Application of High-speed Dual fluoroscopy to study in vivo Tibiotalar and Subtalar Kinematics in Chronic Ankle Instability Patients and Asymptomatic Controls During Dynamic Activities

Abstract

Background

Abnormal angular and translational (i.e. kinematic) motion at the tibiotalar and subtalar joints is believed to cause osteoarthritis (OA) in chronic ankle instability (CAI) patients.

Methods

This preliminary study quantified and compared in vivo tibiotalar and subtalar kinematics in four CAI patients (three female) and ten control subjects (five male) using dual fluoroscopy during a balanced, single-leg heel-rise and treadmill walking at 0.5 and 1.0 m/s.

Results

During balanced heel-rise, 69%, 54% and 66% of mean CAI tibiotalar internal/external rotation (IR/ER), subtalar inversion/eversion (In/Ev), and subtalar IR/ER angles, respectively, were outside the 95% confidence intervals (CIs) of controls. During 0.5 m/s gait, 50% and 60% of mean CAI tibiotalar dorsi/plantarflexion (D/P) and subtalar IR/ER angles, respectively, were outside the 95% CIs of controls. During 1.0 m/s gait, 62%, 65%, and 73% of mean CAI subtalar D/P, In/Ev, and IR/ER, respectively, were outside the 95% CIs of controls. CAI patients exhibited less tibiotalar and subtalar translational motion during gait; no clear differences in translations were noted during balanced heel-rise.

Conclusion

Overall, we found the balanced heel-rise activity to expose more tibiotalar and subtalar kinematic variation between CAI patients and controls. Therefore, weightbearing activities involving large range of motion, balance, and stability may be best for studying kinematic adaptations in CAI patients.

Clinical Relevance

Our preliminary results suggest that CAI patients require more tibiotalar external rotation, subtalar eversion, and subtalar external rotation during weightbearing stability exercises, all with less overall joint translation.

INTRODUCTION

Ankle sprains affect an estimated 32,000 Americans each day and are one of the most common injuries during athletic and recreational activities.^9,19,23,38 Up to 40% of all acute ankle sprains progress to chronic ankle instability (CAI),^12,26 which involves persistent feelings of instability, ankle pain, and subsequent ankle sprains as well as difficulty walking on inclined or uneven surfaces.⁸ CAI is clinically hypothesized to initiate ankle osteoarthritis (OA) by causing abnormal kinematics (i.e. angles and translations) at the tibiotalar and subtalar joints, leading to premature wear of articular cartilage.^{15,16,30,32,33} However, measurements of in vivo motion of the tibiotalar and subtalar joints are not available in CAI patients. These data could clarify the mechanical characteristics of this condition and provide detailed arthrokinematics (i.e. motion relative to the patient’s underlying bony anatomy) to refine current treatment strategies.

Traditional techniques used to quantify joint kinematics track the position of reflective markers adhered to the skin at bony landmarks. Despite widespread use, skin-marker motion capture is limited by errors associated with joint center estimations,¹¹ marker placement,¹⁴ and soft tissue artifact^2,10 and cannot distinguish the independent roles of the tibiotalar and subtalar joints as there are no reliable palpable landmarks for the placement of a skin marker about the talus. As a result, studies using traditional skin marker motion capture represent the ankle as a single joint and measure articulation of the shank relative to the heel.

Dual fluoroscopy (DF) is an imaging modality that allows for 3-D measurement of bone motion, thus providing calculations of angular and translational motion of multiple joints independent of one another. Using DF, prior studies have provided insight as to the functional roles of the tibiotalar and subtalar joints.^22,25 These studies have focused on tibiotalar and subtalar kinematics in healthy adults during overground gait, specifically the portion of the stance phase between heelstrike and heel-off. To our knowledge, the use of DF to study the kinematics of CAI patients has been further limited to measuring the alignment of the tibiotalar joint at pre-defined positions in a quasi-static manner.^4,7,35 The combination of tibiotalar and subtalar joint kinematics has not been established during any type of dynamic activity in patients with CAI. Thus, herein, we evaluated the feasibility of using DF to quantify ankle kinematics during dynamic loading in CAI patients and asymptomatic controls.

MATERIALS AND METHODS

Study Participants

After obtaining Institutional Review Board approval and informed consent, ten healthy volunteers were screened for gross ankle abnormalities and any history of back or lower limb surgery or pain (gender: 5M/5F; age: 30.9 ± 7.2 yo; BMI: 23.6 ± 3.4 kg/m²). In all subjects, standardized radiographic assessments were performed including weightbearing anteroposterior and lateral views of the foot, mortise view of the ankle, and hindfoot alignment view.²⁹ All radiographs were reviewed by an experienced, and fellowship-trained, foot and ankle surgeon (A.B.). Degenerative changes of the tibiotalar and subtalar joint were defined using the Kellgren-Lawrence scale.^18,21 Control subjects were screened for gross abnormalities and significant hindfoot OA (Kellgren-Lawrence score greater than 1). Based on these criteria, no control subjects were excluded. Additionally, patients with ankle pain, feelings of instability and symptoms that limited exercise and activities of daily living were enrolled from the co-author’s (C.L.S.) clinic. Standard radiographic assessments were performed for the affected ankle of each recruited patient.

Patient 1 (CAI-01), a 32 yo male (BMI: 25.9 kg/m²), presented with painful left lateral ankle instability with a positive anterior drawer test.³⁴ Conventional weightbearing radiographs and magnetic resonance imaging (MRI) confirmed the diagnosis. Conventional radiographs demonstrated neutral hindfoot alignment and mild tibiotalar OA with small osteophytes on the talar and tibial side. MRI demonstrated complete disruption of the anterior talofibular ligament (ATFL) and calcaneofibular ligament (CFL), splitting of the peroneus brevis, presence of an accessory type 1 navicular, an osteochondral lesion in the posterolateral talar dome, and a loose body in the lateral ankle gutter.

Patient 2 (CAI-02), a 27 yo female (BMI: 23.9 kg/m²), presented with painful right lateral ankle instability with a positive anterior drawer test. Conventional weightbearing radiographs demonstrated neutral hindfoot alignment and no hindfoot OA. This patient elected not to undergo surgery; therefore, we did not obtain advanced imaging on this patient.

Patient 3 (CAI-03), a 36 yo female (BMI: 30.4 kg/m²), presented with painful right ankle instability with a positive anterior drawer test and substantial tenderness over the ATFL and CFL. Conventional weightbearing radiographs demonstrated neutral hindfoot alignment and no hindfoot OA. MRI demonstrated complete disruption of the ATFL, thickened and elongated CFL, partial lesion of the deltoid ligaments, and a small osteochondral lesion of the lateral talar dome.

Patient 4 (CAI-04), a 28 yo female (BMI: 22.8 kg/m²), presented with painful right ankle instability with a positive anterior drawer test and substantial tenderness over the lateral and anterolateral ankle. Conventional weightbearing radiographs demonstrated neutral hindfoot alignment and no hindfoot OA. MRI demonstrated chronic injury of the ATFL and CFL and anterolateral tibiotalar impingement due to the Bassett ligament.^1,31

Dual Fluoroscopy and Skin-Marker Motion Capture

A custom high-speed DF system validated to a mean rotational and translational bias of 0.25 ± 0.81 degrees and 0.03 ± 0.35 mm, respectively, was used to measure tibiotalar and subtalar kinematics.³⁶ A ten-camera near-infrared motion analysis system (Vicon Motion Systems, Oxford, UK) was temporally and spatially synced with the DF system.^27,36 Reflective skin markers were applied to each subject prior to data capture per a modified Helen-Hayes configuration.²⁰ These reflective markers were used to track the position of the pelvis and bilateral thighs, shank and foot segments in 3-D.²⁰

All subjects completed three activities: a single-leg balanced heel-rise, treadmill walking at 0.5 m/s, and treadmill walking at 1.0 m/s. Two trials were captured of each activity. All activities were performed barefoot. Subjects practiced each activity prior to data acquisition. The balanced heel-rise activity was selected as it likely requires coordination, balance and stability. Additionally, similar heel-rise activities are used as clinical diagnosis tools.²⁴ During balanced heel-rise, subjects were instructed to perform the activity in a comfortable position at their desired speed to promote natural movement. DF images and tracking of skin-marker trajectories were acquired simultaneously throughout the entire heel-rise activity. Walking was chosen as it is a frequent activity of daily living. During treadmill walking, each subject was allowed to ambulate for at least 30 seconds prior to DF acquisition. Skin-marker trajectories were recorded starting several strides prior to and ending at least one stride after DF acquisition. For walking, subjects were not informed when data acquisition would begin.

The DF emitter beam energy settings were subject-specific and determined prior to the dynamic imaging of each subject to optimize the quality of the fluoroscopy images. Beam energy settings ranged from 62 to 78 kVp and from 1.2 to 2.2 mA·s, and depended on the size of the subject’s bones and orientation of their foot within the DF field of view. Dual fluoroscopy images and skin marker motion data were acquired at 300 Hz as described previously²⁷ The balanced heel-rise activity was captured in its entirety. The treadmill moved the foot out of the DF field of view prior to the completion of the gait cycle.²⁷ Thus, heelstrike and toe-off were imaged as separate trials and midstance was not captured in its entirety. The combination of the heelstrike and toe-off portions of the gait cycle that were imaged was referred to as ‘captured stance’. The fluoroscopy time of each subject was limited to 60 seconds.

Computed Tomography and Model-based Markerless Tracking

A computed tomography (CT) scan of each control and CAI subject was acquired (SOMATOM Definition AS, Siemens Medical Solutions, Malvern, PA) from mid-tibia through toe-tips at 1.0 mm slice thickness, 355 ± 59.2 mm square field of view, 512 X 512 acquisition matrix, 80 or 100 kVp, 20 to 93 mA·s. The use of CT image segmentation and model-based markerless tracking required several steps (Figure 1). Briefly, the tibia, talus, and calcaneus of each subject were semi-automatically segmented from the respective CT images (Amira 5.5, Visage Imaging, San Diego, CA). Ray-traced projection through the CT volumes of these segmentations was used to generate digitally reconstructed radiographs of each bone. Model-based markerless tracking³ was used to semi-automatically align the digitally reconstructed radiographs of each individual bone with the DF images from each time point. The combined DF and CT radiation exposure did not exceed 0.11 mSv, which was equivalent to 10 days of natural background radiation. This radiation exposure estimation was calculated based on radiation dosages from dosimeters implanted in phantoms²⁸; the value of 0.11 mSv was anticipated to represent the maximum possible dose to the subject.

Figure 1.

Flowchart of the methodological approach. A computed tomography (CT) scan was obtained of the subject’s ankle and foot and segmented to create three-dimensional (3-D) reconstructions of the tibia, talus and calcaneus bones. A digitally reconstructed radiograph (DRR) was created from each segmented bone. Anatomical coordinate systems were defined for the tibia, talus and calcaneus based on landmarks visible on the 3-D surfaces. Separately, dual fluoroscopy images were acquired. The dual fluoroscopy images and digitally reconstructed radiographs were used by model-based markerless tracking software to quantify the position and orientation of each bone. Bone positions and orientations were used to calculate angles and translations for the tibiotalar and subtalar joints.

Data Analysis

Joint angles and translations during each activity were calculated as detailed previously.²⁷ Briefly, the CT segmentations of each bone were used to create 3D reconstructions of the tibia, talus, and calcaneus. Landmarks were identified on the 3D reconstructions of each bone and used to define subject-specific anatomical coordinate systems for the tibia, talus, and calcaneus. A weightbearing neutral position was determined by using DF to align the tibia, talus, and calcaneus to their respective positions during midstance. The talus and calcaneus coordinate systems were then aligned with the tibia coordinate system, but maintained their respective joint center locations.

Skin-marker data were used to determine gait events such as heelstrike and toe-off. Specifically, heelstrike was defined as the frame corresponding to the minimum height of the heel marker following a downward trajectory. Toe-off was defined as the frame corresponding to the minimum height of the toe marker prior to an upward trajectory. The duration of stance phase was determined as the time between heelstrike and toe-off and used to normalize each trial. All gait trials were aligned at either heelstrike (0% of normalized stance) or toe-off (100% of normalized stance). The balanced heel-rise activity was normalized and aligned across subjects using inflection points from the D/P angles calculated between the calcaneus in relation to the tibia.²⁷

The anatomical coordinate system of each bone was applied to the bone orientations and locations determined via markerless tracking and used to calculate dynamic tibiotalar and subtalar joint angles.³⁶ A fourth-order bi-directional low-pass Butterworth filter was applied to the dynamic joint angles and translations. A cutoff frequency of 10 Hz was selected using the residual analysis method of Winter.³⁷ Joint angles were reported as dorsi/plantarflexion (D/P), inversion/eversion (In/Ev) and internal/external rotation (IR/ER). Dorsiflexion, eversion and external rotation were considered positive. Joint translations were reported in the medial-lateral (ML), anterior-posterior (AP) and superior-inferior directions.²⁷

The mean and 95% confidence intervals (CIs) of joint angles and translations for the two trials of each activity were calculated across the ten control subjects.²⁷ The two trials of each CAI patient were averaged for each activity. For each activity, the mean trial of each CAI patient was plotted against the mean and 95% CIs of the control subjects for qualitative comparison. To facilitate comparison with the controls, the portion of each mean CAI trial that fell outside the 95% CIs was calculated and expressed as a percentage of the entire trial. For treadmill walking at 0.5 and 1.0 m/s, we evaluated the percentage of CAI joint angles that fell outside the 95% CIs of the controls during captured stance. We then calculated this analysis for the individual heelstrike and toe-off portions of gait. In doing so, we were able to investigate whether differences between CAI patients and controls were more evident during ankle loading or unloading. We considered CAI joint angles to be different compared to controls and reported these differences in the text, if a majority (50% or more) of CAI joint angles during balanced heel-rise, captured stance, or the heelstrike or toe-off portions of gait fell outside the 95% CIs of the controls. Finally, using a paired t test, we compared the percentage of CAI joint angles that fell outside the 95% CIs of the controls during captured stance to determine if there were statistically significant differences by joint, gait speed, or portion of gait (heelstrike or toe-off).

Tibiotalar and subtalar range of motion (ROM) was calculated for control subjects and CAI patients for balanced heelrise and 0.5 and 1.0 m/s captured stance as described previously.²⁷ For each joint and activity, the rotational and translational ROM of each CAI patient was plotted relative to the mean and 95% CIs of the controls. The ROM for a CAI patient was considered different if outside the 95% CIs of the controls.

RESULTS

Balanced Heel-rise

The mean (± standard deviation) time to complete the balanced heel-rise activity was 0.87 ± 0.20 seconds and 1.16 ± 0.19 seconds for controls and CAI patients, respectively. During balanced heel-rise, the tibiotalar (Figure 2) and subtalar (Figure 3) joint angles of the CAI patients were different than the control subjects and frequently exhibited opposing trends. Tibiotalar and subtalar IR/ER angles and subtalar In/Ev angles of the CAI patients differed from the control subjects. Here, the mean percentage of CAI joint angles that fell outside the 95% CIs of the control subjects during balanced heel-rise was 69%, 54% and 66%, for tibiotalar IR/ER, subtalar In/Ev and subtalar IR/ER, respectively (Table 1). There were no statistically significant differences between the tibiotalar and subtalar joint angle percentages outside the 95% CIs of the controls during heel-rise.

Figure 2.

Tibiotalar dorsi (+)/plantarflexion (top), inversion/eversion (+) (middle), and internal/external (+) rotation (bottom) mean joint angles of patients with chronic ankle instability (CAI) compared to asymptomatic control subjects during a single-leg balanced heel-rise activity. Data are plotted per normalized balanced heel-rise. The joint angles of the asymptomatic control subjects are presented as the mean (white line) ± 95% confidence intervals (gray). Each colored line represents the mean joint angles of a different CAI patient.

Figure 3.

Subtalar dorsi (+)/plantarflexion (top), inversion/eversion (+) (middle), and internal/external (+) rotation (bottom) mean joint angles of patients with chronic ankle instability (CAI) compared to asymptomatic control subjects during a single-leg balanced heel-rise activity. Data are plotted per normalized balanced heel-rise. The joint angles of the asymptomatic control subjects are presented as the mean (white line) ± 95% confidence intervals (gray). Each colored line represents the mean joint angles of a different CAI patient.

Table 1.

Individual and mean percentages of the balanced heel-rise activity for which the tibiotalar and subtalar joint angles of the chronic ankle instability (CAI) patients fell outside the 95% confidence intervals of the control subjects.

	Tibiotalar Joint			Subtalar Joint
	D/P	In/Ev	IR/ER	D/P	In/Ev	IR/ER
CAI01	0.00	0.00	0.55	0.00	0.48	0.43
CAI02	1.00	0.92	0.67	0.33	0.48	0.65
CAI03	0.37	0.70	0.74	0.73	0.71	1.00
CAI04	0.59	0.18	0.78	0.21	0.47	0.56
Mean	0.49	0.45	0.69	0.32	0.54	0.66
St dev	0.42	0.43	0.10	0.31	0.12	0.24

Two trials for each CAI patient have been averaged. St dev = standard deviation; D/P = dorsi/plantarflexion; In/Ev = inversion/eversion; IR/ER = internal/external rotation. Percentages have been expressed as a ratio.

Treadmill Walking

Although CAI tibiotalar joint angles often exhibited trends that were similar to the controls during captured stance CAI joint angles were not always within the 95% CIs of the controls (Figure 4). During captured stance, a majority of CAI tibiotalar D/P joint angles fell outside the 95% CIs of the controls at the 0.5 m/s walking speed (50%). Differences between the CAI patients and controls were more notable when the heelstrike and toe-off portions of gait were compared separately. For the heelstrike portion of 1.0 m/s gait, a majority of CAI tibiotalar joint angles were outside the 95% CIs for In/Ev and IR/ER (63% and 58%, respectively) (Table 2). For the toe-off portion of 0.5 and 1.0 m/s gait, a majority of CAI tibiotalar joint angles were outside the 95% CIs for D/P (73% and 50%, respectively) (Table 2).

Figure 4.

Tibiotalar dorsi (+)/plantarflexion (top), inversion/eversion (+) (middle), and internal/external (+) rotation (bottom) mean joint angles of patients with chronic ankle instability (CAI) compared to asymptomatic control subjects during 0.5 m/s (left) and 1.0 m/s (right) gait. Data are plotted per normalized stance with all subjects aligned at heelstrike (0%) and toe-off (100%). The heelstrike and toe-off portions of stance were collected as separate trials, since the movement of the treadmill caused the foot to move out of the combined field-of-view of the fluoroscopes prior to the completion of the stance phase of gait. The joint angles of the asymptomatic control subjects are presented as the mean (white line) ± 95% confidence intervals (gray). Each colored line represents the mean joint angles of a different CAI patient.

Table 2.

Individual and mean percentages of the heelstrike (HS) and toe-off (TO) portions of captured stance (at 0.5 m/s and 1.0 m/s) for which the tibiotalar joint angles of the chronic ankle instability (CAI) patients fell outside the 95% confidence intervals of the control subjects.

	Dorsi/plantarflexion				Inversion/Eversion				Internal/External Rotation
	0.5 m/s		1.0 m/s		0.5 m/s		1.0 m/s		0.5 m/s		1.0 m/s
	HS	TO	HS	TO	HS	TO	HS	TO	HS	TO	HS	TO
CAI-01	0.00	1.00	0.24	1.00	0.00	0.00	0.00	0.31	0.50	0.00	0.59	0.31
CAI-02	1.00	0.91	0.85	0.00	0.81	0.47	0.89	1.00	0.00	0.05	0.00	0.00
CAI-03	0.00	0.00	0.00	0.00	0.56	0.15	0.95	0.60	0.71	0.38	1.00	0.69
CAI-04	0.00	1.00	0.61	1.00	0.09	0.00	0.66	0.00	0.02	0.07	0.74	0.00
Mean	0.25	0.73	0.43	0.50	0.37	0.16	0.63	0.48	0.31	0.13	0.58	0.25
St Dev	0.50	0.49	0.38	0.58	0.39	0.22	0.44	0.43	0.35	0.17	0.42	0.33

Two trials for each CAI patient have been averaged. St dev = standard deviation; D/P = dorsi/plantarflexion; In/Ev = inversion/eversion; IR/ER = internal/external rotation. Percentages have been expressed as a ratio.

Subtalar joint angles of the CAI patients were often different compared to controls during captured stance (Figure 5). The most notable differences were observed in subtalar IR/ER during 0.5 m/s and 1.0 m/s captured stance. During captured stance, a majority of CAI subtalar joint angles fell outside the 95% CIs of the control subjects for IR/ER at the 0.5 m/s walking speed (60%) and for D/P, In/Ev and IR/ER at the 1.0 m/s speed (62%, 65%, and 73%, respectively). Differences in subtalar joint angles between CAI patients and controls were most evident when the heelstrike and toe-off portions of gait were independently compared, particularly during toe-off. For the heelstrike portion of gait, a majority of CAI subtalar joint angles fell outside the 95% CIs of the controls for D/P, In/Ev, and IR/ER (77%, 81%, and 97%, respectively) at the 1.0 m/s speed (Table 3). During the toe-off portion of gait, a majority of CAI subtalar joint angles fell outside the 95% CIs of the controls for In/Ev and IR/ER at the 0.5 m/s speed (65% and 64%, respectively) and for D/P, In/Ev, and IR/ER at the 1.0 m/s speed (62%, 82%, and 91%, respectively) (Table 3).

Figure 5.

Subtalar dorsi (+)/plantarflexion (top), inversion/eversion (+) (middle), and internal/external (+) rotation (bottom) mean joint angles of patients with chronic ankle instability (CAI) compared to asymptomatic control subjects during 0.5 m/s (left) and 1.0 m/s (right) gait. Data are plotted per normalized stance with all subjects aligned at heelstrike (0%) and toe-off (100%). The heelstrike and toe-off portions of stance were collected as separate trials, since the movement of the treadmill caused the foot to move out of the combined field-of-view of the fluoroscopes prior to the completion of the stance phase of gait. The joint angles of the asymptomatic control subjects are presented as the mean (white line) ± 95% confidence intervals (gray). Each colored line represents mean joint angles of a different CAI patient.

Table 3.

Individual and mean percentages of the heelstrike (HS) and toe-off (TO) portions of captured stance (at 0.5 m/s and 1.0 m/s) for which the subtalar joint angles of the chronic ankle instability (CAI) patients fell outside the 95% confidence intervals of the control subjects.

	Dorsi/plantarflexion				Inversion/Eversion				Internal/External Rotation
	0.5 m/s		1.0 m/s		0.5 m/s		1.0 m/s		0.5 m/s		1.0 m/s
	HS	TO	HS	TO	HS	TO	HS	TO	HS	TO	HS	TO
CAI-01	0.19	0.70	0.32	0.90	0.96	1.00	0.49	1.00	1.00	1.00	1.00	1.00
CAI-02	0.87	0.06	0.76	0.69	0.85	0.00	1.00	0.94	0.96	0.06	0.96	0.74
CAI-03	0.00	0.91	1.00	0.87	0.00	1.00	0.90	1.00	0.00	0.72	1.00	0.89
CAI-04	0.46	0.00	1.00	0.00	0.08	0.60	0.86	0.35	0.00	0.79	0.90	1.00
Mean	0.38	0.42	0.77	0.62	0.47	0.65	0.81	0.82	0.49	0.64	0.97	0.91
St Dev	0.38	0.46	0.32	0.42	0.50	0.47	0.22	0.32	0.57	0.41	0.05	0.12

Two trials for each CAI patient have been averaged. St dev = standard deviation; D/P = dorsi/plantarflexion; In/Ev = inversion/eversion; IR/ER = internal/external rotation. Percentages have been expressed as a ratio.

There were no statistically significant differences between heelstrike and toe-off for the two gait speeds for the portion of tibiotalar (Table 2) or subtalar (Table 3) joint angles outside the 95% CIs of the controls.

Range of Motion

The CAI patients exhibited angular ROM similar to that of the control subjects during the captured stance of gait and balanced heel-rise, but there were some exceptions (Figure 6). Notably, CAI patients often exhibited less In/Ev ROM than the controls during captured stance and balanced heel-rise, although the individual ROM values were not always outside the 95% CIs of the controls. Two patients during 0.5 m/s gait and three patients during 1.0 m/s gait exhibited subtalar In/Ev ROM below the lower 95% CI of controls. Also of note, CAI-01 and CAI-03 exhibited subtalar angular ROM that was consistently less than the mean and often less than the lower 95% CI of the controls during captured stance. In addition, CAI-03 exhibited tibiotalar D/P ROM that was less than the lower 95% CI of the controls throughout all activities.

Figure 6.

Joint angle range of motion (ROM) values for chronic ankle instability (CAI) patients (colored symbols) plotted against the mean (black dots) ROM and 95% confidence interval (CI) (black bars) of asymptomatic controls for the tibiotalar (top) and subtalar (bottom) joints during balanced heel-rise (left), 0.5 m/s captured stance (middle), and 1.0 m/s captured stance (right). D/P = dorsi/plantarflexion; In/Ev = inversion/eversion; IR/ER = internal/external rotation.

Many CAI patients exhibited decreased translational ROM relative to controls, especially during the captured stance of gait (Figure 7). During 1.0 m/s gait, all CAI patients demonstrated tibiotalar translational ROM in each direction that was less than the mean values of the controls. In fact, a majority of the CAI patients exhibited tibiotalar translational ROM in each direction that was less than the lower 95% CI of the controls during 1.0 m/s captured stance. Furthermore, there were between one and three CAI patients that displayed tibiotalar translational ROM in the AP and SI directions that was less than the lower 95% CI of the controls during each gait speed. During 0.5 and 1.0 m/s gait, all CAI patients exhibited subtalar translational ROM in the AP direction that was less than the mean of the controls, with one and three CAI patients, respectively, falling below the lower 95% CI of the controls.

Figure 7.

Joint translation range of motion (ROM) values for chronic ankle instability (CAI) patients (colored symbols) plotted against the mean (black dots) ROM and 95% confidence interval (CI) (black bars) of asymptomatic controls for the tibiotalar (top) and subtalar (bottom) joints during balanced heel-rise (left), 0.5 m/s captured stance (middle), and 1.0 m/s captured stance (right). Direction of translational movement: ML = medial-lateral; AP = anterior-posterior; SI = superior-inferior.

DISCUSSION

To our knowledge, this was the first study to investigate differences in tibiotalar and subtalar kinematics between CAI patients and controls during dynamic activities. Overall, we found larger differences in kinematics between CAI patients and controls during the balanced heel-rise activity. The CAI heel-rise trials showed a distinctly different trend than the controls for tibiotalar IR/ER and subtalar In/Ev and IR/ER. During captured stance, CAI joint angles typically exhibited trends similar to the controls, but were often shifted higher or lower than the 95% CIs of the controls, which became more evident when the toe-off portion of stance was analyzed separately. Collectively, our results imply that weightbearing activities involving larger ROM, balance, and stability may be most effective for evaluating kinematic differences in CAI patients.

During balanced heel-rise and 0.5 m/s captured stance, CAI-02 demonstrated values for tibiotalar plantarflexion that could have been considered outliers. CAI-02 was the only patient that did not undergo surgery, and thus, instability in CAI-02 may have been less severe than the other patients. This may have enabled CAI-02 to achieve greater peak plantarflexion and D/P ROM. During balanced heel-rise, CAI-02 exhibited tibiotalar angular ROM that was similar to or greater than that of the controls and subtalar angular ROM that was less than the lower 95% CI of the controls. This may indicate that the tibiotalar joint provided additional motion to compensate for the limited ROM at the subtalar joint.

CAI-03 demonstrated kinematics that differed from those of the other CAI patients, notably with subtalar In/Ev and IR/ER trends and ROM values during balanced heel-rise that were more akin to the control subjects than the other CAI patients. These trends exhibited by CAI-03 during balanced heel-rise were accompanied by translational ROM values that were on the lower end of the control values. CAI-03 was the only patient with torn deltoid ligaments, possibly causing medial ankle instability in addition to the lateral instability demonstrated by the other CAI patients. This may explain why the subtalar joint of CAI-03 became more inverted and internally rotated during balanced heel-rise, a motion similar to the controls, yet with greater In/Ev and IR/ER values. The increased tibiotalar In/Ev ROM of CAI-03 may indicate possible changes to the articular surface, since the articular surface is primarily responsible for In/Ev stability when the ankle is loaded.³¹ Furthermore, the deltoid is the main contributor to internal rotation restraint, especially when loaded and plantarflexed, as the foot would be during a balanced heel-rise.³¹ With a damaged deltoid ligament, it is possible that CAI-03 had less control over internal rotation during loaded plantarflexion, thus causing increased internal rotation throughout balanced heel-rise.

CAI-01 and CAI-04 demonstrated relatively similar joint kinematics throughout the balanced heel-rise and captured stance, potentially as a result to their injury patterns to the ATFL and CFL. During balanced heel-rise, CAI-01 and CAI-04, along with CAI-02, exhibited tibiotalar IR/ER, subtalar In/Ev, and subtalar IR/ER that differed from those of the controls. During captured stance, CAI-01 and CAI-04 often exhibited joint angles that were within the 95% CIs of the controls. However, both CAI-01 and CAI-04 demonstrated tibiotalar D/P during toe-off and subtalar IR/ER angles during captured stance that were greater than the upper 95% CI limit of the controls. These differences may provide insight into the effects of CAI due to ATFL and CFL injuries on tibiotalar and subtalar kinematics during weightbearing activities requiring plantarflexion.

The joint angles of CAI patients differed most from the controls during motions requiring plantarflexion such as toe-off and heel-rise. The largest differences were found for subtalar In/Ev and IR/ER. These results may suggest that substantial differences between CAI patients and controls are only present in specific kinematic metrics and/or are activity dependent. The plantarflexion motion involved during toe-off and heel-rise is a position that is more susceptible to rolling of the ankle.^6,13 Given that CAI patients are more prone to rolling of the ankle (and subsequent sprains), kinematic differences may be most evident during activities that demand substantial weightbearing plantarflexion.

Perhaps surprisingly, translational ROM was often decreased for the CAI patients in our study. Previous studies found that CAI patients had significantly greater translation of the talus in the anterior direction compared to their uninjured ankle.^4,7,35 However, these studies evaluated motion of the talus during static weightbearing poses,^4,7,35 not overall ROM throughout an activity where active muscle control is highly engaged, and may not be directly comparable.

Although direct comparisons of results in our study and data in the literature cannot be made, some generalizations can be deduced. Prior studies determined tibiotalar kinematics in CAI patients quasi-statically at 25%, 50%, 75% and 100% of body weight, but did not evaluate dynamic gait.^4,7,35 Caputo et al. reported no significant differences in tibiotalar D/P or In/Ev between patients with unilateral ATFL or combined ATFL and CFL injuries and intact ankles,⁷ a finding that is supported by our balanced heel-rise and captured stance results. Both Caputo et al. and Wainright et al. found that the talus in an ankle with instability (injury to the ATFL or ATFL and CFL) had significantly more internal rotation than an uninjured ankle when loaded at specific percentages of body weight.^7,35 Data from Bischof et al. indirectly supported these findings, where it was determined that patients with lateral ankle instability (injury to the ATFL or ATFL and CFL) had significantly higher peak cartilage contact strains on the medial side of the tibiotalar joint.⁴ Conversely, our study demonstrated more tibiotalar external rotation in CAI patients than controls during balanced heel-rise. One explanation for this discrepancy may be that the CAI patients had more relaxed neutral positions with more internal rotation than the control subjects, causing the CAI patients to exhibit more external rotation. Additionally, three CAI patients in the current study had an injured CFL, which provides external rotation restraint,³¹ and may have contributed to external rotation values that were higher than controls.

There were limitations to this study that warrant discussion. First, gait was performed at relatively slow speeds which may have been unnatural for some subjects, especially controls. These speeds were chosen to ensure that all subjects could perform the same gait speed. Per our institutional review board approval, each subject was limited to 60 seconds of fluoroscopy time, which prevented us from capturing additional activities, such as a self-selected walking speed, and/or additional trials. We could have sought approval to increase the allowable radiation exposure. However, model-based tracking is a time-consuming endeavor. In addition, we wanted to minimize radiation exposure. For these reasons, we chose not to analyze multiple trials. Although treadmill walking allowed us to have a more consistent step cadence and stride length across subjects, it caused the foot to exit the combined field of view of the fluoroscopes prior to the completion of stance and required us to capture heelstrike and toe-off as separate trials. For this reason, we analyzed heelstrike and toe-off separately. Although our imaging technique prevented us from analyzing a majority of stance, we successfully investigated kinematic trends during peak loading (heelstrike to foot-flat) and unloading (late midstance to toe-of) of the ankle during stance. An additional limitation was our small sample size. With only four CAI patients, our results should be interpreted as case studies. Nevertheless, we believe these preliminary results provide conceptual proof that our DF approach can assess kinematics in a complex ankle patient population, which was the objective herein. A final limitation is that joint angles were considered different if 50% or more of a trial fell outside the 95% CIs of the controls; this criterion was not based on a clinically-meaningful difference, and thus, we advocate for caution when interpreting these results.

In conclusion, the results of this exploratory study demonstrate that DF is a viable modality to study the subject-specific kinematics of CAI during dynamic activities. However, coupling DF with a treadmill may not be ideal for evaluating gait, as it does not image all of stance. Thus, we recommend imaging over-ground gait. The balanced heel-rise helped to expose kinematic differences between CAI patients and controls, but differences were more subtle during walking. Therefore, demanding activities, such as stair-climbing, or high impact activities such as jumping/landing, should be examined in future kinematic studies of CAI patients.