Quality of life predicts rehabilitation prognosis in Parkinson's disease patients

Abstract Background Rehabilitation has been reported to improve the quality of life (QoL) of patients with Parkinson's disease (PD). Nevertheless, not all patients are satisfied with rehabilitation outcomes and could achieve a significant improvement in QoL. Objective To detect possible predictors of QoL improvement in patients with PD after rehabilitation. Methods A total of 86 PD patients were included and followed up for 3 months with a 39‐item Parkinson's Disease Questionnaire summary index (PDQ‐39 SI) as the primary endpoint. All patients received 2 weeks of multidisciplinary intensive rehabilitation treatment (MIRT). Changes in patients' QoL were assessed using the PDQ‐39 at baseline and at the 3‐month follow‐up. The reliable change index (RCI) was adapted to determine the individual QoL outcome. The predictors of QoL outcome were detected using logistic regression analysis. Results After a 3‐month follow‐up, PDQ‐39 SI decreased significantly from 22.95 ± 9.75 to 18.73 ± 10.32 (P < 0.001). Scores for QoL improved (RCI>10.9) after rehabilitation for 18.6% of the patients, and 74.4% of patients reported an unchanged QoL (−10.9≤RCI≤10.9), while 7.0% of patients reported a worsening of QoL (RCI<−10.9). Among the baseline parameters, the PDQ‐39 SI was a baseline predictor for changes in QoL in the logistic regression model (OR: 1.15, CI: 1.07–1.24, P < 0.001). Conclusions MIRT could improve QoL for some patients with PD, and PDQ‐39 score at baseline is the most important predictor for QoL improvements after rehabilitation for this patients.


INTRODUCTION
Parkinson's disease (PD) is a progressive neurodegenerative disease that is characterized by motor symptoms such as bradykinesia, resting tremor, rigidity, and a sequence of nonmotor symptoms, such as cognitive decline, anxiety, and depression (Bloem et al., 2021). It can lead to a decline in quality of life (QoL) very early and deteriorates as the disease progresses (Carod-Artal et al., 2007), even under the best medical treatment or with deep brain stimulation .
Therefore, rehabilitation therapies are considered an adjuvant to pharmacological and surgical treatments to maximize functional abilities and to minimize secondary complications (Armstrong & Okun, 2020).
Meanwhile, rehabilitation has also been reported to improve the QoL related to PD (Clarke et al., 2016;Ferrazzoli et al., 2018;Monticone et al., 2015;Morris et al., 2009;Rodrigues De Paula et al., 2006;Tickle-Degnen et al., 2010), but not all studies yield the same results (Wade et al., 2003). Although several reasons could explain the lack of benefit, such as previously satisfactory management leaving little room for improvement, lack of psychological intervention, etc. (Playford, 2003).
Therefore, this revealed that not all patients are satisfied with the rehabilitation results and can achieve a significant improvement in QoL.
The reasons may be multidimensional, but strategies to identify which patient could benefit from rehabilitation seem to be important for optimized treatment outcomes. For instance, we can adopt different rehabilitation strategies for patients with different outcomes to reduce the waste of medical resources and improve the therapeutic effect of individual patients. On the other hand, for those who may benefit less, other treatment methods should be actively adopted to improve the patient's QoL as much as possible.
Therefore, we performed this study to understand the effects of multidisciplinary intensive rehabilitation treatment (MIRT) for mild to moderate PD patients. In particular, the relative contributions of motor and nonmotor baseline parameters to the rehabilitation outcome on QoL were analyzed. We attempted to detect the possible predictors of QoL improvement for PD patients after rehabilitation with a 3-month follow-up.

Study population
In this ongoing, prospective single-center cohort study (multidisciplinary rehabilitation registration study on PD, registration number: ChiCTR2000033768), a total of 86 PD patients attending our inpatient rehabilitation project in Beijing Rehabilitation Hospital from June 2020 to July 2021 were included and followed up for 3 months. All patients received 2 weeks of MIRT. None of the patients underwent drug adjustment from training to 3 months of follow-up. Inclusion criteria were: (1) idiopathic PD as confirmed by a neurologist using the Movement Disorder Society criteria (Postuma et al., 2015); (2) no deep brain stimulation or in vivo implantation treatment; (3) were able to understand each item of the informed consent and willing to sign the informed consent.
The exclusion criteria were as follows: (1) atypical parkinsonism, such  (Hoehn & Yahr, 1967), and MDS-UPDRS score (Goetz et al., 2007). The evaluation of the H&Y stage and MDS-UPDRS score were carried out under the medication ON stage.

MIRT procedure
MIRT is specifically designed for rehabilitation for PD patients and is constituted with a multidisciplinary, aerobic, motor-cognitive, intensive, and goal-based rehabilitation treatment (Ferrazzoli et al., 2018;Frazzitta et al., 2012;Frazzitta et al., 2015). All patients received a 2 weeks of MIRT. The 2-week program comprised four sessions.
The first session is one-on-one physical therapy by a physical ther-

Clinical assessments
QoL was assessed by the Chinese version of the Parkinson's Disease Questionnaire (PDQ-39) (Neff et al., 2018). The PDQ-39 is a validated disease-specific HRQoL measure in PD (Neff et al., 2018) Jacobson & Truax, 1991). A PDQ-39 score decrease of more than 10.9 points was termed "improved," and a score increase of more than 10.9 points was termed "worsen"; changes between −10.9 and 10.9 were termed "unchanged" ( Witt et al., 2011). Both the "worsen" and "unchanged" groups were assigned to the "nonimproved" group.

Statistical analysis
Data distribution and normality were evaluated with the Shapiro-Wilk test. Normally distributed data are expressed as the mean ± standard deviation (SD), while nonnormally distributed data are reported as the median (interquartile range). The PDQ-39 SI and subgroup score at baseline and with follow-up for 3 months were compared by the paired sample t test. Spearman correlations or Pearson correlations were used to explore the relationship between changes in QoL scores and demographic and clinical parameters, including age, age at diagnosis, disease duration, LEDD, H&Y stage, MDS-UPDRS score, GDS, MAES, PFS-16, and PDQ-39 SI score at baseline. Spearman's correlation was also used to evaluate changed nonmotor parameters, including GDS, MAES, and PFS-16 scores, associated with changes in QoL.
Furthermore, demographic and clinical parameters at baseline were compared between the "improved" and "nonimproved" groups according to the PDQ-39 RCI using univariate analysis. Demographic and clinical parameters at baseline as revealed by the first-step analyses (less strictly selected: P < 0.10, one-tailed) were included in a logistic regression analysis as independent variables with the dichotomous variable improved/nonimproved in terms of the PDQ-39 RCI as the dependent variable. P values < 0.05 (two-tailed) were regarded as significant. Receiver operating characteristic (ROC) analyses were used to evaluate the predictive utility of logistic regression models on patients' QoL improvement (Schrag et al., 2017). These statistical data were analyzed using SPSS version 21 (SPSS Inc., Chicago, IL).

RESULTS
Of 155 patients screened, 86 patients (36 males) were included in the final analysis (see Figure 1). The mean age at baseline was 60.24 years (SD = 7.82), and the median disease duration was 6.00 years (interquartile range = 4.00). The patient baseline characteristics are shown in Table 1.  (P < 0.001). Bodily discomfort, social support, and activities of daily living showed no significant changes. In the "improved" group, the bod-ily discomfort (P = 0.029), cognition (P = 0.002), stigma (P = 0.002), emotional wellbeing (P < 0.001), activity of daily living (P = 0.002), and mobility (P < 0.001) domains decreased significantly. There is a downward trend in the social support and communication domain. In the "nonimproved" group, the mobility domain decreased significantly (P < 0.001); however, there was no significant change or increase in other domains. The PDQ-39 subgroup analysis was shown in Figure 3.

Correlation analyses
PDQ-39 SI at baseline was significantly associated with changes in the PDQ-39 SI after a 3-month follow-up (r = 0.422, P<0.001). Meanwhile, the changes in the PDQ-39 SI were significantly associated with GDS score changes (r = 0.537, P < 0.001) and MAES score changes (r = 0.284, P < 0.01) after a 3-month follow-up. The correlation analysis results are shown in Appendix 1.

F I G U R E 3
The effects of MIRT for mild to moderate PD patients in the PDQ-39 subgroup. The comparison of PDQ-39 subgroup scores for all patients, the unimproved group, and the improved group at baseline and after rehabilitation is separately shown on the left of the dotted line, between the two dotted lines, and on the right of the dotted line. The cognition, stigma, emotional wellbeing, and mobility domains decreased significantly, and the communication domain increased significantly. In the "improved" group, bodily discomfort, cognition, stigma, emotional wellbeing, activities of daily living, and mobility domains decreased significantly. In the "nonimproved" group, the mobility domain decreased significantly, and there was no change or increase significantly in other domains (T, total; NI, nonimproved; I, improved. *P < 0.05, **P < 0.01, ***P < 0.001). MIRT, multi-disciplinary intensive rehabilitation treatment  showed that the same is not true for all PD patients. QoL improved steadily in only a fraction of patients with a relatively conservative index to define the change in QoL after rehabilitation . Therefore, we developed a model that could be used to predict improvement in QoL after rehabilitation in PD patients. We found that impaired QoL was the most important predictor of benefit in mild to moderate PD after rehabilitation in our prognostic model. Furthermore, we found that changes in the PDQ-39 SI were significantly associated with PDQ-39 SI at baseline and were significantly associated with GDS score changes and MAES score changes.

Evidence of the positive effects of rehabilitation treatments in PD is
increasing (Clarke et al., 2016;Ferrazzoli et al., 2018;Monticone et al., 2015;Morris et al., 2009;Rodrigues De Paula et al., 2006). Ferrazzoli et al.'s study (Ferrazzoli et al., 2018) suggested that MIRT, which consisted of a multidisciplinary, aerobic, intensive, and motor-cognitive rehabilitation treatment, could improve QoL in patients with PD that could last for 3 months (Ferrazzoli et al., 2018). Monticone et al.'s study (Monticone et al., 2015) found that PD patients could obtain a benefit from multidisciplinary rehabilitative care for at least 1 year after the intervention. Multidisciplinary rehabilitative care could change the course of motor impairment, balance, activities of daily living, and QoL.
Morris et al.'s study (Morris et al., 2009) suggested that inpatient rehabilitation produces short-term reductions in disability and improve-ments in QoL in people with PD for at least 3 months. As mentioned above (Ferrazzoli et al., 2018;Monticone et al., 2015;Morris et al., 2009), our finding also suggest that the improvement of QoL was considered to be significant by MIRT in mild to moderate PD patients. QoL scores significantly improved for 18.6% of the patients with a more conservative method . Another study showed that QoL improved in only 7% of patients who received the best medical treatment with the same definition of improvement   et al., 2008). Rehabilitation therapy could improve the motor and nonmotor symptoms of PD (Lamotte et al., 2015;Rafferty et al., 2017), and consistent with previous reports (Rafferty et al., 2017), patients with advanced PD could obtain more benefit than those with mild PD from rehabilitation therapy.
There are some limitations to this study that must be acknowledged.
First, an important limitation of our findings is the selected patient population. We collected only mild to moderate PD patients, which might lead to selection bias. PD patients in the H&Y:4-5 stage could not complete MIRT due to their serious condition, and they may need to adopt other rehabilitation strategies, so we did not include those patients.
Second, due to the COVID-19 outbreak, patients were unable to get to the hospital for a detailed kinematic evaluation, such as UPDRS-III score and UPRDS part IV score, at the 3-month follow-up, which could reflect the improvement of the patient's motor function and motor complications (Goetz et al., 2007). Last, our follow-up time was only 3 months. Although there is evidence that PD patients can benefit from rehabilitation treatments, it is unclear how long the benefits could last (Abbruzzese et al., 2016). A long-term observation still is needed.
In conclusion, the present study shows that MIRT could improve the QoL of PD patients, especially mild to moderate PD patients with impaired QoL. Impaired QoL at baseline could predict the benefit of rehabilitation for PD patients. Patients with impaired QoL should actively participate in rehabilitation exercises to improve their QoL.
Further research is needed in more PD patients.