Altered Biomechanics in Individuals With Chronic Ankle Instability Compared With Copers and Controls During Gait

A total of 54 volunteers aged between 18 and 35 years were recruited from the university setting and surrounding community for this case-control study (Table). Participants were assigned to the CAI (n = 18), coper (n = 18), or control (n = 18) group.

Recruits who met the inclusion criteria, based on recommendations from the International Ankle Consortium,⁷  were assigned to the CAI group. An LAS had to include symptoms of inflammation and at least 1 day of interrupted physical activity that was confirmed during the recruitment and screening session. Inclusion criteria for the CAI group were as follows: (1) a history of >1 LAS and persistent symptoms, including recurrent feelings of giving way, recurrent sprains, or feelings of instability; (2) initial LAS occurring at least 12 months before study enrollment; (3) the last sprain occurring ≥3 months before this study; (4) ≥5 responses of yes to questions on the Ankle Instability Instrument (AII), including a response of yes to question 1; and (5) scores of <90% and <80% on the Foot and Ankle Ability Measure (FAAM)–Activities of Daily Living (ADL) and the FAAM-Sport scales, respectively.⁷ 

Inclusion criteria for the coper group consisted of the following conditions: (1) a history of 1 LAS resulting in at least 1 day of interrupted physical activity, without residual symptoms caused by the initial LAS; (2) initial LAS occurring ≥12 months before study enrollment; (3) a response of yes to question 1 of the AII; and (4) scores of >99% and >97% on the FAAM-ADL and FAAM-Sport scales, respectively.¹⁹ 

Inclusion criteria for the control group were as follows: (1) no history of LAS, (2) a response of no to all questions on the AII, and (3) a score of 100% on both the FAAM-ADL and FAAM-Sport scales.

Exclusion criteria for all participants were a history of previous surgery to the musculoskeletal structures, a fracture of the lower extremity, peripheral neuropathy, or an acute injury to the musculoskeletal structures of the lower extremities in the 3 months before the study.⁷  All participants provided written informed consent, and this study was approved by the Institutional Review Board of Yonsei University (No. 7001988-201901-HR-429-04).

Before the study, we collected anthropometric measurements of height, body weight, leg length, knee width, and ankle width. For the kinematic measures, spatial trajectory data from 33 retroreflective markers (version 6, Visual3D; C-Motion, Inc) and tracking markers were collected. The markers were attached to the following landmarks: top of the iliac crest; anterior-superior iliac spine; posterior-superior iliac spine; center of the sacrum; greater trochanter; quadriceps tendon; lateral and medial epicondyles; tibial tuberosity; lateral and medial malleoli; posterior-superior, posterior-inferior, and lateral aspects of the calcaneus; and first, second, and fifth metatarsal heads (https://www.c-motion.com/v3dwiki/index.php?title=Marker_Set_Guidelines). The tracking markers attached to the calcaneus produced a more precise analysis of rearfoot motion in the frontal plane.

All participants were instructed to perform a 5-minute warm-up session and a 5-minute practice session for each task to adapt to walking and jogging. Before collecting kinematic data, we captured static posture for marker calibration using the anatomical position. To record the kinematics of lower extremities during gait, participants walked in a shod condition at a speed of approximately 1.34 ± 0.07 m/s^13,20  (±5%) and jogged at a speed of approximately 2.68 ± 0.14 m/s^15,20  (±5%) along the 8-m walkway with an embedded force plate (model ORG-6; Advanced Medical Technology Inc). When they walked and jogged, gait speed was controlled via a metronome. Three sets of valid trial data were collected for each task. Criteria for a valid trial consisted of a heel-strike pattern, chosen speed ranges, and the stance phase in the force plate embedded in the walkway. All participants wore the same type of running shoe (model Falcon Elite 2; Adidas AG). Retroreflective marker trajectory data were recorded at a sampling rate of 200 Hz using an 8-camera motion-analysis system (model MX-FX20 and Vicon Nexus System, Oxford Metric Ltd). Force-plate data were collected at a sampling rate of 1000 Hz.

After data collection, C3D files for all participants were imported into Visual3D. Spatial trajectory data were then smoothed at 13 Hz for walking and 23 Hz for jogging using a fourth-order, low-pass Butterworth filter. Cutoff frequencies for the low-pass filter were computed using power spectral density to determine the optimal cutoff frequency.²¹  Joint angles were calculated for ankles, knees, and hip joints in the sagittal, frontal, and transverse planes using a Cardan angle equation of flexion-extension, abduction-adduction, and internal rotation-external rotation for the 3 joint centers.²²  Force-plate data were normalized to the body weight of each participant and used to determine net internal joint moments for the 3 joints in the 3 planes. The net internal joint moment was computed from the synchronized joint angle, anthropometric data, and normalized ground reaction force data according to the inverse-dynamics method.²³  Smoothed kinematic and kinetic data were normalized for the full gait cycle from 0% to 100%. Normalized kinematic data represented the stride cycle from initial contact (0%) to terminal swing (100%). Phases during walking were defined as follows: early stance (0%–12%), midstance (13%–31%), late stance or push-off (32%–62%), early swing (63%–75%), midswing (76%–87%), and terminal swing (88%–100%).²⁴  Phases during jogging were early stance (0%–15%), midstance (16%–30%), late stance or push-off (31%–38%), early swing (39%–70%), and terminal swing (71%–100%).²⁴ Initial contact, which indicated the start of the gait cycle, was identified when the vertical ground reaction force was >20 N, whereas toe-off, which reflected the end of the stance phase, occurred when the vertical ground reaction force was ≤20 N.¹³  Although each participant may have exhibited a different toe-off point, we calculated a mean toe-off point for each group. Normalized data were then extracted, and the mean and 95% CI values were determined for each group using Excel (version 2016; Microsoft Corp). Ensemble curves on joint angles and moments were plotted using Gnuplot (version 5.2; http://www.gnuplot.info/).

One-way between-subjects analysis of variance with the Tukey post hoc test and χ² analysis were conducted to compare the characteristics of the 3 groups. Ensemble curve analyses were performed to compare the joint angles and moments of the 3 groups, using the means and 95% CIs.^15,16,20  Any CI bands that did not cross indicated differences between groups. The α level for all analyses was set at .05.^15,16,20  We used SPSS (version 25.0; IBM Corp) for statistical analysis.

Based on the analysis of variance and χ² analysis, we identified no differences among groups for sex, age, height, or mass but noted differences for LAS history and AII, FAAM-ADL, and FAAM-Sport scores (Table). Lower extremity kinematics during walking and jogging are shown in Figure 1. While walking, the CAI group demonstrated reduced dorsiflexion angles compared with the LAS coper (0%–3%, 30%–46%) and control (0%–5%, 28%–45%, 67%–68%; Figure 1A) groups. The CAI group displayed greater inversion angles than the LAS coper (0%–10%, 18%–25%, 44%–46%) and control (6%–9%, 15%–23%, 30%–41%, 48%–50%, 82%–89%; Figure 1B) groups. We observed no group differences in knee- or hip-joint kinematics during walking. While jogging, the CAI group exhibited dorsiflexion deficits compared with the LAS coper (0%–4%, 15%–27%, 100%) and control (0%–5%, 18%–27%, 45%–53%, 88%–93%, 99%–100%; Figure 1J) groups. In addition, the CAI group exhibited greater inversion angles than the LAS coper (0%–2%, 28%–53%, 76%, 91%–95%, 99%–100%) and control (1%–3%, 9%–11%, 16%–18%; Figure 1K) groups. We found no group differences in knee- or hip-joint kinematics during jogging. No differences in lower extremity joint moment were present among the 3 groups during walking (Figure 2). While jogging, the LAS coper group showed increased knee internal-rotation moments compared with the CAI group (6%–22%; Figure 2O). No differences in the ankle- and hip-joint moments were evident during jogging.

Our findings are consistent with those reported in several previous studies of sagittal-^15,25  and frontal-^{13,16,20,25,26}  plane kinematics, indicating dorsiflexion deficits and more inverted ankles in the CAI group compared with the LAS coper and control groups. In addition, the LAS coper group showed greater knee internal-rotation moments than the CAI group during jogging.

The CAI group displayed dorsiflexion deficits during the early stance, midstance, late stance, and early swing of walking (Figure 1A and B), as well as during the early stance, midstance, early swing, and terminal swing of jogging (Figure 1J and K). These results coincide with those reported by earlier researchers.^15,25  Although we did not measure static dorsiflexion range of motion (ROM), previous authors^27,28  have described relationships between limited dorsiflexion ROM and dorsiflexion deficits during dynamic tasks. Therefore, we recommend using joint mobilization with posterior talar glide²⁹  to increase static dorsiflexion ROM and resolve dorsiflexion deficits during gait. Also, the CAI group also exhibited more inverted ankles during the early stance, midstance, late stance, early swing, midswing, and terminal swing of walking (Figure 1A and B), as well as the early stance, midstance, late stance, early swing, and terminal swing of jogging (Figure 1J and K). Our findings of more inverted ankles during walking and jogging are consistent with those in the literature.^{13,16,20,25,26}  Inversion may predispose the ankles to sprain via foot-to-ground contact with the lateral border of the foot during early stance and excessive supination torques generated via ground reaction forces.³⁰  Computer modeling also showed an increased LAS risk due to an inverted ankle at initial contact during a side-shuffle movement.³⁰  Furthermore, the CAI group tended to walk and jog with dorsiflexion deficits, which may increase the vulnerability to inversion sprains because of the open-packed position of the ankle joint. Thus, inverted ankle patterns during gait should be corrected via gait training.

Initial contact is crucial for starting the gait cycle in a safe and stable manner. In the ankle joint, the closed-packed position consists of pronation, especially dorsiflexion. Hence, controlled ankle-movement patterns such as greater dorsiflexion and less inversion may allow individuals to achieve stable initial contact during gait. Musculoskeletal injuries can cause neuromuscular deficits by reorganizing the cortical and spinal levels of the central nervous system.³¹  In addition, many researchers have reported alterations of neuromuscular control in individuals with CAI via reflex latency and lower muscle-activation amplitude after sudden perturbation,^32,33  preactivation before initial contact,³⁴  and increased electromyography amplitude of peroneal muscles after initial contact.¹⁴  Neuromuscular control may be separated into open-loop control, which is feed forward or anticipatory, and closed-loop control, indicating feedback or reflex.³⁵  Given that patients with CAI have disrupted closed-loop control, they may need to use open-loop control, such as preactivation of the peroneal muscles, to prepare stable initial contact with less inversion. The faster the gait speed, the shorter the stride time and the faster the angular velocity of the lower extremities,^36,37  which may indicate that individuals have a shorter time to prepare stable initial contact during jogging compared with walking. Our results showed no differences in joint angles between the CAI group and the LAS coper and control groups during the terminal swing of walking but differences were seen in sagittal- and frontal-plane ankle kinematics between the CAI group and the LAS coper and control groups during the same period of jogging. Based on our results, different results for walking and jogging may be due to insufficient time to prepare for initial contact by individuals with CAI, who have altered neuromuscular control. Although the LAS coper and control groups prepared initial contact using greater dorsiflexion and less inversion during the terminal swing of jogging, the CAI group did not prepare initial contact with controlled ankle-movement patterns. Therefore, individuals with CAI may use a preactivation strategy of the peroneal muscles instead of ankle-movement controls, but future studies are required to evaluate this hypothesis, as it is just a speculation.

For the net internal joint moment, the LAS coper group had a tendency toward an increased knee internal-rotation moment compared with the control group, although the groups did not differ. Moreover, the LAS coper group generated a greater knee internal-rotation moment than the CAI group during jogging (Figure 2O). Intervals with differences in joint moments occurred from early stance to midstance during jogging. The knee joint was internally and externally rotated during midstance and terminal stance. In a similar way to the oblique axis of the subtalar joint, tibial internal rotation is associated with eversion of the rearfoot, and tibial external rotation is related to inversion of the rearfoot.³⁸  Thus, a greater knee internal-rotation moment may act as a facilitator for shock absorption during midstance and possibly as a resistor to knee external rotation with inversion during terminal stance. Although the ankle-joint kinematics differed between the CAI and LAS coper groups, the ankle kinetics did not. These findings may imply that the LAS coper group used proximal joint strategies but not ankle strategies, compensating for incomplete ankle functions to protect the ankle from inversion sprains. These results contradict those of earlier studies.^12,13  Doherty et al¹²  reported decreased knee-flexor moment in participants with CAI from 200 to 180 milliseconds before toe-off. Koldenhoven et al¹³  observed that patients with CAI exhibited increased ankle plantar-flexor moment during late stance to toe-off. Although previous researchers^12,13  identified altered sagittal-plane kinetics in the CAI group, our LAS coper group showed the alteration of knee moment in the in the transverse plane. This may have been caused by differences in experimental methods. All participants in this study walked and jogged on a walkway in a shod condition, whereas those in earlier investigations walked on a treadmill or barefoot. Franklin et al³⁹  systematically reviewed biomechanical differences between barefoot and shod conditions during walking and demonstrated differences in the joint angles and moments of the lower extremities and ground reaction forces between the conditions during walking. In addition, previous authors⁴⁰  identified biomechanical differences between walking overground and on a treadmill.

Given that retroreflective markers were attached to shoes for this study, future researchers should consider using specifically manufactured shoes for more accurate analyses of rearfoot motions. Also, because our biomechanical data were collected at selected speeds but not preferred speeds and participants were instructed to perform a heel-strike pattern, their gait biomechanics might have been affected by the gait speed and foot-strike pattern at initial contact. Although the exclusion criteria for all participants included peripheral neuropathy as well as a history of surgery, fracture, or acute injuries to the lower extremities, we did not exclude a history of concussion, other neurologic conditions, or injuries to the trunk. In future studies, investigators may need to include those conditions in the exclusion criteria, as they may affect the results. We also suggest that electromyographic analyses of the muscles controlling the knee joint may improve the understanding of the LAS copers' proximal joint strategy. Prospective studies may also be required to describe a more accurate mechanism for developing CAI and to characterize the effects of preventive intervention programs. Despite these limitations, our findings of kinematic differences between the CAI group and the LAS coper and control groups and the proximal joint strategy of the LAS coper group are strong points of our work.

During gait, the CAI group exhibited dorsiflexion deficits and more inverted ankles compared with the LAS coper and control groups. During jogging, the LAS coper group used a proximal joint strategy by generating a greater knee internal-rotation moment compared with the CAI group. Our findings provide insights for clinicians and athletic trainers in developing intervention programs. As a result, we advise clinicians and athletic trainers to include a proximal joint strategy and controlled ankle-movement patterns, such as greater dorsiflexion and less inversion, in rehabilitation programs for individuals with CAI.

We thank the International Olympic Committee Research Centre Korea and the Institute of Convergence Science for their assistance in conducting this study.