Two 4-Week Balance-Training Programs for Chronic Ankle Instability

Participants

Eighteen participants (2 females, 16 males; age = 18.38 ± 1.81 years; height = 175.26 ± 6.64 cm; mass = 75.79 ± 12.1 kg) started and completed the study over a 6-month period. Fourteen of the participants were in high school, while 4 were in college. All participants were scholastic or recreational athletes. We used a random-number generator by an outside assessor to randomly place participants into the PHSB (n = 9) or the SLB (n = 9) group. A flow diagram based on the CONSORT statement shows the inclusion and exclusion of participants throughout the study (Figure 1). The inclusion criteria were a history of at least 1 ankle sprain, with the initial sprain occurring more than 1 year before the study, and self-reported functional deficits at the time of the study as noted by the Ankle Instability Instrument²⁶ (answer of yes to at least 5 yes-or-no questions) and a feeling of “giving way” (at least 2 episodes in the 6 months before the study).²⁷ Participants were excluded if they had sustained a lower extremity injury, including an ankle sprain, within the 6 weeks before the study or had a history of lower extremity surgery involving the ankle or disorders known to affect balance.²⁷ Participant demographics are presented in Table 1; the groups did not differ except in weight and FAAM-Sports score. The Office of Research Compliance at the university approved the study, and participants volunteered and provided consent under this approval.

View Figure

• Download as Powerpoint
• Download Figure

Table 1 Participant Baseline Demographics (N = 18; Mean ± SD)

View Fullscreen

Procedures

Eligible participants completed the pretest FAAM-ADL and FAAM-Sports questionnaires, the SEBT, and the weight-bearing JPS blocks. After the last exercise session, posttest measurements were completed within the week (average = 3 days). The posttest was performed to the same specifications as the pretest. Participants received an intervention according to their group allocation. Only the primary researcher was blinded to treatment allocation. Participants met with the researcher 3 times a week over a 4-week period for approximately 30 minutes per session to perform the PHSB or SLB program. The injured ankle was used for the training sessions. If a participant reported bilateral ankle instability, the self-reported worse limb was used for training and analysis. All exercises were performed at a single testing location for environmental control. One researcher administered and supervised all testing and exercising sessions for the primary outcomes of FAAM-ADL, FAAM-Sports, JPS blocks, and SEBT. This researcher was not blinded to the treatment allocation.

Disability Questionnaires: FAAM-ADL and FAAM-Sports

The participants completed the FAAM-ADL and FAAM-Sports as previously described.^28–30 The FAAM scores were based on grading criteria established for the instrument.^28,29 The FAAM-ADL and FAAM-Sports are reliable and valid responsive measures of self-reported physical function in healthy participants and are able to detect deficits associated with CAI.^28,29 The minimally clinically important differences that patients perceived for the FAAM-ADL and FAAM-Sports subscales were 8% and 9%, respectively.²⁸

Dynamic Balance

The SEBT was used to evaluate dynamic balance in 3 directions: A, PM, and PL. Extensive descriptions of test administration have been provided by previous authors.^10,31–33 Individuals performed 3 trials in each direction. The directional order was determined by participants randomly drawing 3 index cards. Each person began with the involved leg in the center of the grid. Each reach distance was measured from a mark on the tape as the distance from the center of the grid to the point of maximum excursion of the reach leg. The trial was discarded and repeated if the researcher felt the participant was using the reach leg for a substantial amount of support at any time, the stance foot did not remain completely on the ground, or the participant was unable to maintain balance on the support leg throughout the trial. The participant's true leg length was measured in accordance with previously established methods so that we could calculate normalized reach distances.^10,31–33 The normalized mean of 3 trials for each direction was used for analysis. The SEBT has been reported to have high reliability and validity in detecting reach deficits between athletes with CAI and healthy athletes.^33–35

Joint Position Sense Blocks

Joint position sense in the involved leg was determined using the weight-bearing, sloped-surface block method (Figure 2).²⁴ A custom-built device consisted of 2 sloped-surface boards 30 cm (length) × 30 cm (width) × 1.5 cm (height) connected by a hinge to allow block placement and a set of blocks with sloped surfaces at angles between 0° and 25° in increments of 2.5°. A total of 11 angles were used for PF, DF, and INV, and 5 angles were used for EV.²⁴ Of 36 total trials, 21 were in the sagittal plane and 15 in the frontal plane; 10 were for PF, 10 for DF, 10 for INV, 5 for EV, and 1 for neutral (0°), with 1 measure in each plane.

View Figure

• Download as Powerpoint
• Download Figure

Following the testing protocol described by Sekizawa et al,²⁴ participants were provided reference angles of 0°, 12.5°, and 25° for PF, DF, and INV as well as 5° and 10° for EV. These reference angles were given again at the halfway point of each trial. Trials were performed for PF and DF first, followed by INV and EV, using a randomized order of angles. Each trial used a combination of angle and direction. Direction was determined by positioning the angled surface board to represent the 4 directions. Participants self-reported the angle and direction. Data for JPS were calculated by taking the estimated angle minus the actual slope angle to equal the estimated angle error.²³ All angles in each of the 4 directions were combined to form a continuum of movement and use the same representation with positive and negative values for the 4 directions as described by Sekizawa et al.²⁴

Interventions

The PHSB Program

We used the 4-week progressive dynamic balance-training program developed by McKeon et al¹⁰ for those with CAI (Table 2). The exercises consisted of single-limb hops to stabilization at 18, 27, and 36 in (46, 69, and 91 cm) in combinations of 4 different directions, hop to stabilization and reach, unanticipated hop to stabilization using a 9-marker grid, and single-limb–stance activities with eyes open and closed and on compromising surfaces. These tasks were designed to challenge the participants' ability to maintain a single-limb stance while landing from a hop under various balance and compromising conditions.¹⁰ Before progression to the next level of difficulty within each of the 4 components, error-free completion of the task had to be demonstrated. The required number of error-free repetitions varied within each of the 4 components. For this intervention, participants did not complete the programs in their entirety. Participants in this group completed single-limb hops to stabilization, hop to stabilization and reach, single-limb activities, and up to level 5 for unanticipated hop to stabilization.

Table 2 Progressive Hop-to-Stabilization Balance Program¹⁰

View Fullscreen

The SLB Program

We used a rehabilitation protocol developed by Hale et al²⁰ for patients with CAI that involved stretching, strengthening, functional tasks, a home exercise program, and the neuromuscular control exercises. For the purpose of this study, only the neuromuscular control component of the rehabilitation program from Hale et al was used; thus, a modified program was followed (Table 3). The exercise components consisted of single-limb stance for 60 seconds with 2 repetitions; single-limb stance with a ball toss; single-leg stance kicking against resistance in 4 directions, 3 times, with 5 kicks in each direction; and step-downs on a 6-in (15.2-cm) step with the single limb in 4 directions, including 2 sets of 5 repetitions in each direction. Progression was based on the exercise component used and varied with the number of tosses and surfaces, increasing resistance, or the height of the step. An error occurred when the participant touched down with the opposite limb, demonstrated excessive trunk motion (>30° lateral flexion), removed the arms from across the chest during specified activities (components 1, 3, and 4), or braced the nonstance limb against the stance limb. For this intervention, participants did not complete the programs in their entirety. Participants in this group went up to step-down with single-limb stance.

Table 3 Single-Limb Balance Program²⁰

View Fullscreen

Statistical Analysis

The analysis was conducted using an intent-to-treat approach and included all randomized participants. Descriptive statistics were means and standard deviations. Two 2 × 2 repeated-measures analyses of variance (ANOVAs; time × group) were used for the FAAM-ADL and the FAAM-Sports scores. Three 2 × 2 repeated-measures ANOVAs (time × group), 1 in each direction, were used for the SEBT measurements. Four 2 × 2 repeated-measures ANOVAs (time × group) were used, 1 in each direction, for the JPS blocks. Post hoc tests were conducted if warranted. The α level for all analyses was .05. Effect sizes with bias-corrected Hedges g were calculated to offset the small sample size and to report on pre-intervention to postintervention group comparisons. Between-groups effect sizes compared each group's pretest-to-posttest change score divided by each group's pooled standard deviation. Hedges g was interpreted as small = 0.20, moderate = 0.50, or large = 0.80.³⁶ We used SPSS (version 24; IBM Corp, Armonk, NY) to analyze the data.

Patient-Oriented Outcomes

For the FAAM-ADL, no interactions or group main effects were observed. However, for the FAAM-Sports, despite no interaction, the SLB group showed more improvement (SLB posttest = 82.65 ± 5.1 versus PHSB posttest = 71.69 ± 9.4; P = .006). A time main effect was noted. After completing the programs, both groups improved from pre-intervention to postintervention for the FAAM-ADL (SLB pretest = 84.3 ± 3.1, posttest = 87.0 ± 2.4; PHSB pretest = 83.6 ± 5.2, posttest = 88.8 ± 4.7; P < .001) and FAAM-Sports (SLB pretest = 75.6 ± 5.7, posttest = 82.6 ± 5.1; PHSB pretest = 65.5 ± 8.4, posttest = 71.6 ± 9.4; P < .001). Large time effect sizes were present for FAAM-ADL, while for FAAM-Sports, the SLB group demonstrated a large effect size and the PHSB group, a moderate effect size. Between-groups effect sizes were small to moderate for both patient-oriented outcomes, and all 95% confidence intervals (CIs) crossed zero.

Clinically Oriented Outcomes

The FAAM Questionnaire

Scores on the FAAM-ADL and FAAM-Sports increased from pre-intervention to postintervention based on group membership, but the groups did not differ. Despite the increase in the pre-intervention to postintervention scores and the moderate to large effect sizes, clinically relevant changes over the course of the 4 weeks were not noted in the 2 groups, as minimally clinically important difference values were not reached with the FAAM-ADL (PHSB = 5%, SLB = 2.6%) or FAAM-Sports (PHSB = 6.1%, SLB = 7%) based on values reported by Martin et al.²⁸ Except for the magnitude of change, these changes were somewhat consistent with the results of other researchers who studied balance training using the FAAM-Sports or the Foot and Ankle Disability Index-Sport. Hale et al²⁰ demonstrated more improvement in the CAI rehabilitation group than the CAI control group for the Foot and Ankle Disability Index-Sport. However, the magnitude of change was greater at 20% improvement compared with the 7% improvement we found. We used a modified Hale et al program (lacking stretching, strengthening, functional tasks, and a home exercise program), so it is plausible to consider including these components, in addition to neuromuscular control, in a traditional rehabilitation program.³⁷ Improvement was consistent with what McKeon et al¹⁰ reported using the dynamic balance-training program, but again, the magnitude of change was greater at 15%. Although the dynamic balance-training program used in the McKeon et al program was identical to ours, the participants were different. We studied high school athletes with CAI, whereas McKeon et al¹⁰ studied recreationally active collegiate participants with perhaps a longer history of CAI. Also, our participants may not have experienced greater environmental or task constraints related to CAI, which would have left little opportunity for improvement. Even though both interventions lasted 4 weeks, it appeared that the participants were starting to trend toward a clinically meaningful change.

Dynamic Postural Control

Both groups increased their pretest to posttest normalized reach distances in all directions, but neither group increased more than the other. This result is similar to the findings of Hale et al²⁰ and McKeon et al,¹⁰ who reported improvements in SEBT reach distances after a 4-week rehabilitation program for participants with CAI. In the dynamic balance-training group, McKeon et al¹⁰ showed that reach distances improved in the PM and PL directions of the SEBT but not in the A direction. In the Hale et al²⁰ study, participants with CAI improved after a self-designed traditional 4-week rehabilitation program; they displayed greater SEBT reach improvements for the PL, PM, and lateral directions versus the uninvolved limb and healthy participants. Improvements were evident in both training groups: the balance-training programs incorporated single-limb stance, especially the PHSB program with a hop to stabilization and then a reach back to the starting point. This exercise mimics the reach task on the SEBT. To enhance range of motion (ROM) and increase reach distances, further studies are needed. Although not used in this study, grade III joint mobilizations as described by Hoch et al³⁷ have increased ROM and reach distances in the A, PM, and PL directions of the SEBT. We incorporated isolated exercises, whereas a more comprehensive rehabilitation program³⁸ addressing mechanical restrictions, plantar cutaneous deficits, strength, and static postural control³⁹ may both improve reach distances and result in better functional outcomes.

The PHSB group demonstrated greater changes based on the large time effect sizes for the A and PL directions; the SLB group had a large time effect only for the PL direction. McKeon et al¹⁰ found a large group effect for the PM and PL directions and a minimal to small effect in the A direction.^10,20 Thus, effect sizes may be less informative than the pre-intervention to postintervention improvement in all 3 directions, which ranged from 3.16 to 5.0 in the PHSB group and from 3.99 to 5.07 in the SLB group. These values are fairly consistent with those of other studies^20,38 after a 4-week rehabilitation program.

Joint Position Sense

Previous authors^16,18,19 hypothesized that CAI was associated with disrupted mechanoreceptors^13,22,23 in the articular surfaces, cutaneous surfaces, lateral ankle ligaments, and joint capsules of the talocrural and subtalar joints. Disruption of JPS is partially due to damaged mechanoreceptors sending corrupt signals to the central nervous system.¹¹ This sensorimotor impairment may become evident when the individual bears weight and the efferent responses are received by the muscles. Upon loading, there may be an alteration in foot biomechanics, leading to a lack of adaptability of the joint to change in surface, predisposing the body to become more unstable and leading to incorrect foot landing.²⁴ This may result in an underestimation or overestimation of foot position on landing.

Deficits in JPS have been noted with aging²³ and in healthy participants.²⁴ When healthy participants with no history of LAS were evaluated for JPS using the weight-bearing, sloped-surface block method, PF and INV were underestimated.²⁴ Thus, because the mechanism of injury for an LAS may involve moments that include PF and INV,⁴⁰ underestimation may be more apparent in a participant with CAI. We observed a decrease in angle error for PF and INV. This is important clinically because the most common positions for the foot and ankle during an LAS are PF and INV.²⁴ With less absolute-angle error underestimation, the positions of the foot and ankle may be more compressed, thereby reducing the risk of injury when moving from an unloaded to a loaded condition upon landing.⁴⁰

Joint position sense has not been evaluated previously using the 2 balance-training programs compared here. Over the course of the 4-week training period, participants became more aware of the differences between the actual angle and the estimated angle. This may be related to the stationary kicking, the hops to stabilization in the sagittal and frontal planes, or the step-downs in 4 directions while in stationary or dynamic single-limb stance. Joint position sense was assessed in a weight-bearing, single-legged stance on the block as the angle and the position of the block changed in randomized order for each trial. This was similar to the single-limb stance maintained during the intervention and perhaps aided in postintervention improvement.

Clinical Implications and Limitations

Both experimental groups improved from pretest to posttest on all patient- and clinically oriented outcomes, which may have more relevance than statistical significance in the clinical setting when considering which rehabilitation program to use. In 4-week programs for participants with CAI, improvements were evident in self-reported ADL and sport function, dynamic postural control, and JPS. Thus, the isolated static and dynamic exercises used in this study should be incorporated into a program when only 1 area of impairment, namely sensorimotor impairment, is present. Combining multiple treatment techniques, as is normally done in the clinical setting, should be the mainstay of rehabilitation for patients with CAI. Therefore, we advocate for other areas of impairment to be addressed, including functional activities, ROM, joint mobilization, strength, gait retraining,³⁸ plantar cutaneous deficits,³⁹ and static postural control.

The primary limitation of our study is that the results cannot be generalized to other participants with CAI due to the sample size. Our sample of convenience consisted of injured, physically active individuals from only 1 high school and 1 university. Scheduling sessions during data collection was challenging due to snow days and poor road conditions, which prevented the participants from keeping some of their appointments. A 1-week delay occurred between pretesting and the start of the 2 programs because of school closures, during which the participants were inactive; when school resumed, the 4-week program proceeded as planned. Adaptations of traditional and dynamic balance-training programs might vary based on individual factors including previous injury, rehabilitation history, and current fitness level. Although the lengths of the traditional and dynamic balance-training programs were similar to those tested in previous studies, more randomized controlled studies evaluating the 2 programs as well as other multiple treatment techniques for patients with CAI should be conducted.