Study design

We conducted a ten-week, unblinded trial of twice-weekly, structured, non-contact karate classes tailored for individuals with mild- to moderate-stage PD following the Transparent Reporting of Evaluations with Nonrandomized Designs (TREND) statement [14]. Our primary objective was to evaluate the feasibility and acceptability of community-based karate classes among individuals with mild- to moderate-stage PD. We also hypothesized that karate may offer positive effects on mobility, balance, and quality of life in this population. The feasibility and acceptability of this exercise intervention were measured via participant withdrawal/dropout, adverse effects and falls, and class attendance. As exploratory secondary outcomes, we assessed quality of life, patient global impression of change, and mobility at pre- and post-intervention study visits. We assessed sustainability via six-month follow-up phone calls regarding continuation of karate practice. The Institutional Review Board at Rush University Medical Center approved this study (Clinicaltrials.gov: NCT03555695). Written informed consent was obtained from each participant.

Participants

We recruited individuals with PD from the Rush University Parkinson’s Disease and Movement Disorders Program and other Chicago-area institutions between March-June 2018. Recruitment methods included direct outreach to intramural and extramural neurologists, senior centers, and libraries throughout Chicago. We placed recruitment postings in institutional newsletters and fliers in the neurology offices. Within Rush, we queried our Institutional Review Board-approved data repository for a list of English-speaking individuals diagnosed with PD.

Inclusion criteria were as follows, gleaned first from chart review and subsequent telephone screening: PD diagnosed according to UK Parkinson’s Disease Society Brain Bank Clinical Diagnostic Criteria [15]; between 30 and 90 years of age; English-speaking; seen as an outpatient within the last two years; and able to attend classes twice weekly at one of two karate studios located in the Chicago suburbs. Study team members (JF, BS, CN, ML, CW) prescreened and contacted potential participants based on the inclusion criteria above, utilizing the most recent outpatient visit. For those recruited through community outreach, all were screened for eligibility and we required confirmation of PD diagnosis and Hoehn & Yahr stage from the referring neurologist prior to study enrollment. We included participants regardless of treatment status, including pharmacologic or surgical treatment for PD. Participants were encouraged but not required to maintain their treatment regimen for the 10-week study.

Exclusion criteria were as follows: Individuals living >20 miles from participating karate studios; a diagnosis of atypical parkinsonism, including Progressive Supranuclear Palsy, Multiple System Atrophy, Dementia with Lewy Bodies, Corticobasal Syndrome, drug-induced or vascular parkinsonism, or other atypical parkinsonism not otherwise specified; Hoehn and Yahr stage 4 or higher (either directly noted in chart review at the most recent visit, or indicated during the screening phone call by the individual requiring an assistive device to ambulate); severe psychiatric disorders (defined as persistent hallucinations or delusions with lack of insight, or severe anxiety or depression requiring inpatient hospitalization or suicidal ideation in the preceding 30 days); participation in martial arts and/or boxing exercise programs in the 30 days prior to study entry; or inability to commit to attending or traveling to two classes weekly for 10 weeks. Although subjects were encouraged to attend all 20 sessions, we did not impose a minimal attendance exclusion criterion.

Due to the greater availability of instructors and space at the karate studios during the summer, and the availability of medical student study team members (BJS, ML, CMW) to support participants by attending classes, convenience sampling was used to establish the study cohort. The goal was to recruit and enroll all participants in order to conduct the 10 weeks of classes during the summer months, for the reasons above.

Sample characterization, assessments, and outcomes

Prior to the start of classes, participants attended a pre-intervention group visit at Rush University. At this session, participants completed a series of self-report measures for sample characterization, starting with a brief demographic and history questionnaire covering PD diagnosis, prior involvement with martial arts, and expectations related to karate class. To assess baseline fall frequency, the study team defined a fall as any “unintentional going to the ground” before subjects responded to the following question: “How frequently have you fallen in the last twelve months?”, with response choices of “Never”, “Rarely”, “Monthly”, “Weekly”, or “Daily”. To assess baseline activity level, participants completed the International Physical Activity Questionnaire (IPAQ), a validated questionnaire that asks participants to reflect on the past seven days and indicate the number of days, and number of minutes per day, spent in vigorous physical activity, moderate physical activity, walking, and sitting [16]. The IPAQ defines “vigorous” versus “moderate” activity using examples such as bicycle pace and lifting heavy or light loads.

Feasibility assessments included withdrawal rate from the study, adverse effects and fall frequency during the ten-week study period, attendance, achievement of personal goal, likelihood of continuing, and likelihood of recommending karate to a friend with PD. Attendance records were maintained by karate studio staff at each class and sent to the study team. Attendance was calculated as the percentage of classes attended out of a total of twenty possible classes. The study staff contacted any participants missing >2 classes in a row to assess the reason for missed classes, including any adverse effects or intent to withdraw from the study. Study staff attempted at least 3 phone calls to gather data on missingness or withdrawal. Adverse effects were noted either by study or karate studio staff identifying potential adverse effects during class and following up by phone call with the participant, or elicited during phone calls for missingness or withdrawal. At the pre-intervention visit, participants were asked to write down a specific, personal goal that they hoped to achieve through karate. At the post-intervention visit, participants were presented with and asked to indicate whether or not they achieved the personal goal they set at the pre-intervention visit (yes/no). They were asked to provide a binary responses to whether they would recommend the program to someone else with PD, and whether they would continue participation in karate classes, if available.

Secondary outcomes included quality of life and global impression of change. We assessed quality of life using the Parkinson’s Disease Questionnaire-8 (PDQ-8), a brief, validated, PD-specific measure of quality of life. The PDQ-8 yields a summary index score range of 0–100, where higher scores signify worse quality of life [17]. To assess change in overall wellbeing, we administered the Patient Global Impression of Change (PGIC) [18]. The PGIC assesses the individual’s subjective, overall response to an intervention using a seven-point rating scale, ranging from “No change, or worse” (1), “A little better, but no noticeable change” (3), “Moderately better, and a slight but noticeable change” (5), to “A great deal better and a considerable improvement that has made all the difference” (7). Scores of 5–7 are considered clinically meaningful improvements [19,20]. As the PGIC’s single item refers to a change from baseline, it was only administered at the post-intervention visit and six-month follow-up phone call.

Exploratory outcomes included mobility, balance, gait, and flexibility. We measured functional mobility, specifically gait, turning, and rising, with the Timed Up and Go (TUG). The TUG is scored as the number of seconds needed to rise from a chair, walk 10 feet, turn, and return to the seated position. The historical mean TUG score for PD populations is between 10.5–14.8 seconds [21,22]. Balance and gait were assessed with the Tinetti Mobility Test, which has strong psychometric properties and consists of 2 subscales: balance tests (9 items, range: 0–16) and gait tests (7 items, range: 0–12), where higher scores reflect better performance [23]. We also assessed balance and flexibility with the Functional Reach Test, which measures how far forward one can reach when bending from the waist [24]. The TUG, Tinetti Mobility Test, and Functional Reach Test are all recommended tests of these motor domains in PD [25]. Each study team member conducting assessments was trained according to test-specific instructions, with review of simulated participant assessments by the principal investigator.

Additional exploratory outcomes included mood and brief measures of cognition, as these are critical non-motor domains affected by PD which may be amenable to improvement with different exercise modalities. We used the Hospital Anxiety and Depression Scale (HADS), a brief, 14-item, highly validated scale, to measure anxiety and depression (7 items per subscore, each item scored 0–3), where scores >8 on either subscore indicate probable symptoms. To assess cognition, we administered the Digit Span Test (DST), testing the number of digits a participant can recall correctly after verbal presentation [26]. Both forward and backward digit span were measured, with raw scores scaled based on age. DST is a sensitive measure of working memory, a domain frequently affected in PD [27]. The Symbol Digit Modalities Test (SDMT) assesses visual scanning, divided attention, tracking, and motor speed, all of which can be affected in PD [28]. For the SDMT, the participant is presented with a reference key and has 90 seconds to pair specific numbers with given geometric figures, with either written or oral responses.

Following the ten weeks of karate classes, participants attended a post-intervention study visit at Rush University, where the following measures were repeated: Fall frequency, IPAQ, PDQ-8, TUG, Functional Reach Test, and Tinetti Mobility Test. Each objective mobility, balance, and gait assessment was administered by the same study team member, respectively, at the pre- and post-intervention visits to avoid inter-rater variability. The study team members (JF, BS, CN, ML, CW) administering the post-intervention assessments were blinded to pre-intervention assessment results and attendance records.

Six months post-intervention, a study team member telephoned all completers to determine whether they were still practicing karate, either in formal classes, on their own, or both. We assessed reasons for stopping, if applicable, and repeated the PGIC. A different team member than had interfaced with the participants at the pre- and post-intervention visits was selected for the follow-up calls to limit social desirability bias.

Intervention

The intervention consisted of ten weeks of twice-weekly, non-contact, PD-specific karate classes in the Shotokan style, for a total of twenty classes, each one hour long. Participants attended group classes at one of two community karate studios in the Chicago suburbs based on location and schedule desirability. Each one-hour class followed a pattern of warm-ups, stationary basics, moving stance work, striking, conditioning, and as the curriculum progressed, kata (a traditional, choreographed series of movements), and mindfulness and breathing exercises. Details of the class curriculum are in S1 Appendix. Specifically, the warm-ups consisted of light stretching and joint mobility exercises. Conditioning included agility and coordination exercises, often striking a punching bag or practicing kicks with the help of the instructors. Kata is a traditional, choreographed sequence of movements that combines footwork, armwork, and movement across the floor. The kata elements were taught individually until the entire sequence was learned, and then the choreography was practiced at every class. At each class, students practiced core strengthening exercises on the floor and techniques for falling safely. Although breathing and mindfulness were incorporated subtly throughout all elements of class, each session concluded with a cooldown exercise that brought increased focus to these relaxation techniques. All classes were non-contact. Participants would practice strikes in the air, against a punching bag, or towards a pad held by an instructor.

Classes were led by karate instructors with at least one year of prior experience teaching karate to adults professionally, and who had themselves achieved a black belt in karate, demonstrating mastery. All instructors completed an interactive, in-person training with the principal investigator (JF) on common signs, symptoms, and complications of PD (S2 Appendix). Instructors were encouraged to ask questions, and all questions were answered prior to the beginning of the intervention. Training support continued as needed throughout the study. Two to three instructors and a study team member were present at each class and assisted participants as needed. Modifications were offered when necessary, e.g., if participants were experiencing imbalance, they were encouraged to broaden their stance and focus on arm movements, allowing them to continue participation in that day’s class. Participants were encouraged to attend all classes, and allowed to make up a missed class at the alternate studio location, however, there were no penalties for missed classes. The intervention is described in full as a TIDieR checklist, and attached as (S3 Appendix).

Statistical analysis

This pilot study was intended to capture feasibility, and due to scheduling logistics of the karate studios, karate instructors, study team members, and participants, the class dates were pre-determined. Sample size was primarily defined by the number of individuals able to be recruited ahead of, and available to participate in, the 10-week summer session. Our power analysis was based on the TUG, assuming a mean score of 14.8 seconds and standard deviation of 5.8 seconds for historical PD populations, minimal clinically important difference of 5 seconds [21,22], correlation between baseline and follow-up assessments of 0.5, alpha of 0.05, and power of 0.9, yielding a sample size of 15.

Study data was collected and managed using REDCap, a secure, web-based, electronic data capture tool hosted at Rush [29]. Data was exported and analyzed in Stata 15 [30]. Anonymized data on pre- and post-intervention clinical characteristics, questionnaires, and assessments will be shared by request with any qualified investigator. We report the basic descriptive statistics of the cohort, and for feasibility, we report on dropout rate, adverse events, falls, and attendance, defined as the percentage of the 20 classes attended. We summarized categorical variables by frequencies and percentages, assessed continuous variables for normality, and summarized by mean and standard deviation, or median and interquartile range, as appropriate. We analyzed whether participants’ individual self-determined goals for the class were achieved, whether they had an interest in continuing karate classes, and whether they would recommend classes to others, with binary response options for each, respectively. We report on pre-intervention assessments, including IPAQ, TUG, Functional Reach Test, Tinetti Mobility Test, PDQ-8, HADS, DST, and SDMT.

We then analyzed the within-participant change in IPAQ, TUG, Functional Reach Test, Tinetti Mobility Test, and PDQ-8, HADS, DST, and SDMT, respectively, from pre- to post-intervention, using paired t-tests or nonparametric Wilcoxon signed rank tests, as appropriate. We calculated effect size (Cohen’s d) where appropriate, with <0.2 indicating trivial, 0.2–0.5 small, >0.5–0.8 medium, and >0.8 large effect sizes, respectively. As all participants chose only two of the five response options to the fall frequency question, (“Never,” “Rarely,” “Monthly,” “Weekly,” or “Daily), this was rendered a binary variable. We analyzed pre- and post-intervention fall frequency using McNemar’s test. Given the ordinal nature of the PGIC, we report on the percentage of completers endorsing each of the 7 response options at the post-intervention visit and six-month follow-up call, and compared the proportion indicating a clinically meaningful change (score of 5–7) at each time point using McNemar’s test to evaluate for a change in the proportion over time.

Baseline characteristics of participants

We assessed 146 individuals for eligibility (Fig 1). Among 127 people excluded, the main reasons were: Hoehn and Yahr Stage >3 at the most recent outpatient visit or by self-report (n = 45), scheduling conflicts with karate classes (n = 16), atypical parkinsonism/non-PD diagnosis (n = 14), no response to screening phone calls (n = 12), or lack of interest (n = 11). Nineteen participants enrolled in the study and completed the pre-intervention visit, 15 from Rush University’s Parkinson’s Disease and Movement Disorders Program, and 4 recruited from outside neurologists. Table 1 demonstrates the baseline characteristics of all participants (n = 19) and all participants completing the study (n = 15). Based on the IPAQ, participants reported a mean of 5.37 days in which they walked at least 10 minutes at a time, with a median of 30 minutes spent walking on each such day (interquartile range 30–90 minutes), and a mean of 287 minutes per day spent sitting (SD 203) in the preceding week. Participants reported engaging in vigorous physical activity, such as aerobics, fast cycling, or heavy lifting, on 2.17 days per week (SD 2.18) and for a mean of 51.25 minutes (SD 31.59) on each such day. Participants also reported engaging in moderate physical activity, such as carrying light loads or cycling at a regular pace, on 2.29 days per week (SD 2.43) and for a mean of 88.89 minutes (SD 81.42) on such days.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Study flow diagram.

Study flow diagram demonstrating potential participants screened for eligibility, participants enrolled and completing the study, and reasons for ineligibility or exclusion.

https://doi.org/10.1371/journal.pone.0237777.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Baseline characteristics of KICK OUT PD cohort.

https://doi.org/10.1371/journal.pone.0237777.t001

Feasibility outcomes

Fifteen participants completed the post-intervention assessments. Among those lost to follow-up, two were due to non-intervention-related illness or injury (appendicitis and pre-existing sciatica, respectively), one due to unexpected scheduling conflicts, and one withdrew prior to starting classes without providing a reason, yielding an overall 21.1% attrition rate. Neither the participant with appendicitis nor the participant withdrawing due to pre-existing sciatica felt that these were adverse effects due to karate, nor did the study team find them to be related. There was no change in fall frequency during karate classes, with the percentage of participants reporting rare falls prior to the study(vs. no falls) decreasing from 53.3% pre- vs. 20% during the study (McNemar’s test statistic = 5, exact p = 0.06). As shown in Table 1, there were no significant differences in sample characteristics between those who withdrew compared with those who completed the study (p = 0.86–1.00). Mean attendance at twice-weekly classes was 86.7% (SD 9.9%).

After 10 weeks, 100% of the 15 remaining participants intended to continue karate, 100% would recommend the classes to a friend with PD, and 100% felt they had achieved their pre-intervention personal goal, respectively. These personal goals revolved around themes including improvement of balance, mobility, and overall improved management of PD symptoms; benefiting from the social aspect of committing to exercise at a time and place with other people; improved mindfulness, and more general goals of having fun, getting exercise, and working towards a better future for themselves. Detailed qualitative analyses of pre- and post-intervention expectations, achievements, facilitators, and barriers are forthcoming.

Change in pre- and post-intervention outcomes

We found a statistically and clinically significant improvement in quality of life of 5.9 points on the PDQ-8 (25.3 vs. 19.3, mean difference -5.9, 95% CI -9.9 to -2, p = 0.01, effect size 0.8, Table 2). There were no changes in physical activity levels compared to baseline, as measured by the IPAQ and shown with additional exploratory outcomes in Table 2. Mean TUG decreased from 9.6 to 9.0 seconds (mean difference -0.6, 95% CI -1.3 to 0.2, p = 0.12, effect size 0.4), however this population was notably faster at baseline compared to historical mild- to moderate-stage PD validation cohorts. We found no change in Functional Reach Test, and a small change in Tinetti Mobility Test.Additional exploratory outcomes are shown in Table 2: Mean TUG improved from 9.6 to 9.0 seconds (mean difference -0.59, 95% CI -1.34 to 0.17, p = 0.12, effect size 0.43), however this population was notably faster at baseline compared to historical mild- to moderate-stage PD validation cohorts. We found no change in Functional Reach Test (10.8 vs. 12.5 inches, mean difference 1.73, 95% CI -2.5 to 5.96, p = 0.39, effect size 0.23), and a statistically significant change in Tinetti Mobility Test (27.07 vs. 27.93, mean difference 0.87, 95% CI 0.24 to 1.49, p = 0.01, effect size 0.77).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Pre- and post-karate secondary outcomes among participants.

https://doi.org/10.1371/journal.pone.0237777.t002

There were no significant changes in the exploratory outcomes of mood or cognition. The mean HADS depression subscore was 4.47 (SD 4.32) pre-intervention, and 4.8 (4.13) post-intervention (p = 0.53). The mean HADS anxiety subscore was 5.53 (3.81) pre-intervention and 5.93 (4.67) post-intervention (p = 0.35). Neither the scaled digit span forward score (12.33, SD 3.37 pre-intervention vs. 11.93, SD 3.65 post-intervention, p = 0.41) nor the scaled backward score (10.67, SD 2.97 pre-intervention vs. 10.47, SD 2.53 post-intervention, p = 0.81) demonstrated change. Similarly, there was no change in the SDMT scaled score (pre-intervention: mean 96.17, SD 9.73; post-intervention 95.67, SD 10.87; p = 0.78).

Global impression of change and sustainability

Fig 2 shows the percentage of participants endorsing each response option on the Patient Global Impression of Change, with outlined segments indicating clinically meaningful change. Overall, 86.7% of completers endorsed clinically meaningful change immediately post-intervention, and 53.3% did so six months later (McNemar’s test p = 0.03). After six months, 8/15 individuals (53.3%) continued practicing karate, with seven still attending classes (mean 1.58 days per week, SD 0.79). Five individuals continued to practice karate on their own for a mean of 3.6 days per week (SD 2.07), 22 minutes per session (SD 8.37). Surprisingly, five participants (33%) competed in a regional tournament during follow-up and all advanced to the next level of training (i.e., from white to yellow belt). Among those who did not continue karate, three cited a lack of time, two cited inconvenient locations, and one each reported classes to be too costly or of insufficient benefit, respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Global impression of change.

Participants’ global impression of change immediately post-participation and six months post-participation.

https://doi.org/10.1371/journal.pone.0237777.g002