Abstract
Background:
Several scales are available for rating the severity of tremor at present. However, the sensitivity to change of these instruments has remained to be clarified.
Objective:
To compare the sensitivity of the Fahn-Tolosa-Marin Tremor Rating Scale, the Part III of the Movement Disorder Society-sponsored Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) and the MDS-UPDRS Tremor Scale to the effects of various antitremor treatments.
Methods:
Enrolling subjects with parkinsonism associated with tremor, we analyzed two scenarios: (1) tremor changes associated with acute levodopa challenge (n = 287) and (2) a 12-month outcome of different treatment options (n = 512) including deep brain stimulation (n = 146), levodopa/carbidopa intestinal gel infusion (n = 30), and initiating (n = 63) or adjusting oral antiparkinsonian medication (n = 273). Changes in tremor scales were assessed by effect size values (Cohen’s d and eta-square).
Results:
Part B of the Fahn-Tolosa-Marin Tremor Rating Scale was the most sensitive to acute levodopa challenge (Cohen’s d = –1.04, η2 = 0.12). However, Part A of the Fahn-Tolosa-Marin Tremor Rating Scale showed the highest effect size, which was a small one (Cohen’s d = –0.33, η2 = 0.03), for detecting a treatment-related change in the severity of tremor during long-term follow-up.
Conclusions:
The Fahn-Tolosa-Marin Tremor Rating Scale has a better ability to capture changes due to levodopa challenge or antiparkinsonian treatment than MDS-UPDRS Part III or MDS-UPDRS Tremor Scale.
INTRODUCTION
Tremor, defined as a rhythmic and involuntary oscillation of one or more body parts, is a characteristic feature of many neurological syndromes. During the last twenty-five years, serious efforts have been made to develop scales which can reliably measure the severity of tremor. As a result of these efforts, several tools are available for rating tremor at present. The Fahn-Tolosa-Marin Tremor Rating Scale (FTMTRS) [1, 2], developed to assess postural, kinetic as well as rest tremor, is one of the most commonly used tremor rating scales for essential tremor. In addition to the FTMTRS, the Motor Part of the Movement Disorder Society-sponsored Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) [3] and the MDS-UPDRS Tremor Scale [4] can also be used to capture the severity of tremor in parkinsonism.
In 2012, the Movement Disorder Society (MDS) established a task force to review tremor rating scales available to that date [5]. Scales assessing tremor as a part of a global evaluation of Parkinson’s disease (PD), dystonia or ataxia were not analyzed. The task force classified the investigated scales as recommended, suggested or listed based on availability, use, acceptability, reliability, and sensitivity to the progression of disease or to the effects of therapeutic intervention. Among others, the FTMTRS was classified as recommended which means that this instrument is widely used and underwent successful clinimetric testing. The task force did not identify a need for the development of new tremor scales; however, it recommended the additional psychometric validation of existing scales, with a special focus on the sensitivity of scales to change.
A subsequent task force by the MDS aimed to develop a tremor scale based on the tremor items of the MDS-UPDRS to specifically capture tremor in PD [4]. The MDS-UPDRS Tremor Scale was found to be a valid and reliable instrument for assessing the severity of Parkinsonian tremor; however, this task force also recommended the evaluation of responsiveness of this new scale.
The present study aimed to analyze the sensitivity of the FTMTRS, the Part III of the MDS-UPDRS and the MDS-UPDRS Tremor Scale for detecting changes in two scenarios, during acute levodopa-challenge and a 12-month follow-up of mixed therapeutic interventions.
MATERIALS AND METHODS
Study population and protocol
The study protocol was similar to the procedure utilized by Goetz et al. to identify the scale which best detects the response of dyskinesia in PD to an established treatment [6]. The present study was conducted between 2017 and 2019 with the approval of the Regional and Institutional Ethical Committee (3617.316-24987/KK41). After obtaining written informed consents, two partly overlapping consecutive series of patients were enrolled in this prospective study: (1) an acute levodopa challenge; and (2) a long-term follow-up for treating PD. All included patients were treated at the Department of Neurology, Pécs, Hungary and had parkinsonism associated with tremor regardless its nature (e.g., rest or postural-kinetic). Patients without tremor (e.g., rigid-akitenic type of PD without any visually detectable tremor) were excluded. However, those rigid-akinetic PD patients, who had detectable and ratable postural and/or kinetic tremor, were included.
The definitions of the UK Brain Bank for PD and the International Parkinson’s Disease and Movement Disorder Society for parkinsonism had been used for clinical diagnoses prior to initiation of an antitremor treatment. To confirm the diagnosis of PD in some patients included in the longitudinal cohort, acute levodopa challenge and/or DaTSCAN examination were also performed. The diagnosis of PD and other forms of parkinsonism was affirmed and the management of included patients was led by a neurologist specializing and being highly experienced in movement disorders (N.K.).
As a part of the clinical management of the patients, acute levodopa challenge was performed (1) to confirm the diagnosis of PD; and/or (2) to differentiate PD from other Parkinsonian syndromes; or (3) to characterize levodopa response more precisely; or (4) to evaluate the time course and type of dyskinesias induced by levodopa. To achieve a “practical” OFF state in non-de novo patients, any type of levodopa was withdrawn at least 12 hours prior to testing, while long-acting dopamine agonists (e.g., pramipexole, ropinirole, rotigotine) and monoamine-oxidase B-inhibitors (e.g., rasagiline, selegiline) were discontinued at least 48 hours before the challenge. To prevent adverse events such as nausea and vomiting, 10–30 mg/day of domperidone was utilized before the administration of levodopa in de novo or untreated patients. However, if an adverse event interfering with the outcome (e.g., nausea, vomiting, hypotension) occurred during the test despite the efforts made for the prevention of such an event, the patient was excluded. To achieve a definitive ON state, 200–400 mg immediate release formulation of levodopa/benserazide pills (Madopar® Dispersible, Roche) were administered. In some patients, who had been previously chronically treated with antiparkinsonian medications and indicated that the achieved ON state did not correspond to their best ON, further 50–100 mg levodopa was administered and this dose was repeated until the best ON state was achieved.
Another partly overlapping group of patients was examined twice: baseline (ON medication) and 12 months later (follow-up). This group of patients received mixed treatment options to control PD including deep brain stimulation (DBS, n = 146), levodopa/carbidopa intestinal gel infusion (LCIG, n = 30), initiating oral antiparkinsonian medication in de novo patients (n = 63), and adjusting oral medication in treated patients (n = 273). We followed both international [7] and national guidelines [8] on treating the patients. The protocols for DBS [9, 10] and LCIG [11, 12] treatments were described in details elsewhere.
Assessed data
In addition to demographic, medication, and disease-related data, the FTMTRS [2] and the MDS-UPDRS [13] were assessed by highly experienced and trained nurse specialists (É.B. and K.T.). Each patient was examined twice: (1) in the acute levodopa-challenge group in both “practical” OFF and “best” ON states on the same day and (2) in the longitudinal group before the initiation of treatment (baseline, medication ON state) and 12 months later (follow-up, medication ON state). To measure neurocognitive performance, the Montreal Cognitive Assessment (MoCA) [14] was also assessed by a trained neuropsychologist (M.K.) at baseline. Major neurocognitive disorder (MoCA <20.5 points) [14, 15] was an exclusion criterion for participation.
Scales evaluated
The FTMTRS, which was introduced in 1988 [1] and later revised in 1993 [2], consists of three sections. In Part A, the examiner rates the amplitude of several tremor components (postural, kinetic and rest tremor) in 9 different anatomic locations including the face, the tongue, the voice, the head, the trunk, and the extremities. In Part B, the clinician evaluates tremor during writing, drawing and pouring; while Part C assesses patient-reported functional disability. In addition to these, both the clinician and the patient score a global assessment item. Each item of the scale can be rated from 0 to 4. This scale provides a clinician- and patient-rated measure of severity of tremor with good clinimetric properties. However, training is a requirement for good inter-rater reliability, and a potential ceiling effect for upper extremity tremor may also be a weakness of this instrument [5].
The MDS-UPDRS, which was first published in 2008, is the successor version of the Unified Parkinson’s Disease Rating Scale designed by the MDS [3, 16]. This scale, which has four parts, provides a comprehensive picture of PD by measuring several aspects of the disease. Part I, which assesses nonmotor experiences of daily living, consists of two subparts: in Part IA, the clinician assesses a number of non-motor symptoms based on the information provided by the patient and/or the caregiver; while Part IB is a self-administered questionnaire completed by the patient. Questions of Part II are also answered by the patient and focus on the motor experiences of daily living. Part III, which is completed by the rater, extensively measures motor symptoms of PD by assessing not just different tremor components, but also other symptoms of the disease such as bradykinesia, rigidity, disturbances of gait and posture, and postural instability. Part IV rates motor complications of PD based on patient-derived information and the observations of the clinician. All items of the MDS-UPDRS can be scored from 0 to 4. Clinimetric analysis of the scale found that it is a reliable and valid instrument with high internal consistency [3]. Since its development, the scale has been validated in several non-English languages. However, training is needed to facilitate interrater reliability, especially regarding Part III [3].
The MDS-UPDRS Tremor Scale [4], which is based on the already available tremor items of the MDS-UPDRS, was developed to specifically capture important aspects of tremor in PD. The developers performed Rasch analysis on the tremor items of the MDS-UPDRS. After response scale rescoring, creation of super-items and deletion of one item; fit to the Rasch model was obtained. Subsequently, raw scores were converted into an interval scale measure. The authors found that this linear measure is more precise than the use of the raw sum scores. They also concluded that this strictly unidimensional, short scale is able to capture lateralization of tremor; has good construct validity; and provides a valid and moderately reliable measure of tremor unbiased that is by age, sex, and education. However, addition of items measuring low tremor levels and representing patients’ perspectives could resolve the floor effect of this scale and improve its reliability. In addition, its responsiveness to change is also unknown [4].
Statistical analysis
To identify the scale with the best sensitivity for detecting changes in tremor to treatment, the scores of Part A and B of the FTMTRS, the total score of Part III of the MDS-UPDRS and the linear scores of the MDS-UPDRS Tremor Scale were utilized. The detailed methods for calculating the linear scores of the MDS-UPDRS Tremor Scale were described previously [4]. To test normality, the Shapiro-Francia test was utilized. As data from the applied scales followed the normal distribution, repeated measures t-test and analysis of variance (ANOVA) were used to assess statistical significance.
Effect size is a scale-free measure of the relative size of the effect of an intervention suitable for standardized interpreting. To determine the magnitude of treatment effect, three different effect size calculations were made: (1) the effect size correlation coefficient values (r), (2) Cohen’s d for repeated measures, and (3) partial eta-squared (η2). The detailed methods for generating these values were described previously [17, 18]. Generally, r is a standardized measure of the strength and direction of the linear relationship between two variables and Cohen’s d is a representation of the standardized difference between two means. Cohen defined effect sizes [19] as small (r = 0.1 or d = 0.2), medium (r = 0.3 or d = 0.5), and large (r = 0.5 or d = 0.8). Partial eta-squared (η2) is also an effect size measurement for the analysis of variance measuring the strength of the effect on a continuous field. It can be defined as the ratio of variance accounted for an effect and that effect plus its associated error variance within an ANOVA [20]. Based on its value, weak (<0.02), moderate (<0.09) and strong (>0.09) effects can be distinguished for repeated measurements. In addition to effect size, the magnitude of relative change was also determined for each investigated scale in both scenarios.
Our a priori hypothesis was that the scale with the highest effect size can be considered as the most sensitive to treatment response. All statistical analyses were performed using the IBM SPSS software package (version 24.0.2, IBM Inc., Armonk, NY, USA).
RESULTS
Data of 287 patients, who underwent acute levodopa challenge, and 512 patients, who were treated with an established antitremor treatment for PD, were finally analyzed. Sociodemographic, clinical and treatment data of the two cohorts at inclusion are shown in Tables 1 and 2.
Sociodemographic, clinical and treatment data of patients undergoing acute levodopa challenge (n = 287) at enrollment
*These scales were assessed in ON (with medication) state. ET, essential tremor; FTMTRS, Fahn-Tolosa-Marin Tremor Rating Scale; LED, levodopa equivalent dosage; SD, standard deviation; MDS-UPDRS, Movement Disorder Society-sponsored Unified Parkinson’s Disease Rating Scale; PD, Parkinson’s disease.
Sociodemographic, clinical and treatment data of patients being followed (n = 512) at enrollment
ET, essential tremor; FTMTRS, Fahn-Tolosa-Marin Tremor Rating Scale; LCIG, levodopa-carbidopa intestinal gel; LED, levodopa equivalent dosage; SD, standard deviation; MDS-UPDRS, Movement Disorder Society-sponsored Unified Parkinson’s Disease Rating Scale; PD, Parkinson’s disease.
Acute levodopa challenge was positive in 234 cases (81.5%), whereas negative results indicated parkinsonism other than PD in 53 patients (18.5%) of this group (Table 1). In the longitudinal cohort, there were no patients who could be diagnosed with another disorder than PD due to the occurrence of atypical symptoms or ineffectiveness of the initiated treatment during the 12-month follow-up. In patients suffering from PD with preexisting ET, the mean interval between the diagnosis of ET and PD was 7 years.
Scores of all rating scales indicated improvements in severity of tremor by decreasing due to an acute levodopa challenge (Table 3) or an antitremor treatment (Table 4). Levodopa, administered during acute testing, led to larger improvements in severity of tremor in patients with PD compared to subjects with non-PD parkinsonism (Table 3). During 12-month follow-up, initiation of oral antiparkinsonian medications in de novo patients and initiation of bilateral subthalamic DBS led to the largest improvements in the severity of tremor (Table 4).
Changes in severity of tremor during acute levodopa challenge and treatment effect sizes for each investigated scale
*These values were calculated using the following formula: (ScoreON-ScoreOFF)/ScoreOFF*100. **r refers to Pearson’s correlation coefficient between the ON and OFF values. ***During the calculation of these values, previously described formulas were utilized [17, 18]. CI, confidence interval; Cohen’s drm, Cohen’s d for repeated measures; FTMTRS, Fahn-Tolosa-Marin Tremor Rating Scale; MDS-UPDRS, Movement Disorder Society-sponsored Unified Parkinson’s Disease Rating Scale; SD, standard deviation.
Changes in severity of tremor during the first year of antitremor treatment and treatment effect sizes for each investigated scale
*These values were calculated using the following formula: (ScoreFollow - up-ScoreBaseline)/ScoreBaseline*100. **r refers to Pearson’s correlation coefficient between the baseline and 12-month follow-up values. ***During the calculation of these values, previously described formulas were utilized [17, 18]. CI, confidence interval; Cohen’s drm, Cohen’s d for repeated measures; FTMTRS, Fahn-Tolosa-Marin Tremor Rating Scale; MDS-UPDRS, Movement Disorder Society-sponsored Unified Parkinson’s Disease Rating Scale; SD, standard deviation.
Effect size values were small in the non-PD parkinsonism group and moderate to large in the PD group for acute levodopa challenge (Table 3), while they were small for the 12-month antitremor treatment when data of all patients in this group were analyzed (Table 4). However, analysis of longitudinal results disaggregated by type of treatment found that effect size values for initiation of oral antiparkinsonian medications in de novo patients and DBS were moderate to large (Table 4), similarly to effect sizes for acute levodopa challenge in the PD group.
Part B of the FTMTRS and Part III of the MDS-UPDRS were the most sensitive to acute levodopa challenge in patients with PD (Cohen’s d = –1.04, η2 = 0.12 and Cohen’s d = –0.96, η2 = 0.11, respectively), whereas Parts B and A of the FTMTRS indicated changes in severity of tremor due acute levodopa in patients with non-PD parkinsonism most sensitively (Cohen’s d = –0.16, η2 = 0.02 and Cohen’s d = –0.17, η2 = 0.02, respectively). During a 12-month follow-up, Part A of the FTMTRS showed the highest effect sizes when data of all subjects in this group (Cohen’s d = –0.33, η2 = 0.03) and only data of patients treated with DBS (Cohen’s d = –0.82, η2 = 0.14) were analyzed. In the “de novo” patients group, however, Part III of the MDS-UPDRS was the most sensitive to change due to the initiation of oral antiparkinsonian treatment (Cohen’s d = –0.31, η2 = 0.02; Table 4).
DISCUSSION
To identify the scale which best detects the response of tremor to treatment, the present study applied three instruments to measure changes in severity of tremor of patients undergoing an acute levodopa challenge and/or receiving an antitremor treatment during a 12-month period. Three different effect size values (Pearson’s correlation coefficient, Cohen’s d and partial eta-squared) were calculated for each scale, and the scale with the highest effect sizes were considered the most sensitive to change in severity of tremor.
Based on our data, all scales showed satisfactory sensitivity to change. However, the objective parts of the FTMTRS (Part A and B) had the highest effect size values in both of the scenarios we examined. A possible explanation for the highest sensitivity to change of the FTMTRS could be that this tool provides a unique opportunity for a specific assessment of tremor. In contrast, Part III of the MDS-UPDRS also measures other motor symptoms of parkinsonism than just tremor. These symptoms of parkinsonism can also improve due to treatment, and this may have an impact on sensitivity to change in tremor of this scale. Nonetheless, the MDS-UPDRS-based Tremor Scale was also inferior to the FTMTRS in sensitivity for detecting changes in the severity of tremor. This may result from the fact that the FTMTRS is a more detailed scale measuring tremor also in specific body parts (e.g., tongue, head, voice, and trunk) and during activities which are not scored by the other investigated scales.
Effect size values showed that the effect of acute levodopa on the severity of tremor was more pronounced than that of the 12-month antitremor treatment. Less marked longitudinal changes after significant acute changes have also been observed in some previous studies investigating other aspects of parkinsonism [21, 22] or other movement disorders [23]. Patients of the present study who were followed received several treatment options leading to different levels of improvement in severity of tremor. Effects of every treatment option may have partly been responsible for the magnitude of the overall change achieved. However, most of the patients in this group were followed after the adjustment of oral medications which was associated the lowest effect size values, and lower number of patients were followed after the initiation of oral medications and received DBS, which had higher effect size values. This may have decreased the magnitude of the overall change in the severity of tremor in this group. The low effect sizes of the adjustment of oral medications may partly result from that all scales were assessed in ON state at both baseline and 12-month follow-up, the adjustment of oral treatment, therefore, resulted in only small changes in scores. However, previous studies also found efficacy of LCIG [24, 25] and DBS [24] superior to conventional optimized medical treatment in PD which is consistent with the effect size values found by the present study. Another possible explanation for the less marked longitudinal changes compared to acute changes induced by levodopa may be that movement disorders progress and efficacy of treatments may decrease over time [22, 27].
The strength of the present study partly lies in the high number of enrolled patients with parkinsonism resulting from several etiologies and representing a wide range of disease severity. This may be of high importance in longitudinal studies investigating progressive disorders, because disease progression, which should also be taken into account for the correct interpretation of results, seems to depend more on the baseline characteristics of the study cohort than on the duration of the follow-up period [28]. The study provides useful information about the efficacy of different antitremor treatments measured by several instruments for rating tremor. The calculation of effect size values and relative changes highly improve interpretability of our results. However, the relatively small number of patients who were followed after the initiation of oral medications and treated with LCIG is a limitation of our study. Another limitation may be that the sample was enrolled in a single movement disorders unit, not randomly selected in cooperation with other centers.
To conclude, our study provides data in favor of the FTMTRS as the preferred primary outcome in clinical research focusing on tremor and its treatment. The present results move the comparative scale program established by Goetz et al. [6] forward, and they may also help the selection of the most eligible primary outcomes in future clinical trials investigating movement disorders associated with tremor. However, instruments assessing patient-reported changes in severity of tremor may also be helpful in evaluating effectiveness of different medical interventions. Their sensitivity to change should be examined in future studies.
CONFLICTS OF INTEREST
The authors have no conflict of interest to report.
Footnotes
ACKNOWLEDGMENTS
We would like to thank Márton Kovács, Éva Balázs and Katalin Takács for their assistance in study-related examinations. This study was supported by the Hungarian Brain Research Program (2017-1.2.1-NKP-2017-00002), NKFIH EFOP-3.6.2-16-2017-00008, NKFIH SNN125143 and the New National Excellence Program (ÚNKP-17- 4 -I.- PTE-311) government-based funds. Our research was partly financed by the Higher Education Institutional Excellence Program of the Ministry of Human Capacities in Hungary, within the framework of the 5th thematic program of the University of Pécs, Hungary (20765/3/2018/FEKUSTRAT).