Secondary Logo

Journal Logo

Original Research Articles CME Article . 2018 Series . Number 3

Trunk Exercises Improve Gait Symmetry in Parkinson Disease

A Blind Phase II Randomized Controlled Trial

Hubble, Ryan P. PhD; Naughton, Geraldine PhD; Silburn, Peter A. PhD, FRACP; Cole, Michael H. PhD

Author Information
American Journal of Physical Medicine & Rehabilitation: March 2018 - Volume 97 - Issue 3 - p 151-159
doi: 10.1097/PHM.0000000000000858


Postural instability is one of the most disabling symptoms of Parkinson disease (PD) and contributes to falls in this population. Unfortunately, these symptoms are poorly managed by current pharmacological and surgical interventions,1 which emphasizes the need for more effective alternative therapies for improving the overall management of these patients. Exercise-based interventions are known to improve many motor symptoms and clinical measures of balance, mobility, and falls risk in people with PD.2,3 However, despite the proven efficacy of exercise-based interventions for improving the clinical symptoms of PD, there is currently conflicting evidence regarding whether these methods are capable of reducing the rate and number of falls in PD populations. For example, some research has reported a significant reduction in the number and rate of falls experienced by people with PD after an exercise-based intervention,4 whereas others have reported no significant change in these outcomes.3 Despite these inconsistencies, a recent systematic review and meta-analysis of eight randomized controlled trials has added important weight to this debate; indicating that although exercise does not reduce the number of fallers in the PD population, it is effective at reducing the rate of falls experienced by each individual.5

During locomotion, the maintenance of equilibrium relies upon one's ability to produce smooth and rhythmic movements of the head and trunk, which collectively comprise almost 60% of the body's mass.6 Although clinical assessments of axial motor symptoms have been shown to provide some insight into the effects of the disease on one's ability to safely ambulate, most tests that assess postural stability, trunk rigidity, and gait disability rely on Likert scales, which seems to make them less sensitive to subtle changes in function.7 To improve the assessment of postural stability during dynamic activities, researchers have started using lightweight wearable sensors to examine gait stability8–10 and muscle activation11 in people with PD. Specifically, scientists have used such devices to measure the medial-lateral (ML, side-to-side), anterior-posterior (AP, front-to-back), and vertical (VT, up-and-down) movement patterns of the head, trunk, and pelvis during walking to assess disease-related changes in gait stability. Of the measures reported, the harmonic ratio (HR) is one of the more common8–10 and, in the context of walking, provides a measure of the step-to-step symmetry of segmental accelerations.12 Higher HRs describe improved step-to-step symmetry and are typically considered to represent a more stable gait pattern. The HR has previously been used to identify differences in movement symmetry between PD patients and controls,9,10 PD freezers and nonfreezers,13 and PD patients who have a history of falling and those who have not previously fallen.8,9 Interestingly, recent research has shown that these deficits in segmental control are accompanied by specific alterations in the activation patterns of the superficial trunk muscles.11 Specifically, people with PD who prospectively reported falling had greater peak and baseline levels of erector spinae activity during walking compared with age-matched controls.11 Interestingly, these differences in baseline activity were shown to be significant predictors of the ML pelvis, trunk, and head displacement11 that has been linked with future falls in previous research.14,15 The authors argued that the increased baseline activation of the erector spinae may have been indicative of an underlying dysfunction of the deeper and more fatigue-resistant muscles involved in postural control (i.e., multifidus, transverse abdominus).

Given the altered trunk muscle activations exhibited by PD fallers have been linked to larger and less symmetrical head11,14,15 and trunk11 movements, it is possible that exercises that target trunk mobility and endurance may assist with improving the step-to-step symmetry of head and trunk movements in this population. Therefore, it was the purpose of this phase II randomized controlled trial to determine whether a 12-wk intervention incorporating both exercise and falls prevention education was more effective than falls prevention education alone at improving gait symmetry in people with PD. It was hypothesized that the exercise group would have improved step-to-step symmetry after the intervention, whereas participants in the education group would exhibit no improvements in gait symmetry after the 12-wk intervention period.



This phase II randomized controlled trial was developed in accordance with the Consolidated Standards of Reporting Trials guidelines (see Checklist, Supplemental Digital Content, Individuals from a metropolitan neurology clinic diagnosed with idiopathic PD, based on the UK Brain Bank Criteria,17 were sent a letter outlining the details of the study inviting them to volunteer. Prospective participants were initially screened over the telephone and were excluded if they had the following: (a) an inability to ambulate independently; (b) uncontrolled hypertension; (c) a prescription for psychotropic medications; (d) significant limitations due to osteoporosis; (e) orthopedic surgery within the previous year; (f) serious neck, shoulder, or back injuries (including spinal fusions); (g) received deep brain stimulation surgery for symptom management; (h) a neurological condition other than PD; or (i) no history of falls or near misses within the past year. For the purposes of this study, a fall was defined as a coming to the ground or lower level not as the result of a major intrinsic event or overwhelming hazard.3 Similarly, near misses were defined as events during which an individual felt that they were going to fall but did not.3 The study's protocol18 was registered with the Australian and New Zealand Clinical Trials Registry (ACTRN12613001175763) and approved by the University's Human Research Ethics Committee (2013 223Q). The recruitment and assessment of all participants were completed between February 2014 and December 2015, and all volunteers provided written informed consent in accordance with the Declaration of Helsinki.

On the basis of an a priori sample size calculation using ML trunk HRs recorded for people with PD during walking,10 it was determined that a minimum of 11 participants was required per group to confidently report any significant changes in the step-to-step symmetry of trunk motion (diff = 0.05, SD = 0.04, Cohen d = 1.25, Power = 80%, P = 0.05).

Clinical Measures

Before randomization, participants completed a battery of baseline assessments including the following: (1) Addenbrooke Cognitive Examination; (2) Bailey-Lovie high-contrast visual acuity test; (3) Timed Up and Go test; (4) Activities-specific Balance Confidence scale; (5) 39-item Parkinson's Disease Questionnaire (PDQ-39); (6) Part III of the Unified Parkinson Disease Rating Scale (UPDRS III); (7) Hoehn & Yahr stage score; and (8) Schwab & England Activities of Daily Living Scale. Furthermore, participants completed a previously developed questionnaire to collect details about their medical history (e.g., date of diagnosis) and prescription medication use. Using the information provided, it was possible to calculate each participant's levodopa equivalent daily dose using previously described methods.19 All baseline measures were taken 1 to 2 hrs after the patient's scheduled dose of anti-parkinsonian medication to ensure that results were representative of similar tasks performed in the real world. Participants with significant visual (Bailey-Lovie high-contrast visual acuity >0.30 logMAR) and/or cognitive (Addenbrooke Cognitive Examination score <82) impairment were excluded before baseline testing.

Gait Analysis

After the clinical assessments, participants completed four walking trials separated by a rest break of at least 30 secs along a 10-meter walkway at a self-selected pace. While performing this task, head and trunk accelerations were measured at 1500 Hz using two microelectromechanical system three-dimensional accelerometers (Noraxon Inc, Scottsdale, AZ), which were statically calibrated using previously described methods.20 To facilitate the assessment of head and trunk accelerations during walking, the accelerometers were firmly attached over the (1) occipital protuberance of the skull via a sport headband and (2) the spinous process of the 10th thoracic vertebra (T10) using double-sided tape. Walking speed was measured using a pair of Speedlight timing gates positioned 6 meters apart (SWIFT Performance Equipment, Alstonville, Australia).

In addition to the acceleration patterns of the head and trunk, bilateral activation of the thoracic and lumbar erector spinae was measured using surface electromyography. The skin overlying the muscles of interest was prepared with an abrasive gel (NuPrep; Weaver & Company, Aurora, CO) and cleaned thoroughly with isopropyl alcohol to improve myoelectric signal quality. Where necessary, excessive body hair was removed with a razor before skin abrasion to improve signal quality and to enhance electrode adhesion. After skin preparation, four pairs of silver/silver chloride pregelled surface electrodes (AMBU Blue Sensor, Ballerup, DK; 34-mm diameter, 10-mm2 sensing area) were placed with a center-to-center interelectrode distance of 34 mm over the thoracic (5 cm lateral to the T10 spinous process) and lumbar (2 cm lateral to the L3 spinous process) erector spinae.11 Raw electromyogram (EMG) data were collected at a rate of 1500 Hz using wireless transmitters containing integrated preamplifiers (gain: 500, common mode rejection ratio: >100 dB, input impedance: >100 MΩ). To facilitate synchronization of head and trunk accelerations with trunk muscle activations, both data sets were wirelessly telemetered to a Telemyo DTS belt receiver and to a laptop running the MyoResearch XP software (Noraxon Inc).

To allow for intergroup and interday comparisons, trunk muscle activity was expressed as a percentage of the peak activation recorded for each muscle during three maximum voluntary isometric contractions.11 To perform the maximum voluntary isometric contraction tasks, participants were required to lie prone/prostrate on a padded table with their hips flexed and their feet on the floor with a Velcro strap placed over the lower torso to secure them to the table for safety. During each maximal effort, the patients simultaneously extended both hips to raise their legs to a horizontal position (i.e., 180 degrees) at which point their movements were actively resisted by the researcher. This method was chosen in preference to the traditional Biering-Sørensen test to limit the potential difficulties that older participants may have with this more complex movement pattern.21 Participants were verbally encouraged by the researchers, and the maximum value recorded for each muscle during the three trials was used for normalization of electromyography data collected for that muscle during the walking trials.

Randomization and Blinding

After baseline assessment, participants were assigned by a member of the research team (RPH) to one of two 12-wk intervention groups using a random allocation sequence (block size = 2; 1:1 ratio) that was generated by a team member who was not involved in participant allocation or assessment (GAN). To minimize the risk of bias and to eliminate the potential for interrater reliability issues, all clinical assessments were conducted by an experienced movement disorders scientist who was blinded to participant group assignment (MHC). A flow diagram of participant recruitment and group assignment is represented in Figure 1.

Summary of the number of participants contacted, screened, assessed, and randomized to each group for the purposes of this study.


Participants were randomly assigned to receive either 12 wks of falls prevention education or 12 wks of exercise and falls prevention education and were required to commence their assigned intervention within a week of completing the baseline assessments. The use of an exercise-based intervention is supported by systematic evidence, which demonstrates that exercise is one of the best methods for reducing falls risk in older adults.22 In contrast, previous research has reported little to no evidence regarding the efficacy of falls prevention education strategies with respect to their capacity to reduce falls risk in ageing populations.22 As such, those assigned to receive exercise and falls prevention education comprised the treatment group, whereas those receiving the falls prevention education represented the placebo group. Because both groups received the same falls prevention education, it was possible to discriminate the changes resulting from the exercise-based intervention from those related to the education program.

During the 12-wk intervention, participants in the education group received a weekly multidisciplinary educational brochure that explained how factors, such as exercise, nutrition, and/or sleep quality, may influence their risk of falling. The education brochures provided a combination of written and illustrative materials and were developed using information freely available from community-based support groups. Participants assigned to the exercise group received the same weekly education brochures but also completed a 12-wk exercise program aimed at improving trunk mobility and endurance. This program involved one supervised 90-min session each week with a trained exercise scientist in groups of up to three participants. This exercise program was designed to conform with current recommendations for exercise-based interventions that target stability23,24 and was informed by programs previously described for older adults25 and people with PD.26 In short, the exercise-based intervention comprised the following three parts: (1) a warmup focusing on trunk mobility exercises to improve range of motion; (2) an exercise routine focusing on the endurance and stability of the trunk muscles (multifidus, erector spinae, obliques, transverse abdominus, rectus abdominus); and (3) a cooldown involving stretching and walking in a real-world environment. An in-depth description of the specific endurance and mobility tasks involved in the exercise-based intervention has been previously published elsewhere.18 Participants were reassessed 12 and 24 wks after the baseline assessment, 1 to 2 hrs after their scheduled dose of anti-parkinsonian medication, to ensure a fair comparison with baseline assessments. Where possible, the 12-wk follow-up assessment was scheduled to occur within 1 wk of the participants completing their allotted intervention.

Data Analyses

Primary Outcome (Gait Step-to-Step Symmetry)

Raw accelerations were transformed to a horizontal-VT orthogonal coordinate system to remove the effect of gravity from the AP and ML axes of the sensors.20 After transformation, accelerations were low-pass filtered using a bidirectional fourth-order Butterworth filter, with a cutoff frequency of 30 Hz.8 The time series of the filtered AP, ML, and VT head and trunk accelerations were then divided into individual gait cycles by identifying the peaks in VT trunk accelerations, which coincide with heel contact.8–10 The AP, ML, and VT HRs were then calculated for six successive gait cycles within each walking trial, with the average AP, ML, and VT HRs of these gait cycles used for further analysis.

Secondary Outcome (Muscle Function)

To evaluate trunk muscle function, the raw EMGs were initially processed with an adaptive filter to attenuate any influence of the electrocardiogram on the trunk EMGs.11 Data were then full-wave rectified and low-pass filtered using a fourth-order Butterworth filter with a 20-Hz cutoff frequency. The root mean square method was then used to process the rectified and filtered EMGs for consecutive 50-millisec windows (i.e., 75 samples) with a 74-sample overlap. All processed EMGs were then normalized by expressing them as a percentage of the peak activation recorded during the maximum voluntary isometric contraction trials.

The secondary outcomes were calculated for three successive gait cycles for each leg (i.e., 6 gait cycles total), which were taken from the middle of the trial to minimize the influence of acceleration and deceleration at the beginning and end of the trial. For the assessment of muscle function, the three gait cycles completed for each leg produced eight peaks of muscle activity (i.e., 4 left and 4 right footfalls yield 3 left and 3 right gait cycles; 1 peak per footfall). The eight peaks derived from the left- and right-side thoracic and lumbar erector spinae muscles (i.e., 16 peaks per vertebral level) were subsequently averaged, yielding a single peak value for each vertebral level during each walking trial. Similarly, to evaluate the extent to which the superficial trunk muscles “switched off” between strides, the minimum EMG amplitude between successive heel contacts (i.e., within the 7 troughs between the 8 activation peaks) was determined.11 Similar to the methods used for the peak activation data, the 7 troughs for the left- and right-side thoracic and lumbar erector spinae (i.e., 14 troughs per vertebral level) were averaged to represent the minimum activation of these muscle groups during each walking trial. All EMG processing was performed using the MyoResearch XP software (MR 3.6.20), whereas custom programs developed in Matlab R2015b (The Mathworks, Natick, MA) were used to identify peaks and troughs in the processed EMGs and to transform and process the raw accelerations.

Statistical Analysis

The one-way analysis of variance procedure was used to compare the groups at baseline for differences in continuous demographic variables (e.g., age), whereas the χ2 test was used to compare groups for categorical outcomes (e.g., sex). Where the assumptions of normality (Shapiro-Wilks test) or homogeneity of variance (Levene test) were violated, the nonparametric Kruskal-Wallis test was used to compare the continuous measures.

To determine the efficacy of the two 12-wk interventions, the analyses were based on a modified intention to treat approach.27 In accordance with this approach, two participants who withdrew from the study after randomization were excluded from the analyses, because they did not receive treatment and, hence, were unable to contribute meaningful data regarding the efficacy of the interventions. Furthermore, when participants who completed the 12-wk intervention were unable to return for follow-up testing, their data were imputed using the last observation carried forward method.28 Linear mixed model analyses were conducted to determine whether the trunk-specific exercise program was more effective than the education program at improving the step-to-step symmetry of head and trunk movements and muscle activation. These models included two repeated factors (day; 3 levels; trial; 4 levels), one fixed factor (group; 2 levels) and two covariates (levodopa equivalent daily dose and walking speed). Walking speed and levodopa equivalent daily dose were both included as covariates, because they were significantly correlated with the primary and secondary outcomes. Furthermore, it is well known that walking speed influences accelerations8,9 and trunk muscle activations11 and that levodopa significantly improves some motor symptoms in PD.1,29 When a significant group-day interaction was identified, the Tukey least significant difference post hoc procedure was used to identify where the differences lay.

To provide insight into the clinical meaningfulness of any changes in step-to-step symmetry, muscle activations, and/or the clinical rating scales, the minimal detectable change for each measure was also derived. The minimal detectable change score represents the minimum change in a particular outcome measure that would be considered to result in a meaningful change in patient function and, hence, provides useful information regarding the clinical importance of the reported findings. All statistical analyses were completed with Statistical Product and Service Solutions (SPSS v21.0) and the level of significance was set at P < 0.05.


Study Population Retention and Compliance

Of the 24 participants assessed at baseline, 22 completed the 12-wk intervention and two withdrew citing changes in circumstances that made them unable to commit to the project. Comparison of the remaining 22 patients at baseline indicated that the exercise and education groups did not differ with respect to measures of cognition, vision, neurological function, or mobility. However, individuals in the exercise group had a greater body mass index at baseline than the education group (Table 1).

Baseline demographics and disease-specific scores for the participants randomized to the education and exercise interventions

Although all 22 participants were reassessed at 12 wks (mean 12-wk follow-up time: exercise = 94.6 [2.0] days, education = 92.1 [3.0] days; P = 0.49), four participants (2 exercise; 2 education) did not complete the 24-wk follow-up (mean 24-wk follow-up time: exercise = 188.6 [7.0] days, education = 186.4 [7.4] days; P = 0.84). Of these patients, two reported having recently undergone deep brain stimulation surgery, one was unavailable to complete the follow-up and one was not contactable via telephone or e-mail. As such, the 24-wk data for these four participants were imputed from the 12-wk assessment using the last observation carried forward method. Average participant compliance for the exercise sessions was 90%, with individual compliance ranging from 8 (67%) to 12 (100%) of the 12 supervised sessions. Participants reported no discomfort or adverse effects associated with either intervention.

Primary Outcome (Step-to-Step Symmetry)

Statistical comparison of the two groups indicated that they did not differ with respect to the step-to-step symmetry of head and trunk movements at baseline. The linear mixed model analyses indicated no significant main effects for group, but significant group-day interactions were reported for AP (P = 0.038) and VT (P = 0.004) head movements and AP (P < 0.001), ML (P = 0.003), and VT (P = 0.024) trunk movements. Pairwise comparisons revealed that the exercise group demonstrated improved step-to-step symmetry for VT head movements (P = 0.009) and AP trunk movements (P < 0.001) at 12 wks compared with baseline. Furthermore, at 24 wks, improvements in the step-to-step symmetry of AP head movements (P = 0.040) and AP trunk movements (P < 0.001) were evident compared with baseline values, whereas AP head movements were also better at this time point, relative to the 12-wk assessment (P = 0.011).

In contrast, the post hoc analyses indicated that the education group exhibited poorer step-to-step symmetry, with respect to AP (P = 0.005) and VT (P = 0.035) head movements and AP (P = 0.049), ML (P < 0.001), and VT (P < 0.001) trunk movements at 12 wks relative to the baseline measures. In addition, step-to-step symmetry of VT trunk movements (P = 0.024) was reduced at 24 wks compared with baseline, whereas the step-to-step symmetry of ML trunk movements was improved at 24 wks compared with the 12-wk assessment (P = 0.010).

When assessing the clinical importance of the significant improvements/declines in the step-to-step symmetry of head movements reported for the groups (via the minimal detectable change score), it was determined that the recorded changes were not substantial enough to be considered clinically meaningful. Despite this, the improved step-to-step symmetry of AP trunk movements recorded for the exercise group at 12 and 24 wks was not only statistically significant but sufficiently large to be considered clinically important. Similarly, the reduced step-to-step symmetry of ML and VT trunk movements in the education group at 12 wks was substantial enough to be considered a clinically important change (Table 2).

Estimated marginal means (and standard errors) from the linear mixed model analyses performed for the HRs and muscle activation patterns of the education and exercise groups at baseline, immediately after intervention (12 wks), and after the 12-wk retention period (24 wks)

Secondary Outcomes (Trunk Muscle Activation and Clinical Outcomes)

At baseline, participants assigned to the education group exhibited significantly greater peak activations for the lumbar erector spinae but otherwise did not differ from those in the exercise group with respect to the other secondary outcomes. Analysis of the trunk muscle activations of the two groups identified significant group-day interactions for peak (thoracic: P < 0.001; lumbar: P = 0.032) and trough (thoracic: P < 0.001; lumbar: P = 0.010) erector spinae activity. Post hoc analyses indicated that peak activation of the thoracic erector spinae was increased at 12 wks relative to the baseline assessment for the exercise group (P = 0.026). In contrast, thoracic erector spinae activity within the troughs was significantly reduced at baseline (P = 0.039) and 12 wks (P = 0.049), relative to the 24-wk assessment, whereas the lumbar erector spinae exhibited less activation in the troughs at 12 wks, relative to baseline (P = 0.011). Nevertheless, despite the statistical significance of these outcomes, the minimal detectable change values indicated that the recorded changes in trunk muscle activation for the exercise group were insufficient to be deemed clinically meaningful.

For the education group, the pairwise comparisons revealed that peak activation of the thoracic and lumbar erector spinae was significantly reduced at 12 wks (thoracic: P < 0.001; lumbar: P = 0.008) and 24 wks (thoracic: P < 0.001; lumbar: P < 0.001) compared with baseline. Furthermore, peak activation of the lumbar erector spinae was significantly reduced at 24 wks relative to the 12-wk assessment for those patients in the education group (P = 0.010). With respect to the minimum activation levels of the thoracic erector spinae, participants in the education group demonstrated significantly reduced activation during the 12- (P < 0.001) and 24-wk (P < 0.001) assessments compared with baseline. Similarly, lumbar erector spinae activation within the troughs was significantly reduced at the 24-wk time point compared with the baseline (P < 0.001) and 12-wk (P < 0.001) assessments for these patients. Despite of the large number of changes recorded in erector spinae activation for the education group, only the reduction observed in peak lumbar erector spinae activation between the baseline and 24-wk assessments and the changes in minimum levels of activation were sufficiently large to be considered clinically important (Table 2).

Unlike the analyses conducted for step-to-step symmetry and trunk muscle activation, statistical analysis of the clinical scores identified no significant group-day interactions (Table 3), suggesting that neither intervention led to a measurable change in the clinical outcomes.

Estimated marginal means (and standard errors) from the linear mixed model analyses performed for the clinical assessments of mobility, balance confidence, symptom severity, and quality of life for the education and exercise groups at baseline, immediately after intervention (12 wks) and after the 12-wk retention period (24 wks)


This phase II randomized controlled trial represents the first study to examine the efficacy of a 12-wk trunk-specific exercise program for improving step-to-step symmetry in people with PD. The results support the hypothesis that trunk-specific exercises may improve (or at the very least, maintain) step-to-step symmetry of trunk movements and trunk muscle function in this population. These outcomes are commensurate with previous research, which demonstrated improvements in step-to-step symmetry of VT trunk movements in people with mild cognitive impairment after a 6-mo multicomponent exercise program.30 In addition, these results extend existing knowledge by suggesting that measures of step-to-step symmetry, such as the HR, may be suitable for assessing subtle changes in postural control during dynamic tasks, such as walking. The lack of significant changes in step-to-step symmetry between the 12- and 24-wk assessments for the exercise group also suggests that the benefits of the weekly exercise program may be retained for up to 12 wks after the cessation of the training regime. In contrast, the findings presented for the education group suggest that without specifically focusing on maintaining mobility and core strength, the step-to-step symmetry of trunk movements may decline in as little as 12 wks. These findings should be considered with some caution, however, because the poorer step-to-step symmetries reported for the education group at 12 wks had generally returned to near baseline values by the 24-wk assessment. Collectively, these findings tend to suggest that the exercise-based intervention was more effective than the education-based intervention at improving gait symmetry in people with PD. However, it should be reiterated that the mode of delivery used for the education intervention required participants to be proactive and to seek guidance and/or additional information, if needed. As such, it is possible that the efficacy of such an approach could be improved if it were delivered in a more structured and closely monitored fashion.

The reported changes in step-to-step symmetry were complemented by changes in trunk muscle function. Specifically, the education group experienced significant and clinically meaningful declines in the step-to-step symmetry of trunk movements at 12 wks that were matched with declines in peak and minimum levels of erector spinae activity at 12 and/or 24 wks. However, it should be noted that the education group exhibited significantly greater peak activation of the lumbar erector spinae at baseline compared with the exercise group. Furthermore, despite the peak activation of this muscle group reducing for the 24-wk period of the study, its level of activation at the 24-wk time point was not dissimilar to that reported for the exercise group. Nevertheless, considering the erector spinae are bilaterally activated around initial heel contact to control the forward flexion moment experienced by the trunk,31 it may be argued that the reduced peak thoracic erector spinae activation may have contributed to the poorer trunk control exhibited by the education group at 12 wks. Support for this notion was provided by the results presented for the exercise group that exhibited significant (although not clinically meaningful) improvements in trunk muscle function after the intervention. Considering these results in conjunction with the improved AP symmetry of head and trunk movements during the 12- and 24-wk tests, it could be argued that the improved trunk muscle function exhibited by the exercise group served to resist the large anteriorly directed torque imposed upon the body at heel contact.31 Collectively, these results seem to provide evidence to suggest that a targeted exercise program may assist with maintaining trunk muscle function in people with PD, which has important implications for clinical practice.

As with any study, there are a number of potential limitations that should be considered when interpreting these outcomes. First, a slow rate of recruitment resulted in a relative small sample size (from a statistical perspective). Although the comparisons reported for the primary outcome measure were supported by an a priori power calculation, the generalizability of these findings to a larger cohort is unknown. Second, an ancillary aim of this project was to evaluate whether three weekly exercise sessions offered greater improvements in gait symmetry and trunk muscle function than one weekly exercise session.18 However, difficulties with participant recruitment and retention made it necessary to discard this secondary aim and focus on the primary aim. Given the encouraging outcomes of this study, future research might seek to establish whether increasing the frequency of this exercise program offers greater improvements in step-to-step symmetry and/or has the potential to reduce the rate of falls in people with PD. It should also be noted that the design of this study meant that participants in the exercise group received a higher treatment dose than the education group for the 12-wk intervention period (i.e., 12 × 90-min exercise sessions = 1080 additional mins). As such, although one could argue that the improvements in AP head and trunk step-to-step symmetry resulted from the targeted trunk muscle exercises, it is possible that these improvements were attributable to the greater level of activity of the exercise group for this period. Finally, although every effort was made to ensure that patients were assessed at a similar time of day for each testing session, logistical constraints meant that some participants had be tested at a different time of the day for one or more of the follow-up sessions. Although this may have influenced the reported outcomes, care was taken to ensure that participants were tested 1 to 2 hrs after a scheduled dose of anti-parkinsonian medication to minimize the influence of any motor fluctuations that patients may experience throughout the medication cycle.

In conclusion, the results presented suggest that by performing as little as one focused exercise session per week, it may be possible to improve or at the very least maintain step-to-step symmetry in people with PD. Such improvements in function are likely to have significant implications for an individual's self-confidence and independence, which ultimately should contribute to an improved quality of life. Given these findings, exercises that target trunk muscle function should be considered when developing an exercise program that seeks to improve balance and reduce falls risk in people with PD.


1. Bloem BR, Hausdorff JM, Visser JE, et al.: Falls and freezing of gait in Parkinson's disease: a review of two interconnected, episodic phenomena. Mov Disord 2004;19:871–84
2. Allen NE, Canning CG, Sherrington C, et al.: The effects of an exercise program on fall risk factors in people with Parkinson's disease: a randomised controlled trial. Mov Disord 2010;25:1217–25
3. Ashburn A, Fazakarley L, Ballinger C, et al.: A randomised controlled trial of a home based exercise programme to reduce the risk of falling among people with Parkinson’s disease. J Neurol Neurosurg Psychiatry 2007;78:678–84
4. Sparrow D, DeAngelis TR, Hendron K, et al.: Highly challenging balance program reduces fall rate in Parkinson disease. J Neurol Phys Ther 2016;40:24–30
5. Shen X, Wong-Yu IS, Mak MK: Effects of exercise on falls, balance, and gait ability in Parkinson's disease: a meta-analysis. Neurorehabil Neural Repair 2016;30:512–27
6. Winter DA: Biomechanics and Motor Control of Human Movement, 4th ed. New York, John Wiley & Sons Inc., 2009
7. Hubble RP, Naughton GA, Silburn PA, et al.: Assessing stability in mild and moderate Parkinson's disease: can clinical measures provide insight? Gait Posture 2016;49:7–13
8. Cole MH, Sweeney M, Conway ZJ, et al.: Imposed faster and slower walking speeds influence gait stability differently in Parkinson fallers. Arch Phys Med Rehabil 2017;98:639–48
9. Latt MD, Menz HB, Fung VS, et al.: Acceleration patterns of the head and pelvis during gait in older people with Parkinson's disease: a comparison of fallers and nonfallers. J Gerontol A Biol Sci Med Sci 2009;64:700–6
10. Lowry KA, Smiley-Oyen AL, Carrel AJ, et al.: Walking stability using harmonic ratios in Parkinson's disease. Mov Disord 2009;24:261–7
11. Cole MH, Naughton GA, Silburn PA: Neuromuscular impairments are associated with impaired head and trunk stability during gait in Parkinson fallers. Neurorehabil Neural Repair 2017;31:34–47
12. Bellanca JL, Lowry KA, VanSwearingen JM, et al.: Harmonic ratios: a quantification of step to step symmetry. J Biomech 2013;46:828–31
13. Weiss A, Herman T, Giladi N, et al.: New evidence for gait abnormalities among Parkinson's disease patients who suffer from freezing of gait: insights using a body-fixed sensor worn for 3 days. J Neural Transm (Vienna) 2015;122:403–10
14. Cole MH, Silburn PA, Wood JM, et al.: Falls in Parkinson's disease: evidence for altered stepping strategies on compliant surfaces. Parkinsonism Relat Disord 2011;17:610–6
15. Cole MH, Silburn PA, Wood JM, et al.: Falls in Parkinson's disease: kinematic evidence for impaired head and trunk control. Mov Disord 2010;25:2369–78
16. Schulz KF, Altman DG, Moher D: CONSORT Group: CONSORT 2010 Statement: updated guidelines for reporting parallel group randomised trials. J Clin Epidemiol 2010;63:834–40
17. Hughes AJ, Daniel SE, Kilford L, et al.: Accuracy of clinical diagnosis of idiopathic Parkinson's disease: a clinico-pathological study of 100 cases. J Neurol Neurosurg Psychiatry 1992;55:181–4
18. Hubble RP, Naughton GA, Silburn PA, et al.: Trunk muscle exercises as a means of improving postural stability in people with Parkinson’s disease: a protocol for a randomised controlled trial. BMJ Open 2014;4:e006095
19. Deuschl G, Schade-Brittinger C, Krack P, et al.: A randomized trial of deep-brain stimulation for Parkinson's disease. N Engl J Med 2006;355:896–908
20. Moe-Nilssen R: A new method for evaluating motor control in gait under real-life environmental conditions: Part 1 - the instrument. Clin Biomech (Bristol, Avon) 1998;13:320–7
21. Vera-Garcia FJ, Moreside JM, McGill SM: MVC techniques to normalize trunk muscle EMG in healthy women. J Electromyogr Kinesiol 2010;20:10–6
22. Chang JT, Morton SC, Rubenstein LZ, et al.: Interventions for the prevention of falls in older adults: systematic review and meta-analysis of randomised clinical trials. Br Med J 2004;328:680
23. Li F, Harmer P, Fitzgerald K, et al.: Tai chi and postural stability in patients with Parkinson's disease. N Engl J Med 2012;366:511–9
24. Smania N, Corato E, Tinazzi M, et al.: Effect of balance training on postural instability in patients with idiopathic Parkinson's disease. Neurorehabil Neural Repair 2010;24:826–34
25. Hesari AF, Mahdavi S, Abadi MRH, et al.: Comparisons of Berg Balance Scale following core stabilization training in women elderly. Ann Biol Res 2012;3:1499–504
26. Bridgewater KJ, Sharpe MH: Trunk muscle training and early Parkinson's disease. Physiother Theory Pract 1997;13:139–53
27. Gupta SK: Intention-to-treat concept: a review. Perspect Clin Res 2011;2:109–12
28. Streiner D, Geddes J: Intention to treat analysis in clinical trials when there are missing data. Evid Based Ment Health 2001;4:70–1
29. Schaafsma JD, Giladi N, Balash Y, et al.: Gait dynamics in Parkinson's disease: relationship to Parkinsonian features, falls and response to levodopa. J Neurol Sci 2003;212:47–53
30. Doi T, Makizako H, Shimada H, et al.: Effects of multicomponent exercise on spatial-temporal gait parameters among the elderly with amnestic mild cognitive impairment (aMCI): preliminary results from a randomized controlled trial (RCT). Arch Gerontol Geriatr 2013;56:104–8
31. Winter DA, Yack HJ: EMG profiles during normal human walking: stride-to-stride and inter-subject variability. Electroencephalogr Clin Neurophysiol 1987;67:402–11

Exercise; Parkinson Disease; Accident Prevention; Gait

Supplemental Digital Content

Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.