Whole-Body–Vibration Training and Balance in Recreational Athletes With Chronic Ankle Instability

Participants

Fifty recreational athletes with self-reported CAI volunteered for the study. They were assigned by concealed random allocation using random numbers generated by online software (http://www.randomization.com) to 1 of 3 groups: vibration (VIB; 11 men, 6 women; age = 22.4 ± 2.6 years, height = 172.0 ± 8.3 cm, mass = 70.2 ± 8.2 kg), nonvibration (NVIB; 10 men, 6 women; age = 21.8 ± 2.1 years, height = 171.3 ± 9.0 cm, mass = 66.2 ± 10.1 kg), or control (CON; 12 men, 5 women; age = 23.6 ± 3.4 years, height = 172.7 ± 10.8 cm, mass = 70.6 ± 11.7 kg; Table 1). Sample size was calculated based on the work of Sefton et al,¹⁸ who measured posteromedial reach in participants with CAI. The minimal number of participants required to attain a power of 0.8 and a bilateral α level of .05 was calculated to be 16 per group.

Table 1 Baseline Characteristics, Mean ± SD (N = 50)

View Fullscreen

Following the selection criteria for patients with CAI,¹⁹ we included participants if they reported a history of at least 1 substantial ankle sprain (the most recent injury must have occurred more than 3 months before study enrollment), 2 or more episodes of the ankle “giving way” in the 6 months before the study, and a score ≤24 on the Spanish version of the Cumberland Ankle Instability Tool (CAIT).²⁰ In participants with bilateral ankle instability, the ankle with the lower score was selected. Exclusion criteria were a history of previous surgery to the musculoskeletal structures of either lower extremity; fracture in either lower extremity requiring realignment; or acute musculoskeletal injury to the joints of the lower extremity in the 3 months before the study that affected joint integrity and function, resulting in at least 1 lost day of desired physical activity.¹⁹ Participant flow is presented in Figure 1.

View Figure

• Download as Powerpoint
• Download Figure

All participants provided written informed consent. The study was approved by the Ethics Committee of Clinical Research at the Hospital Complex in Toledo (Spain) and was registered as trial NCT02794194 at ClinicalTrials.gov on June 8, 2016.

Procedures

A clinical trial was performed using a randomized, between-groups design. Participants were assessed at 3 times: pretraining (Pre), posttraining 1 (Post1; 48 hours after the last training session), and posttraining 2 (Post2; 6 weeks after the last training session). Measurements were performed in the following order: body-composition analysis, Biodex Balance System test (BBS; Biodex Medical Systems, Shirley, NY), and Star Excursion Balance Test (SEBT). Assessors (R.S.G., F.J.D., C.R., P.E.) and the researcher (J.A.V.) who performed the statistical analysis were blinded to group allocation.

Training Protocol

Participants followed a 6-week balance-training protocol for an unstable ankle based on previous research (Table 2).^6,21 Exercises were performed barefoot on a BOSU Balance Trainer (BOSU, Ashland, OH) 3 days each week (with 48 hours between sessions; Figure 2). All exercises were carried out only on the unstable ankle and were the same for both experimental groups. Participants in the NVIB group trained with the BOSU on the floor, whereas participants in the VIB group trained on an Excel Pro vibration platform (Fitvibe, Bilzen, Belgium). Synchronous WBV was applied. The training program consisted of 3 series of four 45-second exercises with a 45-second rest between exercises. A repetition-based balance-training–progression style¹⁴ was followed. The level of difficulty of all exercises was increased after 3 weeks.²² Frequency was also increased by 5 Hz every 2 weeks. Amplitude was increased from 2 to 4 mm after the first week and then maintained for the remainder of the study. Participants in the CON group were encouraged to continue their levels of physical activity and refrain from beginning a new training program.

Table 2 Exercise Scheme of the 6-Week Balance-Training Program With Whole-Body Vibration

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

Biodex Balance System Test

Ankle balance was assessed using the BBS, which consists of a mobile platform that allows up to 20° of surface tilt in 360° range of motion. The platform, which is interfaced with computer software (version 1.32; Biodex Medical Systems) generates the Overall Stability Index (OSI), Anterior-Posterior Stability Index (APSI), and Medial-Lateral Stability Index (MLSI) from the degree of tilt. The APSI and MLSI represent platform displacements from the horizontal in the sagittal (Y) and frontal (X) planes, respectively, and the OSI is a composite of the APSI and MLSI. The following formulas were used to generate the indices: OSI = [(Σ(0 – Y)² + Σ(0 – X)² / samples)]^0.5, APSI = [(Σ(0 – Y)² / samples)]^0.5, and MLSI = [(Σ(0 – X)² / samples)]^0.5.²³ Higher values represented poorer stability, whereas lower values represented better stability.

The test was performed at level 8 with participants barefoot in single-legged stance. They were instructed to step on the BBS platform with their eyes open, assume a comfortable position while keeping their knees slightly flexed (15°), look straight ahead at the monitor, and place their hands on their hips. Foot-position coordinates were registered to ensure that the same position was used for all tests. We instructed participants to keep a cursor, which represented the center of the platform, in the center of the bull's eye on a visual feedback screen. Only 3 practice trials were performed to reduce any learning effects,²³ and 3 test evaluations were then performed. Each trial lasted 20 seconds with a 10-second rest between trials. The average of the 3 test evaluations was used for data analysis. Failed trials were not recorded and were removed from the data analysis. A trial was considered a failure if the participant used the handlebars of the platform to maintain balance, put the free foot on the platform, or completely lost his or her balance.

Star Excursion Balance Test

The SEBT was performed as previously described.^3,24 We created a grid with 8 tape measures extending from the center at 45° from each other.³ Each measure was labeled according to the direction of excursion in relation to the standing limb. Oral and video instructions were given to the participants before the test. Participants stood barefoot with the stance foot centered on the grid and aligned with the anterior and posterior directions. While maintaining single-legged stance on the unstable ankle, they reached with the contralateral limb to lightly touch as far as possible in the chosen direction with the most distal part of the foot and returned to a bilateral stance. Only 5 directions of the SEBT were analyzed to avoid redundancy among the 8 directions.³ The anteromedial, medial, and posteromedial have been proposed to be the most sensitive directions in participants with CAI.^3,18,25 The anterior and posterolateral directions have also been used in some studies^2,26 involving participants with CAI. A composite score based on the 5 directions was calculated to quantify the overall change. A standardized protocol of 4 practice trials, followed by 3 test trials, was performed in each of the 5 directions to minimize the learning effect.²⁴ Reach distances were measured by the same researcher (R.S.G.), who marked the tape measure. The average of the 3 test trials was normalized for length of the stance limb and used for analysis. While participants were lying supine, we measured their limbs from the anterior-superior iliac spine to the distal tip of the medial malleolus using a standard tape measure. A trial was discarded and repeated if (1) balance was lost, (2) any part of the foot was lifted or moved from the center grid, (3) the hands did not remain on the hips, or (4) the reach limb provided support when touching down.

Body-Composition Analysis

Height was measured using a wall-mounted stadiometer (model 220; Seca, Hamburg, Germany). Body mass (with the participant in underwear) was measured using a digital balance (model 707; Seca) with a weighing accuracy of 0.1 kg. Total and regional body compositions were measured by dual-energy x-ray absorptiometry (DXA; model Lunar iDXA; General Electric Healthcare, Fairfield, CT) using a standardized protocol specified by the manufacturer. Lean mass, fat mass, percentage of fat, bone mineral content, and bone mineral density for the total body and for the limb with CAI were obtained using enCORE software (version 13.40; General Electric Healthcare). Standard calibration procedures were performed before each testing session by the same technician (P.E.).

Statistical Analysis

The normality of each variable was initially tested using the Shapiro-Wilk test. All variables displayed normal distributions. A 2-way repeated-measures analysis of variance was performed for all outcome variables to analyze the interaction among groups (VIB, NVIB, CON) and the time of assessment (Pre, Post1, Post2). When differences were established, we applied a post hoc Bonferroni multiple-comparisons test. The effect size (ES) was calculated for all pairwise comparisons according to the formula proposed by Glass et al.²⁷ When a pairwise comparison was performed between the NVIB and VIB groups, we used a pooled standard deviation for the calculations. The magnitude of the ES was interpreted using the scale of Cohen²⁸: small (<0.2), medium (0.5), and large (>0.8). All data were presented as means ± standard deviations. The α level was set at .05. Statistical analysis was performed using SPSS (version 22.0; IBM Corp, Armonk, NY).

Biodex Balance System Test

Results for the BBS are presented in Table 3. We observed no differences among the 3 groups for any of the 3 measurements (P > .05). Within-group analysis showed decreases in the VIB group between Pre and Post1 in the OSI of −18.69% ± 21.58% (P = .01; ES = −0.96; 95% confidence interval [CI] of the mean difference [MD] of the score = −0.57, −0.06) and the APSI of −13.28% ± 25.34% (P = .02; ES = −0.68; 95% CI of MD = −0.33, −0.03) and between Pre and Post2 in the OSI of −20.14% ± 24.62% (P = .003; ES = −0.92; 95% CI of MD = −0.56, −0.09) and the APSI of −15.34% ± 27.42% (P = .03; ES = −1.33; 95% CI of MD = −0.42, −0.02). Whereas we did not demonstrate differences in the MLSI, we found large ESs between Pre and Post1 in the VIB (ES = −0.87; 95% CI of MD = −0.43, 0.06) and NVIB (ES = −0.91; 95% CI of MD = −0.50, 0.00) groups.

Table 3 Balance Test Results (Mean ± SD)

View Fullscreen

Star Excursion Balance Test

Results for the SEBT are presented in Table 3. We observed no differences (P > .05) among the 3 groups for any of the 3 measurements. However, moderate to large ESs were present in several directions at Post1 between the VIB and CON groups (medial direction: ES = 0.61; posteromedial direction: ES = 0.73; composite score: ES = 0.54) and NVIB and CON groups (anteromedial direction: ES = 0.82; medial direction: ES = 0.58; posteromedial direction: ES = 0.75; composite score: ES = 0.80).

Within-group analysis of the VIB group showed increases (P < .05) with moderate to large ESs (Figure 3) between Pre and Post1 in the medial direction of 4.93% ± 3.78% (P = .008; ES = 0.85; 95% CI of MD = 0.97, 7.69 cm), posterolateral direction of 5.21% ± 9.43% (P = .04; ES = 0.52; 95% CI of MD = 0.07, 7.88 cm), and composite score of 3.72% ± 4.09% (P = .01; ES = 0.68; 95% CI of MD = 0.60, 5.63 cm) and decreases between Post1 and Post2 in the medial direction of −2.93% ± 4.97% (P = .03; ES = −0.43; 95% CI of MD = −5.30, −0.27 cm), posterolateral direction of −3.09% ± 6.07% (P = .04; ES = −0.38; 95% CI of MD = −5.75, −0.14 cm), and composite score of −2.30% ± 2.64% (P = .007; ES = −0.47; 95% CI of MD = −3.68, −0.48 cm).

View Figure

• Download as Powerpoint
• Download Figure

In the NVIB group, increases with moderate to large ESs (Figure 3) were shown between Pre and Post1 in the medial direction of 7.36% ± 10.34% (P < .001; ES = 0.78; 95% CI of MD = 2.34, 9.27 cm), posteromedial direction of 8.75% ± 13.53% (P = .002; ES = 0.83; 95% CI of MD = 2.38, 11.79 cm), posterolateral direction of 5.32% ± 7.93% (P = .03; ES = 0.43; 95% CI of MD = 0.32, 8.37 cm), and composite score of 5.51% ± 6.61% (P < .001; ES = 0.58; 95% CI of MD = 2.04, 7.22 cm). Decreases between Post1 and Post2 were noted in the anterior direction of −3.70% ± 6.70% (P = .01; ES = −0.40; 95% CI of MD = −6.25, −0.70 cm), anteromedial direction of −3.05% ± 4.59% (P = .002; ES = −0.39; 95% CI of MD = −4.76, −0.96 cm), medial direction of −4.54% ± 5.09% (P = .001; ES = −0.47; 95% CI of MD = −6.81, −1.64 cm), posteromedial direction of −4.03% ± 4.91% (P = .002; ES = −0.40; 95% CI of MD = −6.60, −1.27 cm), posterolateral direction of −3.64% ± 5.07% (P = .01; ES = −0.35; 95% CI of MD = −6.37, −0.60 cm), and composite score of −3.89% ± 3.92% (P < .001; ES = −0.41; 95% CI of MD = −5.24, −1.95 cm).

Body-Composition Analysis

Body-composition variables are presented in Table 4. No differences occurred among groups before or after the training (P > .05). We also observed no group-by-time interaction effect for any of the variables (P > .05).

Table 4 Body-Composition Variables (Mean ± SD)

View Fullscreen

Balance

Deficits in balance have been reported in participants with CAI.^1,29 Therefore, improving their balance could decrease the risk of sustaining a new ankle sprain. Schiftan et al⁴ showed that balance training could effectively reduce these deficits.

The BBS has been proposed as a useful method to assess balance in participants with CAI.³⁰ Poor performance on the OSI at level 3 has been noted in participants with CAI as compared with healthy participants.³¹ Kim and Heo³² found that 4 weeks of balance training using virtual reality exercises enhanced the BBS balance indexes in participants with CAI. Similarly, after 12 weeks of proprioceptive training, healthy, young speed skaters displayed improvements.³³ In our study, we found that better performance on the OSI and APSI was maintained only by the VIB group at Post2. These results may indicate that adding vibration was more effective than balance training alone. Cardinale and Bosco⁷ noted that WBV training could enhance muscle-spindle sensitivity and excitability of the α and γ motoneurons. These adaptations could lead to reduced reaction time of the ankle-stabilizing muscles and the motor-unit recruitment thresholds,⁸ which may offer an advantage when adjusting position during the BBS test. To our knowledge, only Cloak et al⁶ combined WBV and balance training in participants with CAI and reported progress in those who trained with WBV and a wobble board compared with the wobble board alone. However, given that we did not use electromyography during the balance tests, we can only assume that these differences were caused by the addition of WBV.

We also found postintervention improvements in balance. However, improvements occurred in both the VIB and NVIB groups and were reported by Cloak et al⁶ after the combined training. Therefore, we cannot conclude that adding WBV training provided extra benefits compared with balance training alone.

The SEBT has also been described as useful for identifying deficits in participants with CAI who may benefit from rehabilitation to reestablish dynamic ability.¹ Participants with CAI displayed poorer performance on the SEBT than copers with lateral ankle sprain.² We found postintervention benefits in both training groups in the medial and posterolateral directions and the composite score. We also observed better performance in the posteromedial direction by the NVIB group. These results are similar to those found by previous researchers,²⁹ who saw improvements in the medial, posteromedial, and posterolateral directions after 4 weeks of rehabilitation that addressed range of motion, strength, neuromuscular control, and functional tasks. Similarly, Sefton et al¹⁸ demonstrated progress in the anteromedial, medial, and posteromedial directions after 6 weeks of balance training. In our study, we analyzed the persistence of the effects after the intervention ceased. Values tended to return to baseline levels, with decreases between Post1 and Post2 in the medial and posteromedial directions and the composite score in the VIB group and in the medial, posteromedial, and posterolateral directions and the composite score in the NVIB group. Han et al³⁴ also evaluated the persistence of the effects and reported that balance improvements were retained after training ceased. However, it is difficult to compare their results with ours because they used a force platform to assess balance and the intervention was based on elastic-tubing exercises. Furthermore, they recruited participants according to a history of ankle sprains, but they did not use a valid instrument, such as the CAIT, to determine the presence of CAI and performed the Post2 after 4 weeks instead of 6 weeks.

Body Composition

We observed no changes in total body or lower limb lean mass in any group after 6 weeks of training. Information on the effects of WBV training on body composition in young participants is limited, and the results conflict. On the one hand, researchers^35,36 have determined that WBV in young, healthy participants does not improve lean mass. Measures of vibration frequency and amplitude were similar to those in our study; however, we found some differences in the protocols. Our participants performed specific balance exercises, whereas participants in previous studies performed body-weight exercises targeted to the lower extremities and the trunk³⁵ or half-squat exercises combined with resistance training.³⁶ Still, increases in lean mass have been reported in other studies.^10,11,37 Roelants et al¹⁰ noted an increase in fat-free mass after participants completed a 24-week WBV training program in which they performed unloaded static and dynamic lower limb and upper limb exercises. However, given that body composition was assessed via underwater weighing instead of DXA analysis, it is difficult to compare results. Improvements in lean mass with only 6 weeks of WBV between sets of squat exercises have also been reported.¹¹ Comparing 2 amplitudes (2 mm and 4 mm) of WBV during 6 weeks of training, Martinez-Pardo et al³⁷ determined that lean body mass increased only with high amplitudes. In our study, high amplitudes were applied, but frequency values were between 30 and 40 Hz; Martinez-Pardo et al³⁷ maintained frequency at 50 Hz throughout the study.

Other researchers have also related WBV training to a decrease in the percentage of fat mass. Artero et al³⁶ found a decrease in the percentage of body fat after 8 weeks of WBV. However, their results must be interpreted with caution because they used skinfold-derived equations to measure body composition. Using DXA analysis, Lamont et al¹¹ reported a decrease in the percentage of leg fat, but their vibration measures (from 4 to 6 mm of amplitude and 50 Hz) were higher than ours. Contrary to these studies, we did not demonstrate changes in fat mass, supporting the results of Martinez-Pardo et al,³⁷ who also confirmed that 6 weeks of WBV training applied at 50 Hz with high-amplitude values did not lead to a decrease in fat mass. Regarding bone mass, Gilsanz et al³⁸ suggested that WBV could lead to increased bone mass in young women with low bone mineral density after 1 year of training. We observed no changes in bone mass after 6 weeks of WBV training, consistent with the findings of Torvinen et al,³⁹ who also concluded that 8 months of WBV training had no effect on the bones of young healthy adults.

Limitations

Whereas all participants had homogeneous characteristics, the intervention might not have challenged their sensorimotor systems equally. Furthermore, the vibration load was the same for all participants rather than being determined individually.⁴⁰ Adding a BOSU to the vibration platform could also have attenuated part of the effective energy produced by the vibration. The WBV has been associated with increases in testosterone and growth hormone levels; however, no physiological measurements were taken in our study. Furthermore, the training protocol was targeted at single-legged balance ability. Therefore, a hormonal change induced by the exposure to WBV would not necessarily have led to morphologic advancements without adding some type of resistance training. Another limitation was the small sample size, which might not have been adequate to detect postintervention differences among groups. It was also not possible to blind participants to group allocation, which may have influenced our results. Finally, given that the CAIT was administered only to select the participants, our results should be interpreted cautiously. We cannot affirm that our training protocol diminished the feeling of instability assessed by the CAIT.

Practical Applications

Balance impairments have been reported in participants with CAI. We observed that balance training on an unstable surface improved performance on the SEBT. Also, we found that the same intervention combined with WBV benefitted balance ability on the BBS. Therefore, implementing this kind of training should be considered in future interventions to reduce the risk of recurrent ankle sprains by improving the dynamic balance ability of participants with CAI.