Factors associated with motor complications in Parkinson's disease

Abstract Objectives Levodopa is the most effective therapy for treating Parkinson's disease (PD); however, side effects such as dyskinesias and motor fluctuations may occur after some years of its usage. The aims of this study were to assess the frequency of and factors associated with motor complications among PD patients on levodopa treatment. Methods In a cross‐sectional study carried out in 2010–2013, clinical data and treatment details were collected. Logistic regression expressed by odd ratios (OR) and 95% confidence intervals (CI) was conducted to examine the effects of several independent variables on the occurrence of motor complications. Results A total of 455 patients were enrolled, among whom 374 were on levodopa. Analysis was performed in 328 patients whose exact duration of levodopa treatment was known. Among patients included in the analysis, 25.9% experienced motor complications; of these, 21% had dyskinesias and 20.1% had motor fluctuations. Based on logistic regression, statistically significant factors associated with the occurrence of motor complications were younger age at onset of the disease, higher levodopa equivalent daily dose (LEDD), shorter time to levodopa initiation, and akinetic‐rigid dominant phenotype of PD. Conclusions This study suggests that postponing the start of levodopa therapy and maintaining low daily doses of levodopa might reduce the risk of motor complications. Our results confirm that due to higher risk of motor complications, effectively treating patients with akinetic‐rigid dominant phenotype of PD might be more challenging than for patients whose dominant symptom is tremor.

usually starts at night or early in the morning. Motor fluctuations are alterations of clinical state with "off"-periods manifesting in worsening of motor symptoms (Aquino & Fox, 2015;Vijayakumar & Jankovic, 2016).
Based on the findings of various observational surveys using different study designs, the frequency of dyskinesias ranges from 26% to 44% and that of motor fluctuations from 22% to 64% (Nicoletti et al., 2016;Yoritaka et al., 2013;Martínez-Martín et al., 2014;Hashim et al., 2014;Larsen, Karlsen, & Tandberg, 2000;Schrag & Quinn, 2000). Several studies have shown the prevalence of motor complications to increase with the duration of the disease (Nicoletti et al., 2016;Hashim et al., 2014;Schrag & Quinn, 2000;Scott, Macleod, & Counsell, 2016;García-Ruiz, Del Val, Fernández, & Herranz, 2012;Stocchi et al., 2014). Motor complications may affect more than half of patients within 5 years of diagnosis (Bjornestad et al., 2016) and by the end of the first decade, up to 90% of patients may exhibit dyskinesia and 60% motor fluctuations (García-Ruiz et al., 2012). One of the longest follow-up studies demonstrated that by the end of the second decade, motor complications were present in 100% of patients who were on at least 300 mg levodopa per day (Hely, Reid, Adena, Halliday, & Morris, 2008). Risk of development of motor complications may also rise in a dose-dependent manner (Schrag & Quinn, 2000;Scott et al., 2016;Warren Olanow et al., 2013) but also in a levodopa duration-dependent manner (Hashim et al., 2014;Schrag & Quinn, 2000).
Additionally, a low body weight (Warren Olanow et al., 2013) presence of tremor at diagnosis (Scott et al., 2016), higher levodopa equivalent dose, (Nicoletti et al., 2016) and higher MMSE score (Scott et al., 2016) have been shown to predict a higher rate of dyskinesias, while younger age (Stocchi et al., 2014;Bjornestad et al., 2016) and UPDRS Part III score (Bjornestad et al., 2016) predict a higher rate of motor fluctuations.
The association between the emergence of motor complications and the time from the first symptoms of PD until levodopa initiation has been the subject of debate. There is growing evidence suggesting that delaying the start of levodopa therapy might not be associated with smaller risk of motor complications in the long term (Scott et al., 2016;Cilia et al., 2014). The relationship between motor complications and distinct clinical motor phenotypes of PD has not been studied enough. There is some evidence that tremordominant subtype may indicate a lower probability of motor complications (Nicoletti et al., 2016). To extend the understanding between motor complications and a range of clinical factors, the current study was carried out.
The objectives of this paper were (1) to describe the frequency of motor complications among patients with PD who receive levodopa treatment and (2) to assess the effect of various demographic and clinical factors on the occurrence of motor complications.

| Study design and patients
A cross-sectional observational study was conducted in three medical centers in different regions of Estonia (Tartu, Pärnu, and Tallinn) between 2010 and 2013. As it was primarily an epidemiological study, all the patients in the region were included, representing different stages and phenotypes of the disease. PD was diagnosed based on the diagnostic criteria of the Queen Square Brain Bank Criteria (Gibb & Lees, 1988;Lees, Hardy, & Revesz, 2009). No specific exclusion criteria were set. From the initial cohort of 486 parkinsonian patients, 16 were excluded due to correction of diagnoses including other tremors, secondary or atypical parkinsonism, or Lewy body dementia, and 15 more were excluded due to refusal or incomplete testing. A total of 455 patients were included in the final analysis.
Information from all available sources was obtained for the case ascertainment. A majority of the study participants were outpatients (n = 390). A few were inpatients (n = 14), in nursing homes (n = 23) or visited at their own homes (n = 28). Ethical approval was obtained from the Research Ethics Committee of the University of Tartu. All patients provided signed informed consent.

| Data collection and outcome measures
A semistructured interview based on a special case report form containing items on demographic and clinical data was performed, along with a neurological examination. Total levodopa equivalent daily dose (LEDD) was calculated using the standardized conversion formula (Tomlinson et al., 2010). The data on the prevalence of motor complications was obtained from two sources: patient cards and Part IV of the MDS-UPDRS (Goetz et al., 2008). Patient cards were provided with binary questions assessing motor fluctuations, dyskinesias, and off-dystonia that had occurred at any time during the disease. All items in the MDS-UPDRS are scored on a scale of 0 to 4 (normal/mild/slight/moderate/severe). The presence of motor complications over the past week was indicated by a score of ≥1 on MDS-UPDRS items 4.1 and 4.6 (off-dystonia was included among dyskinesias) and item 4.3 (motor fluctuations). All other parts of the MDS-UPDRS were also performed: Part I-nonmotor aspects; Part II-motor experiences of daily living; and Part III-motor examination.
During the neurological examination, motor phenotype was determined based on the most prevalent symptom: the dominance of tremor with other motor symptoms of only a mild level; nontremordominant PD which includes syndromes with akinesia and rigidity; or postural instability and gait disorder (PIGD). The clinical stage was assessed using the Hoehn and Yahr scale (HY) (Hoehn & Yahr, 1967).
To measure depression, the Beck Depression Inventory (BDI) was performed (Beck, Ward, Mendelson, Mock, & Erbaugh, 1961). For the screening of cognitive impairment, the Mini Mental State Examination (MMSE) was performed (Folstein, Folstein, & McHugh, 1975). For the purpose of evaluation of quality of life, the Parkinson's Disease Quality of Life Questionnaire (PDQ-39) was used (Peto, Jenkinson, Fitzpatrick, & Greenhall, 1995).   Table 1 shows, prevalence of motor complications was relatively low, earlier in the course of PD, but increased with the duration of levodopa treatment.

| Characteristics of patients
The clinical profile of the entire study sample, as well as the group comparisons of those with and without motor complications, is depicted in Table 2. The ratio of females to males was similar in patients with and without motor complications (approximately 60% of females and 40% of males in both groups). Patients with motor complications were younger than patients without motor complications (71.7 ± 8.9 vs. 75.2 ± 8.2 years; t-test, p = .002). The differences were most evident when regarding age at PD onset and duration of the disease. The outcomes of most of the clinometric scales differed more or less between patients with and without motor complications. No differences were seen in the severity of PD assessed by HY, with a median value of 3 in both groups. No statistically significant differences were found in the MMSE scores, with the median value of 27 among those without and 28 among those with motor complications (Mann-Whitney U test, p = .08). Duration of levodopa treatment was higher among patients with motor complications compared to those without (8.8 ± 5.3

Patients on levodopa therapy
Motor fluctuations Dyskinesias

| Logistic regression analysis
The results of the logistic regression can be found in Table 4. and having akinesia-rigid phenotype of the disease (as opposed to the dominance of tremor) increased the odds by 156% (p = .033).
To confirm that the effect of clinical phenotype and LEDD is not due to possible confounders (duration of PD and duration of levodopa treatment), we conducted an analysis using the second model, in which we adjusted for these two possible confounders. LEDD (OR = 0.003; 95% CI: 1.001 to 1.004, p = .0004) and akinesia-rigid phenotype as opposed to tremor-dominant phenotype (OR = 2.549; 95% CI: 1.1 to 6.2, p = .04) still emerged as significant independent variables associated with the occurrence of motor complications with almost the same OR coefficients as in the first model.
As a post hoc analysis, we examined the LEDD of patients stratified by age and found that younger patients tended to be on a higher LEDD than older patients.

| DISCUSSION
This study aimed to address two main questions. Our first goal was to assess the prevalence of motor complications among a representative sample of PD patients in Estonia. Our second aim was to examine T A B L E 3 Antiparkinsonian treatment of PD patients with and without motor complications the association between multiple factors and the development of the disabling side effects of levodopa.
With respect to the first research question, we found a slightly lower frequency of motor complications in our patients than several other studies (approximately 20% of our study participants with either of the motor complications vs. more than 22% of motor fluctuations and more than 26% of dyskinesias in other studies) (Nicoletti et al., 2016;Yoritaka et al., 2013;Martínez-Martín et al., 2014;Hashim et al., 2014;Larsen et al., 2000;Schrag & Quinn, 2000). It is possible that the somewhat lower frequency of motor complications was partly due to the profile of the study participants. Regarding the usage of levodopa, the mean daily levodopa dose of 434 mg in our study was lower than that reported in other studies (500 mg in the study by Hashim et al. and 548 mg in the study by Yoritaka et al.) (Yoritaka et al., 2013;Hashim et al., 2014). As this study was part of an epidemiologic survey of PD, all patients in the region were enrolled, included early cases. Among patients with early stages of PD who have been on levodopa therapy for a short time, motor complications are rather unexpected, a phenomenon demonstrated in Table 1. This finding contrasts with the results of Stocchi et al. (Stocchi et al., 2014); in which more than half of patients under levodopa therapy for 1 to 2 years were diagnosed as having wearing-off symptoms, whereas only a mi- to be somewhat lower than in other countries, levodopa daily dose may underestimate the total dopaminergic load. Furthermore, approximately half of patients (57%) with motor complications were on dopamine agonists, which were used more than twice as than they were by patients without motor complications (Mann-Whitney U, p < .0001).
Our findings on LEDD emerged from the multivariate regression analysis as a significant predictor of motor complications that was not found for the levodopa daily dose. It may support our observation that the levodopa dose alone may underestimate the cumulative dopaminergic burden. Although, motor complications are traditionally considered to result from long-term use of levodopa, (Aquino & Fox, 2015) the total burden of antiparkinsonian treatment may play a significant role in the emergence of motor complications.
One of the key findings of this study is that a shorter duration of disease until the initiation of levodopa appears to increase the odds of motor complications. This finding is supportive of a few studies indicating that early initiation of levodopa might be a risk factor for motor complications (Schrag & Quinn, 2000;Denny & Behari, 1999) but contrary to some other surveys suggesting that delaying the start of levodopa therapy is not associated with a smaller risk of motor complications in the long term (Scott et al., 2016;Cilia et al., 2014). Our results support the opinion that younger age at onset of PD seems to be a risk factor for motor complications per se (Hashim et al., 2014; T A B L E 4 Factors associated with motor complications in PD: the logistic regression model García-Ruiz et al., 2012;Bjornestad et al., 2016;Warren Olanow et al., 2013). Post hoc analysis revealed that the younger patients tended to be on a higher LEDD than older patients. They might be more prone to the higher usage of dopaminergic treatment in order to reduce their parkinsonian symptoms than older patients as a result of having more commitments concerning their employment status and families than older patients. Levodopa may be postponed in patients who are not troubled by their PD-related motor symptoms and in patients with young-onset PD who are at the highest risk for developing levodoparelated complications (Jankovic & Poewe, 2012). At the same time, the possibility of the different effects of dyskinesias and motor fluctuations on quality of life should be taken into account. Some studies have found that PD patients with fluctuations have lower quality of life than patients without fluctuations (Stocchi et al., 2014;Skorvanek et al., 2015) but no such association was found between quality of life and dyskinesias in another study (Hechtner et al., 2014). In our survey, a majority of patients with dyskinesias reported that drug-induced complications impair their functional ability only mildly, and severe dyskinesias were rarely described. This finding may provide some support for the possibility that patients themselves might not be particularly troubled by their dyskinesias, at least when those features are mild. In summary, we conclude that the traditional clinical approach of balancing efficacy and risk of motor complications is needed for each patient, using the lowest dose of levodopa that will provide satisfactory clinical control (Warren Olanow et al., 2013). Still, improved quality of life should be the ultimate goal of our treatments.
In disagreement with the data of recent studies, we did not find female gender to be associated with higher occurrence of motor complications (Yoritaka et al., 2013;Scott et al., 2016;García-Ruiz et al., 2012;Stocchi et al., 2014;Bjornestad et al., 2016;Warren Olanow et al., 2013). It has also been suggested that the severity of motor PD symptoms predicts a higher risk of developing motor complications (Bjornestad et al., 2016). However, this does not appear to be the case in our study, supporting the results of another study based on a cohort of incident PD patients (Scott et al., 2016). Although, we found more severe MDS-UPDRS Part II scores for those experiencing motor complications, according to logistic regression analysis, everyday activities were not significantly associated with the presence of levodopa-induced side effects, which does not support the observation by one other study based on a cohort of early PD patients (Warren Olanow et al., 2013).
One rather important finding was that the akinetic-rigid phenotype of PD emerged as a significant independent associated factor in the higher occurrence of motor complications. This association remained significant even after including possible confounders, such as duration of PD and levodopa therapy, to the second model.
Patients with nontremor-dominant types of PD are characterized as being more often depressed and having a more severe clinical picture, higher LEDD, and longer disease duration than patients with tremor-dominant subtype (Burn et al., 2012). In a recent casecontrol study in Italy investigating the relationship between clinical phenotype and the risk of developing dyskinesias, a significant negative association between tremor-dominant phenotype as an initial PD manifestation and levodopa-induced dyskinesia was established (Nicoletti et al., 2016). Another case-control study conducted in Finland investigated differences in the binding of striatal dopamine transporter and the extent of caudate dopamine terminal loss, and the consequent dopamine function was relatively more preserved in PD patients with tremor compared to akinetic-rigid patients (Kaasinen, Kinos, Joutsa, Seppänen, & Noponen, 2014). Our study allows us to confirm information about the clinical heterogeneity of distinct PD phenotypes. Due to the higher risk for motor complications, achieving effective treatment results in patients whose dominant symptoms are akinesia and rigidity might be more challenging than in patients with the tremor-dominant disease.
Our study has several strengths. First, the case identification in our study was based on all available sources in the area, and we accordingly enrolled a relatively large sample of patients with PD, including early cases as well as both moderately and severely ill patients. Thus, the results may be extrapolated to the Estonian PD population as a whole. Secondly, the evaluation of the participants was thorough, and wide-ranging aspects (motor, nonmotor, functional, cognitive, and emotional) were assessed based on validated clinometric scales and questionnaires. Therefore, we believe that the measurements are accurate and reflect the profile of study participants correctly.
Some limitations of this study need to be taken account when interpreting the findings. The main weakness of the study is a relatively small sample size of patients with motor complications, which can be a source of low power for the logistic regression analysis. The second limitation might be a potential underestimation of motor complications with subtle presentations, especially among patients with impaired cognitive ability. Finally, due to the cross-sectional design of our study, each patient was assessed once, so no pattern of progression of the disease could be estimated and causes and effects could not be determined.
In conclusion, the most important findings of this study were that patients receiving a higher LEDD and having akinetic-rigid dominant phenotype as opposed to tremor-dominant phenotype of the disease are at greater risk of having motor complications. The results of the research do not support the idea that levodopa should be initiated as early as possible; instead, postponing the start of levodopa, together with prescribing optimal daily doses, might reduce the odds of motor complications.

ACKNOWLEDGMENTS
The study has been supported by the Grants GMVCM1239P and IUT2-4 of the Estonian Research Council.

CONFLICT OF INTERESTS
The authors declare that they have no financial or non-financial competing interests.