;
RA: rigid-akinetic. *Significant relationship (p < 0.05) between the variables and
the physical activity level (HAP).
Two models were proposed by the multiple linear regression analysis. The results of
these proposed models are shown in [Tables 3] and [4]. In the first model, age, UPDRS-ADL score, and modified HY stage were found to explain
76% of the variance of the HAP score. In the second model, age, UPDRS-ADL score, and
lower limb bradykinesia were found to also explain 76% of the HAP score variance.
Table 3
Coefficients of regression and 95% confidence intervals (95%CI) of the multivariate
regression analysis, with inclusion of HY stage.
|
Variable
|
Regression coefficient
|
95%CI
|
|
Age
|
-0.81
|
(-1.11, -0.51)*
|
|
UPDRS-ADL
|
-1.20
|
(-1.78, -0.62)*
|
|
HY
|
-10,30
|
(-17.24, -3.36)*
|
|
Intercept
|
159.28
|
|
|
R2
|
0.76
|
|
Regression equation: Y = 159.28 – 0.81 Xi – 1.20 Xu – 10.30 Xh.
With Xi = age; Xu = UPDRS-ADL; and Xh = HY.
UPDRS: Unified Parkinson’s Disease Rating Scale; ADL: activities of daily living;
HY: Hoehn and Yahr disability stages; R2: Explained variance. *p < 0.05.
Table 4
Regression coefficients and 95%CI of the multivariate regression analysis, with inclusion
of lower limb bradykinesia.
|
Variable
|
Regression coefficient
|
95%CI
|
|
Age
|
-0.81
|
(-1.11, -0.50)*
|
|
UPDRS-ADL
|
-1.10
|
(-1.75, -0.45)*
|
|
Lower limb bradykinesia
|
-1.72
|
(-2.94, -0.50)*
|
|
Intercept
|
144.03
|
|
|
R2
|
0.76
|
|
Regression equation: Y = 144.03 – 0,81 Xi – 1.10 Xu – 1.72 Xb.
With Xi = age; Xu = UPDRS-ADL; and Xb = bradykinesia.
UPDRS: Unified Parkinson’s Disease Rating Scale; ADL: activities of daily living;
R2: Explained variance. *p < 0.05.
DISCUSSION
We proposed two models to explain the variance in the level of physical activity in
individuals with PD, because it would be impossible to statistically define which
variable, namely disease severity or lower limb bradykinesia would best predict the
level of physical activity in such individuals. The two models were able to predict
76% of the variance in the level of physical activity, as evaluated by the HAP.
Although few studies have assessed factors affecting physical activity level of individuals
with PD, some authors have indicated factors associated with the disease and/or the
individual. Nimwegen et al.[2] observed that a higher severity of the disease, walking deficits, and an inability
to perform ADL predicted 24% of the variance of the patients’ daily physical activity
level. Our results also indicated the inability to perform ADL and disease severity
as predictors of the level of physical activity in individuals with PD. However, after
adding these factors to age, a higher percentage of this variance was explained (76%).
Furthermore, in a recent study, age, levodopa daily dose, distance covered in the
6-min walk test, and the total UPDRS score were found to predict only 28% of the daily
physical activity level in sedentary PD patients[5]. As in this previous study, we found that age was a determinant of the level of
physical activity. Increased age was associated with increased progression of motor
impairment; decreased responsiveness to levodopa treatment; and more severe postural,
walking, and cognitive deficits in patients with PD[13]
,
[14]
,
[15]. The strong influence of age on the physical activity level and the non-influence
of disease duration, as observed in this study, have been pointed out by many authors[15]
,
[16]. Thus, the decrease in the physical activity level seems to be highly associated
with the immobility and sedentarism of older individuals[16], since the aging process by itself is characterized by a decline in the physical
ability and functional performance[17].
The ability to perform ADL was found to be a predictive factor for the level of physical
activity in patients with PD, as previously reported[2]. A higher loss of functional independence was associated with a higher UPDRS score
and a more severe disease as per the modified HY scale, indicating a positive association
between disease severity and the need for help in ADL and instrumental activities[18].
Disease severity was found to be a predictive factor of the physical activity level
in such individuals in other previous studies[2]
,
[5]. Unlike the results of our study, previous studies[2]
,
[5] noted that the total UPDRS score alone or along with the HY stage were predictive
factors of the level of physical activity in PD patients. The HY scale has been widely
used for the clinical evaluation of PD patients, and describes categories of motor
dysfunction. High correlations have been reported between HY progression and motor
decline and deterioration of quality of life in individuals with PD, confirming its
wide use in the classification of PD severity[8].
To our knowledge, ours is the first study to identify that lower limb bradykinesia
is a predictive factor of the level of physical activity in patients with PD. Bradykinesia
is considered the most debilitating symptom of the disease and is associated with
reduced ability of the patient to perform various functional tasks[1]. In addition, an interaction effect between the duration of the disease and age
on bradykinesia has been previously reported[15]. This indicates that the simultaneous effects of age and disease duration on bradykinesia
are not additive, as they are highly dependent on each other. Various clinical scores,
including the UPDRS-motor score, have been reported to show stronger associations
with velocity than with amplitude of movement[19]. Despite the complex relation between these factors, these previous findings, as
well as the findings from this study, indicate that lower limb bradykinesia is associated
with motor deficits, age, and disease progression.
For the definition of bradykinesia, we used the subscore proposed by Allen et al.[9]. In fact, it is a lower limb bradykinesia subscore derived by summing items of the
UPDRS motor examination. As many studies used the same subscore, it allowed us to
make comparisons between them[9]
,
[19]. Besides that, most of the activities of the HAP (more than 75%) involve mainly
the lower limbs.
To our knowledge, this was the first study assessing subtype classification as a possible
predictive factor of the level of physical activity of patients with PD. Although
the clinical heterogeneity among patients with PD has been extensively debated[7], indicating the existence of clinical subtypes of the disease, this classification
remains controversial and requires a better definition. Accordingly, several authors
have proposed different subtype classifications[10]
,
[20]
,
[21], although a clear and accepted classification has not yet been established. This
indefinition may have influenced with the findings of the present study, and this
requires further investigation.
Finally, fatigue was not a predictive factor of the physical activity level in individuals
with PD. It is known that the physiologic stress of daily physical activities is increased
in patients with PD, and this may contribute to the elevated level of fatigue that
is characteristic of the disease[22]. Thus, we expected that fatigue would influence the physical activity level of such
individuals. However, the relationship between fatigue and the measures of physical
activity is intriguing. Garber and Friedman[23] observed that a higher degree of fatigue was associated with a low level of physical
activity. It is evident that patients with increased fatigue move less, but it is
uncertain whether fatigue is the cause of physical inactivity or if it is the result
of a sedentary lifestyle[23]. According to Elbers et al.[24] the variance in the physical activity level explained by fatigue is small (only
2%). These findings, as well as those from our study, suggest that fatigue may be
an independent symptom of PD that cannot predict the level of physical activity and
that may only represent a minor factor within the numerous factors that affect the
level of physical activity in individuals with PD.
The findings of our study indicate that the level of physical activity is in part
explained by factors that cannot be modified, such as age, but also by factors modifiable
by therapeutic intervention, such as ability to perform ADL and lower limb bradykinesia.
Based on our results, bradykinesia and the identification of difficulties in performing
ADL must be target during physical therapy evaluation, particularly when these factors
are present in older individuals with PD. Therefore, rehabilitation programs should
include functional training of activities and gain of speed, which could make such
individuals more physically active and reduce the adverse consequences of inactivity.
Our study has some limitations. Although most activities related to the physical activity
level involve the lower limbs, upper limb bradykinesia was not evaluated. Moreover,
although no significant difficulty has been observed during the tests, cognition was
not formally evaluated in all participants. Most importantly, we were unable to identify
which variable (modified HY stage or lower limb bradykinesia) better predicted the
level of physical activity in patients with PD. This was likely due to a strong association
between then, with both reflecting disease severity, making it difficult to define
which variable is the best predictor in this study, considering the relatively small
sample size. This interaction may also have occurred due to the methodological bias
between the scales itself. The bradykinesia measure is derived from four questions
of the motor UPDRS, which reflects disease severity, such as the HY scale.
In conclusion, up to 76% of the variance in the physical activity level of individuals
with PD can be explained by age, the ability to perform ADL, and the severity of the
disease (as determined by the modified HY scale), or by lower limb bradykinesia. In
contrast to the expected result, the UPDRS-motor score, disease progression time,
clinical subtypes of the disease, and fatigue did not explain this variance. According
to our results, factors that can be modified by therapeutic intervention were identified
and they must be evaluated and considered when proposing rehabilitation programs for
PD.
