Background
Lateral ankle sprains (LAS) are very common in the active population [
1] with nearly two million annual reported cases in the United States of America [
2]. However, the true incidence of LAS is a lot higher considering that 55 to 64% of individuals sustaining a LAS will not seek professional health care [
3,
4]. Seventy percent of individuals will sustain at least one recurrent sprain after the index LAS [
5] and 40% will develop chronic ankle instability (CAI) [
6]. CAI is characterised by a propensity for recurrent LAS at least 12 months after the index LAS, frequent episodes of ankle giving way, persistent symptoms such as pain, swelling, limited motion, weakness and diminished self-reported function [
7]. CAI leads to altered balance control, ankle neuromuscular function, and lower limb biomechanics during locomotion [
8‐
10]. These deficits place individuals with CAI at greater risk of sustaining recurrent sprains [
11], developing post-traumatic ankle osteoarthritis [
12] and consequently decreasing their health-related quality of life [
13].
Since several decades, clinical gait analysis provides objective useful information about the mechanical and neuromuscular deficits in individuals with CAI [
8]. During walking, individuals with CAI exhibit greater ankle inversion angles [
14‐
16] which place more load under the lateral part of the foot [
17,
18]. These biomechanical deficits could predispose their ankle to give way and sustain a recurrent sprain [
8]. To attenuate this cascade of biomechanical events, a greater activity of the peroneus longus muscle, the main evertor of the ankle complex, is observed in these individuals [
14,
17,
18] and could represent a protective mechanism [
14]. Changes in other lower limb muscles’ activity in individuals with CAI are inconsistent in previous studies [
8]. It could be attributed to the spectrum of disability associated with CAI [
7] and by the great disparity in reported EMG analyses [
8]. There is a need to standardise the processing and normalisation methods in studies investigating EMG of the lower limbs in individuals with CAI to increase the external validity of the published results.
Previous studies reported inconsistent knee biomechanics deficits in individuals with CAI [
14‐
16,
19,
20]. A few studies found no differences between the knee biomechanics of individuals with CAI compared to healthy controls [
14,
16,
19]. Also, two recent studies reported conflicting results when comparing knee moments between these two groups [
15,
20]. Moisan et al. [
20] reported greater and Son et al. [
15] reported smaller knee abduction moments in individuals with CAI compared to controls during walking. Further studies are needed to better understand the proximal adaptations to the lower limb biomechanics during walking. Also, only a few studies investigated the biomechanics of the lower limbs of individuals with CAI compared to healthy controls [
15,
20] using a comprehensive approach including kinematic, kinetic and EMG measures. Consequently, the theoretical framework explaining the biomechanical adaptations in individuals with CAI is mostly based on the results of studies including heterogeneous methods and participants’ specificities (e.g., level of disability). More studies using a comprehensive approach are needed to identify the biomechanical adaptations associated with CAI for the same population and investigate how kinematic, kinetic and EMG adaptations interact.
The objective of this case-control study was to identify the kinematic, kinetic and EMG differences between individuals with CAI and controls during overground walking. It was hypothesised that individuals with CAI will exhibit greater ankle inversion angles and peroneus longus muscle activity as well as smaller ankle inversion moments compared to controls.
Discussion
This case-control study aimed to identify lower limb kinematic, kinetic and EMG differences between individuals with CAI and healthy controls during walking. This study was deemed important as only a few studies compared the biomechanics of the lower limbs of individuals with and without CAI using a comprehensive approach including kinematic, kinetic and EMG measures [
15,
20,
30]. Using such comprehensive approach allowed the identification of biomechanics adaptations associated with CAI for the same population and investigation of how kinematic, kinetic and EMG adaptations interacted. Also, knee moments differences between individuals with and without CAI during gait were inconsistent and contradictory in previous studies. Indeed, Moisan et al. [
20] reported greater and Son et al. [
15] reported smaller knee abduction moments in individuals with CAI compared to controls during walking and thus further research was needed to better understand the impact of CAI on this joint. Our main hypotheses were that participants with CAI would exhibit greater ankle inversion angles and peroneus longus muscle activity as well as smaller ankle inversion moments during the stance phase of gait compared to healthy control participants. Our results fully support these hypotheses. The main findings were greater ankle inversion angles, reaching a mean difference of 4.10
o at 20%SP (see Additional file
1), in participants with CAI compared to controls which are consistent with previous results [
14‐
16]. However, in contrary to previous findings, no significant differences in ankle inversion angles [
15,
33] and moments [
15] were observed during the pre-swing phase of gait. Delahunt et al. [
14] hypothesised that to prevent excessive ankle inversion movements, the peroneus longus muscle of individuals with CAI activates strongly during the stance phase. Our results are consistent with this hypothesis. Indeed, peroneus longus muscle activity was greater for participants with CAI from 60 to 76%SP, which is about the period during which the between-group differences in ankle frontal movements started to decrease (see Fig.
1). During the pre-swing phase, the ankle frontal movements’ curves of the CAI and control groups nearly overlapped (see Fig.
1), highlighting an efficient biomechanical compensatory strategy in individuals with CAI during the latter half of the stance phase. However, during the first half, participants with CAI could be at greater risk of sustaining a LAS as the ankle is in a more vulnerable position (i.e. more inverted). Indeed, if an external perturbation (e.g., stepping on an uneven terrain) results into increased ankle inversion in individuals with CAI, they would be at more risk of sustaining a LAS as their ankle is already in a more inverted position and their peroneus longus muscle presents a reflex latency and an electromechanical delay of activation [
34]. Thus, clinicians should probably focus on the biomechanical deficits during the first part of stance.
Consistent with Monaghan et al. [
16] results, participants with CAI exhibited ankle eversion moments whereas controls exhibited ankle inversion moments during most of the stance phase. Even though between-group differences for ankle frontal moments were only statistically significant between 40 to 78%SP, one can observe that both curves are separated from each other for most of the stance phase (5 to 85%SP) (see Fig.
2) with moderate to strong Cohen’s d effect sizes (see Additional file
2). Even if the differences are not statistically different throughout this period, such differences could perhaps be clinically meaningful. Changes in kinetic control of the ankle also represent a protective mechanism to prevent LAS and explain the greater peroneus longus muscle activity [
8].
In agreement with previous findings from our research group, participants with CAI exhibited greater knee abduction moments during the early stance [
20]. This result suggests that ankle biomechanical changes in the frontal plane induce a knee compensatory mechanism in individuals with CAI. No between-group knee kinematics differences were observed, but the greater knee abduction moments in individuals with CAI could represent an attempt of the locomotor system to attenuate the load and strain on the lateral structures of the ankle using the proximal segment in the kinetic chain. It could also represent an attempt to stabilise the knee during the early stance and to better attenuate the impact forces, known to be greater in individuals with CAI [
15]. No between-group difference was observed for the vastus lateralis muscle which is consistent with previous work [
20]. However, contrary to our finding, Son et al. [
15] reported a decreased vastus lateralis muscle activity of less than 4% in individuals with CAI compared to controls. As the differences were only observed during a period when the activity of the vastus lateralis muscle is rapidly decreasing (i.e., around 18 to 90%SP), the results most likely have little clinical applicability. Further studies investigating the activity of the abductor and adductor muscles are needed to better understand the role of the knee during walking in individuals with CAI. In our study, we observed no between-group differences for the gluteus medius muscle activity, which is consistent with previous results [
35,
36]. However, other studies reported greater [
17] or lower [
15,
20] gluteus medius muscle activity in individuals with CAI compared to controls. The disparity in EMG normalisation methods, dependent variables as well as the different experimental conditions (i.e., shod and barefoot) may perhaps explain the inconsistencies noted in previous results. Further studies are needed to better understand the differences in gluteus medius muscle activity using standardised EMG assessment methods or novel and promising methods such as ultrasound imaging [
37].
Our study allows validation and precision of previous results regarding the biomechanics of the lower limb of individuals with CAI during walking, namely the increased ankle inversion angles and moments as well as peroneus longus muscle activity. In clinical contexts, both the cause and consequence of CAI should be addressed. Sensorimotor deficits in motor-neuron pool excitability, reflex reactions, muscular strength and proprioception [
38] could be responsible for the changes in gait movement strategy in individuals with CAI [
15]. Accordingly, the therapeutic goal in clinical gait rehabilitation should be to address the faulty ankle movements and restore proper sensorimotor function. Impairment-based rehabilitation programs including gait retraining [
39,
40] or exercise regimen targeting sensorimotor deficits [
41], prescribed independently or together [
42], have shown promising results. External feedback during gait retraining reinforces ideal repetitive actions via optimised sensorimotor loops [
43] and sensorimotor training allows targeting deficits in both sensory and motor aspects of sensorimotor control [
41]. Our results will inform the development of future efficacy trials aiming at determining to what extent addressing the biomechanical particularities of individuals with CAI will result in improved clinical outcomes.
Our results should be interpreted in light of a few limitations. First, forefoot and midfoot segments kinematics were not analysed. Individuals with CAI present kinematic changes for these segments compared to controls during walking [
44,
45]. Thus, between-group differences could have been present in our study but could not be observed using our design. Second, as our results suggest that the ankle biomechanical changes induce proximal biomechanical compensations in individuals with CAI, more studies should also investigate between-group differences at the knee and hip joints. This is especially important as it was recently reported that individuals with CAI exhibit a hip-dominant strategy to generate power allowing forward acceleration of the lower limbs during gait [
15]. Third, the mean age of the recruited participants was 25.5 and 23.7 years for the CAI and control groups, respectively. Our results could perhaps not be generalisable for older individuals with CAI. Fourth, inherently to the chosen research design, it is unclear if the biomechanical differences found in our study are a cause or consequence of CAI and thus prospective studies are warranted. Fifth, even though the IPAQ scores were not significantly different between groups, high within-group variability was observed. The inclusion of individuals with heterogenous physical activity levels in each group may have contributed to decreasing the homogeneity of our data and thus the ability of our analyses to reach the threshold of significance. Sixth, as we did not include a group of copers, our study does not provide insights regarding the biomechanical differences between individuals who sustained a LAS and healed normally and those who developed CAI. Seventh, even though the biomechanical deficits exhibited by individuals with CAI during overground and treadmill walking are similar [
8], our results may perhaps not be entirely generalisable during treadmill walking.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.