Subchondral Bone Degeneration and Pathology 3-15 Years Following Ankle Sprain Injury in Adolescent Sport

Abstract

Background:

Sport-related ankle sprains (SASs) are prevalent in adolescents (ages 10-19), increasing the risk of developing posttraumatic osteoarthritis (PTOA). Although early ankle osteoarthritis (OA) is not well defined, OA eventually includes alterations in bone mineral density (BMD), structural changes, and soft tissue pathology. This study examined the impact of SAS sustained in adolescent sport on bone and soft tissue structural outcomes 3-15 years postinjury.

Methods:

Participants (n = 10) with prior unilateral SAS in adolescent sport (HxAI) were compared to age- and sex-matched controls. To assess injury-related pathologies and BMD, 1.5-tesla (T) extremity magnetic resonance imaging (MRI) and computed tomography scans were used. Semiquantitative scores for injury patterns and OA features from MRI scans were summed and compared between groups. The talus, calcaneus, navicular, and 5% distal tibia were segmented, and BMD was measured for each bone.

Results:

All HxAI participants exhibited MRI injury pathology (median 2; IQR 1-6), whereas only 1 of 10 controls showed pathology (median 0; IQR 0-0), χ²(1, n = 20) = 16.36, P < .001. Both the injured and uninjured ankles in HxAI displayed injury pattern pathology. Additionally, 3 of 10 injured ankles and 2 of 10 uninjured ankles in the HxAI group (median 0; IQR 0-3), but none of the controls (median 0; IQR 0-0), exhibited OA features. In the HxAI group, talus BMD was lower in the injured ankle (502.4 ± 67.9 g/cm³) compared with the uninjured ankle (515.6 ± 70.1 g/cm³) (F = 13.33, P = .002), with no significant BMD differences at the calcaneus, navicular, or 5% distal tibia. No differences were observed between the ankles of the control group.

Conclusion:

The presence of injury pattern pathology, structural changes, and reduced talus BMD suggest that degenerative changes may occur in individuals as early as 3-15 years following ankle injury.

Graphical Abstract.

This is a visual representation of the abstract.

Level of Evidence: Level III, retrospective comparative study.

Background

Osteoarthritis (OA) is the most common form of arthritis and typically manifests in the hip, knee, hand, or ankle, characterized by deterioration in the articular cartilage and changes in the subchondral bone. ² Despite extensive research on hip and knee OA,^9,13 the understanding of ankle OA remains comparatively limited. In contrast to hip and knee OA, where approximately 8% to 12% of cases are attributed to posttraumatic OA (PTOA) resulting from injury,^10,28 approximately 78% of ankle OA is classified as posttraumatic. Ankle PTOA is often associated with ankle sprains. ²⁸ Ligamentous injuries affecting joint structure and stability significantly increase the risk of developing PTOA, and considering the high rates of ankle injuries, sports-related ankle sprains (SAS) may significantly influence the incidence of ankle PTOA. ⁵

SASs are one of the most common musculoskeletal injuries, estimated to represent 40% to 50% of all sports-related injuries presenting at the emergency department.^6,30 Adolescent athletes are at high risk of ankle sprains, which can lead to long-term complications such as OA. ¹¹ In particular, lateral SASs are a common cause of ligamentous ankle PTOA, with persistent instability and varus malalignment being common features. ²⁸ Additionally, research has shown that individuals who have experienced an SAS during adolescence may have more pain or symptoms, reduced function, and a greater risk of developing OA following an injury.^22,23

PTOA is especially debilitating owing to its earlier onset and more rapid progression when compared to traditional OA, ³¹ highlighting the need to identify key etiology features for early identification. PTOA can be diagnosed through a combination of clinical evaluation and imaging techniques. Although there are clinical definitions for hip, knee, and hand OA, there are no formal guidelines for diagnosing ankle OA. ³ Clinical features typically include the presence of joint pain, gait disturbance, and/or restricted range of motion. Imaging techniques such as radiographs, computed tomography (CT), magnetic resonance imaging (MRI), and/or ultrasonography are often used to confirm ankle OA, by determining the presence of joint space narrowing, osteophytes, and cartilage degeneration.^3,21

Although CT can also be used to visualize articular cartilage, this 3-dimensional imaging modality is well suited to evaluating structural changes, subchondral bone changes, and osteophytes. ²¹ More specifically, bone remodeling is thought to play a significant role in the development of OA, involving a complex interplay between bone formation and resorption. The initial stages of OA are characterized by increased bone remodeling ultimately leading to bone loss followed by slower bone turnover, which can result in subchondral sclerosis and osteophyte formation.^7,12 Therefore, bone loss may be indicative of early PTOA development in the ankle, and further research is required to identify these key features following adolescent SAS.

The objective of this study was to examine the impact of SAS sustained in youth sport on structural outcomes associated with early-onset ankle PTOA 3-15 years postinjury. This period of time postinjury was selected because radiographic evidence of PTOA generally presents around 20 years postinjury. ¹⁰ Therefore, this 3- to 15-year time frame postinjury would allow for early identification of individuals considered “at risk” of progressing toward ankle PTOA. MRI was used to determine the presence of joint pathology associated with injuries and OA, whereas CT was used to examine changes in bone mineral density (BMD).

Materials and Methods

Study Design

This was a cross-sectional study that recruited participants (n = 10) from the previous Ankle Pre-OA cohort study conducted at the Sport Injury Prevention Research Centre at the University of Calgary, Canada,^22,23 who indicated their willingness to be contacted for future research and word of mouth (eg, newsletters, social media). All potential participants underwent an interview process, which included collecting details about the SAS. ²⁰ This included details regarding their SAS (eg, first SAS in the case of multiple), date/year of SAS injury, diagnosis, medical attention received, and time-loss duration. For inclusion, participants must have had a history of a significant unilateral SAS 3-15 years prior to enrolment in the study (HxAI group). SAS was defined in accordance with the International Ankle Consortium’s position statement, as a clinical diagnosis of ankle ligament injury, specifically including injuries involving the lateral ligament, ligaments of the tibiofibular syndesmosis (high ankle sprain) and/or medial ligament sprain that resulted in disruption of sports participation for >3 days and required medical consultation (eg, physician, physical therapist, or athletic trainer). HxAI participants must have reported no injury to the contralateral ankle. Exclusion criteria included pregnancy, nonsteroidal anti-inflammatory use, cortisone injection within the 3 months prior to testing, or a musculoskeletal injury within the 3 months prior to testing that resulted in time loss (work, school, or sport). Additionally, individuals were excluded if they had any significant lower limb injuries within the past 3 years or any chronic conditions, either ongoing or within the last 12 months, that could affect mobility and physical function. It is possible that the referent injury in youth sport may not have been the only ankle injury in youth but was the most significant as defined by inclusion criteria. The Godin Leisure Time Exercise questionnaire was used to assess the HxAI groups activity levels (>24 units = active, 14-23 units = moderately active, <14 units = insufficiently active/sedentary). The Foot and Ankle Outcome Score (FAOS) Foot and Ankle Survey was used to assess the HxAI participants’ opinions about their ankle-related problems. The FAOS survey consists of 5 subscales including pain, other symptoms, function in daily living, function in sport and recreation, and foot and ankle–related quality of life. Each item is scored from 0 to 4, and a normalized score (100 indicating no symptoms, and 0 indicating extreme symptoms) was calculated for each subscale.

Age- and sex-matched controls (Control group) with no history of ankle injury (n = 10) were also recruited through invitations made by study coinvestigators, collaborators, posters, and word of mouth via newsletters and advertisements on the University of Calgary Research Participation website. The Controls were age-matched within 2 years of their HxAI counterpart. All participants provided written informed consent before study participation. Ethical approval for this study was obtained from the Conjoint Health Research Ethics Board at the University of Calgary (REB16-2280).

MRI Acquisition

MRI scanning was conducted on the ankle using a 1.5-tesla (T) extremity GE Optima MR430S MRI system (GE Healthcare, Chicago, IL) representative image (Figure 1). MRI scans were performed on both ankles in the SAS group, whereas the control group underwent a single-ankle MRI scan on the ankle that corresponded with the SAS participants’ injured side. The following sequences were obtained for both groups (parameters provided in Supplemental Table S1): (1) T1-weighted fast spin echo (FSE) in the sagittal plane, (2) T1-weighted gradient echo in the sagittal plane, (3) proton density–weighted FSE in the coronal plane, (4) proton density–weighted FSE in the coronal plan with fat saturation, (5) proton density–weighted FSE in the axial plane, and (6) axial short tau inversion recovery (STIR) in the axial plane. Additionally, for the HxAI group, we obtained (7) STIR FSE in sagittal plane and (8) T1-weighted 3D FSE in the sagittal plane. Details of the MRI sequences are outlined in Appendix 1.

Figure 1.

Representative magnetic resonance image (MRI) from the history of ankle injury group (HxAI) of the injured and uninjured ankle. (A) Coronal proton density–weighted fat saturated image of the previously injured ankle of a participant in the HxAI group, showing a tibial bone marrow lesion and tibiotalar cartilage loss compared to the (B) coronal uninjured ankle. (C) Sagittal T1-weighted image (D) compared to the sagittal uninjured ankle.

MRI Analysis

A board-certified fellowship-trained musculoskeletal radiologist (J.L.J.) assessed each ankle MRI for the presence and severity of pathologies associated with acute ankle injuries, blinded to the injured ankle. Specifically, the presence and severity of ligament and tendon injury (score 0-3 per structure), osseous pathology (score 0-2), talar osteochondral lesions (score 0-5), and effusion in the tibiotalar and talocalcaneal joints (score 0-2) were evaluated using a semiquantitative scale for each feature.^13,24 MRI scoring is defined in Appendix 2. In addition, semiquantitative scoring was used to assess the presence and severity of pathology associated with OA. This included osteophytes (score 0-1) and cartilage loss (score 0-1) in the tibiotalar and subtalar joints, as well as severity of bone marrow lesions (BMLs) (score 0-2) in the tibia, talus, and fibula. Sum scores were computed based on (1) the sum of presence and severity of injury pattern pathology defined by the Roemer et al ²⁴ criteria alone (maximum score 97), (2) sum of OA features alone (maximum score 10), and (3) a combination of injury pattern pathology and OA features (maximum score 107). The median and interquartile range (IQR) and range of the sum scores are presented.

CT Image Acquisition

Participants underwent bilateral ankle helical CT scans with a GE Revolution GSI scanner (GE Electric Medical System, Milwaukee, WI). Scans were acquired in the supine position, with the ankles flexed at 90°, resting on a wedge from the distal tibia to the end of the feet. A 6-rod hydroxyapatite (HA) phantom (QRM-BDC/6–200, Quality Assurance in Radiology and Medicine, GmbH, Moehrendorf, Germany) was included in the scan field of view for density calibration. Scan parameters used were 120 kVp, 235 mAs, 50-cm field of view, 512 × 512 matrix, 0.625 mm reconstructed slice thickness, pitch 0.5, revolution time 0.5 seconds, and a bone plus reconstruction kernel with filtered back-projection.

CT Image Analysis

The bilateral calcaneus, talus, navicular, and tibia bones were manually marked in all 3 planes using ITK-SNAP (version 3.8.0) ³³ and segmented using a bone-enhanced graph cut technique ⁴ (Figure 2). The graph cut code is open source and publicly available (https://github.com/Bonelab/FemurSegmentation). Based on the participants’ standing height, anthropometric equations were used to estimate tibia length for both males and females. ²⁵

Figure 2.

Segmentation of the talus (yellow), calcaneus (green), navicular (magenta), and distal 5% of the tibia (red) in the sagittal view. [See online article for color figure.]

$$\mathrm{Male}\mathrm{tibia}\mathrm{length}\left(\mathrm{cm}\right)=\left(\frac{(\mathrm{stature}\left(\mathrm{cm}\right)-85.449}{0.2303}\right)/10$$

$$\mathrm{Female}\mathrm{tibia}\mathrm{length}\left(\mathrm{cm}\right)=\left(\frac{(\mathrm{stature}\left(\mathrm{cm}\right)-86.560}{0.2129}\right)/10$$

The distal tibia was cropped to 5% of the tibia length for BMD analysis. Applying a defined percentage of limb length allows for comparisons between and within individuals by objectively standardizing the distal tibia region of interest. ²⁷ CT attenuation values were converted to gHA/cm³ using the HA phantom. We measured bone volume (cm³) and bone mineral density (BMD, gHA/cm³).

Statistical Analyses

Statistical analyses for MRI pathology were performed using SPSS Statistics (version 29.0.0.0; IBM Corp, Armonk, NY), and graphs were made using R Studio (version 4.3.1); BMD was analyzed using Stata Statistical Software, release 18 (StataCorp, College Station, TX), and graphs were made using GraphPad Prism, version 10.1.1 (GraphPad Software, Boston, MA). Descriptive statistics for participant age and time since injury for the HxAI group, and body mass index (BMI) were determined and reported as mean ± SD. The presence or absence of MRI injury pattern pathology and OA features in the control and HxAI groups were compared using cross-tabulation and Pearson chi-square analyses for the associations and differences between the 2 groups. The MRI median sum score, the interquartile range (IQR) for the 25th and 75th quartiles, and range were reported. The within-subject frequency of MRI injury pattern pathology and OA features for the HxAI group was tested using the McNemar paired samples proportion test. BMD by skeletal site was analyzed using a mixed effects linear regression with restricted maximum likelihood, and Kenward-Rodger degrees of freedom were employed to produce between- and within-group comparisons fitting a full factorial (group by time) model. This model formulation allowed for a between group comparison using average marginal effects while also allowing for a within group, between leg, comparison through linear combinations. Finally, the interaction term compared, across injury groups, the average within group differences between legs (for the injured group this would effectively compare injured to uninjured legs, whereas for the uninjured groups this would result in a side-to-side comparison). Statistical significance was set to P < .05.

Results

Demographics and Clinical Characteristics

Each group included 7 females and 3 males. The mean age at the time of imaging in the HxAI group was 22.1 (SD 3.7) years, with a median of 7.0 (5.0-12.0) years from injury. The mean age of the control group was 23.1 (SD 3.7) years. The mean BMI of the HxAI group was 23.7 (SD 3.9) whereas the BMI of the Control group was 21.9 (SD 2.6). The participants injured their ankle in a variety of sports, including soccer (n = 2), basketball/netball (n = 2), gymnastics (n = 2), volleyball (n = 1), football (n = 1), ski/snowboard (n = 1), and skateboarding (n = 1). At the time of the study visit, within the HxAI group the participants were either classified as active (n = 9) or moderately active (n = 1). Although the HxAI remained relatively active, they also reported poor symptoms and functional limitations (Table 1).

Table 1.

History of Ankle Injury Group (HxAI) Activity Levels and Symptoms/Functional Limitations. ^a

Measure	Median (IQR)	Range
Godin Leisure Time and Exercise	45 (28–69)	15–92
FAOS Foot and Ankle Survey
Symptoms/ Stiffness	6 (4–8)	0–14
Pain	3 (1–5)	0–12
Daily living	1 (0–1)	0–9
Sport and Recreational	3 (1–4)	0–5
Quality of Life	3 (2–4)	0–4
Total FAOS	16 (9–25)	0–39

^aThe Godin Leisure Time & Exercise survey and the Foot and Ankle Outcome Score (FAOS) Foot and Ankle survey. Data are presented as median, 25th and 75th interquartile range (IQR), and range.

MRI Pathology

MRI injury pattern pathology was significantly higher in the HxAI group compared to the Control group. In the HxAI group, all participants exhibited MRI injury pattern pathology with a median sum score of 2 (IQR 1-6, range 1-14) in the injured ankle, whereas 1 of 10 participants in the Control group exhibited MRI injury pattern pathology with a median sum score of 0 (IQR 0-0, range 0-4) (χ² (1, n = 20) = 16.36, P < .001) (Figure 3A). Within the HxAI group, 10 of 10 injured ankles and 8 of 10 uninjured ankles showed observable pathology with a median sum score of 2 (IQR 1-6, range 1-14) in the injured ankle and a median sum score of 2 (IQR 1-5, range 0-7) in the uninjured ankle (P = .16). The HxAI group had significantly greater OA features compared with the Control group. In the HxAI group, 3 of 10 of the participants exhibited OA features with a median sum score of 0 (IQR 0-3, range 0-4) in the injured ankle, whereas none (0 of 10) of the participants in the Control group had observable OA features with a median sum score of 0 (IQR 0-0, range 0-1), χ²(1, n = 20) = 3.53, P = .06 (Figure 3B). Within the HxAI group, 3 of 10 injured ankles and 2 of 10 uninjured ankles presented with OA features, with a median score of 0 (IQR 0-3, range 0-4) in the injured ankle and a median score of 0 (IQR 0-0, range 0-2) in the uninjured ankle (P = .32). MRI results are reported in detail in Appendixes 3 and 4, with representative MRIs depicting various pathology in Appendix 5.

Figure 3.

Boxplots illustrating magnetic resonance imaging (MRI) ankle scoring. (A) The sum injury pattern pathology score is composed of scores for the lateral ankle ligaments, syndesmotic ligaments, deltoid and tibiospring ligaments, spring ligament complex, and sinus tarsi ligaments, along with the peroneal, flexor, extensor retinacula and tendons, bone (fibula, tibia, calcaneus, navicular), talar osteochondral lesions, and effusion in the tibiotalar and talocalcaneal joints. (B) The sum osteoarthritis (OA) features score is composed of scores for osteophytes (tibiotalar, subtalar), cartilage loss (tibiotalar, subtalar), and bone marrow lesions in the fibula, tibia, and talus. (C) The sum MRI pathology is the combined scores from the injury pattern pathology and OA features. HxAI, history of ankle injury.

BMD Analysis

BMD was significantly higher in the HxAI group compared with the control group in the talus (mean difference = 88.5 [SE 17.2] g/cm³, 95% CI 41.0-136.1, P < .001), calcaneus (mean difference = 65.1 g/cm³, 95% CI 22.9-110.1, P < .005), navicular (mean difference = 89.8 g/cm³, 95% CI 45.7-134.0, P < .001), and distal tibia (mean difference = 98.2 g/cm³, 95% CI 34.7-161.8, P = .002). In the talus there was an interaction between group and ankle (t = −2.14, P = .046) for talus BMD. Talus BMD in the HxAI group was significantly lower (mean difference = −13.1 g/cm³, 95% CI −21.4 to 5.0, P = .002) in the injured ankle (502.4 g/cm³, 95% CI 468.5-536.3) compared with the uninjured ankle (515.6 g/cm³, 95% CI 481.7-549.4). In the control group, BMD was not significantly different between limbs for any of the bones investigated (Figure 4A). There was no significant interaction group by limb at the other skeletal sites (Figure 4B-D).

Figure 4.

Bone mineral density (BMD) for the control and ankle injury groups. BMD of the (A) talus, (B) calcaneus, (C) navicular, and (D) distal 5% of the tibia for the Control and ankle injury history (HxAI) groups between the matched uninjured and injured ankles. Because the Control group had no history of ankle injury, the side for the injured and uninjured ankles were matched based on their HxAI counterpart. Data are presented as the median with interquartile range (25th and 75th percentile); whiskers represent the minimum and maximum values.

Discussion

Our study highlights the significant structural changes in the ankle 3-15 years following adolescent SAS, including lower talus BMD in the injured ankle compared with the contralateral ankle and evidence of injury pattern pathology and OA features on MRI. These findings in this young cohort may suggest future detrimental implications as PTOA is typically progressive. ³¹ The lower talus BMD indicates that bone loss may occur and persist following an SAS. Interestingly, injury pattern pathology and OA features were also observed in the uninjured ankle of the HxAI group, suggesting broader implications for a unilateral ankle sprain or overall sports participation.

The injury pattern pathology reported here supports our previous findings that reported adolescent SASs resulted in overall poorer outcomes, which included more ankle pain and symptoms, reduced self-reported ankle function and ankle-related quality of life, reduced sports participation, poorer dynamic balance, and a greater fear of pain compared with controls. ²² Ankle injuries are often perceived as benign, ¹⁶ and clinically assessed strength and range of motion suggest full recovery after 1 month postinjury. ¹ However, the combination of the previously reported health outcomes ²³ and structural changes suggests there are potential long-term health consequences following an SAS during adolescence. Additionally, the use of both MRI and CT is a strength of this study allowing for the comprehensive assessment of the ankle and will help to identify individuals at a greater risk of developing PTOA.

The structural changes assessed using MRI in the injured ankle also coincide with changes in the talus BMD. Our findings suggest that SAS during adolescence may be associated with BMD changes in the talus, with the talus exhibiting a decrease in BMD in the injured ankle compared with the uninjured ankle. There is conflicting evidence on the direction of BMD change in relation to OA as some studies suggest an increase in BMD due to subchondral sclerosis and/or osteophytes,^14,15 whereas other studies have reported lower BMD in OA patients.^17,26 These discrepancies are likely due to differences in disease course when BMD is assessed as well as anatomical location (eg, subchondral bone plate or subchondral trabecular bone).^18,20 Earlier stages of OA are hypothesized to be associated with lower BMD and later stages with higher BMD.^7,20 Our study supports this hypothesis as the talus had lower BMD in the injured ankle of the HxAI group. To put into perspective, the magnitude of the significantly lower BMD in the talus of the injured ankle observed in this study exceeds changes previously reported at the distal tibia due to growth from age 16-19 years ⁹ and the differences between landing and take-off legs in figure skaters. ⁸ Although the control group were age- and sex-matched to the HxAI group, overall, the controls had lower BMD across the ankle skeletal sites. Given the high rates of ankle injury, the controls may potentially not have been as physically active as the HxAI group, leading to lower BMD due to less loading of the bones compared with those who participated in youth sports.

A surprising finding was the presence of MRI pathology in the uninjured ankle in the HxAI group. This may reflect the high level of sports participation in this group, and potentially previous minor to moderate ankle sprains that did not meet our inclusion criteria. There is also a possibility of participants having kinesiophobia, potentially resulting in less loading on the index ankle and compensating with more loading on the uninjured ankle, thereby predisposing it to less severe sprains over the years. This finding will require a larger sample size and longer follow-up to determine whether it is relevant to OA progression.

Differences in disease progression may exist between ankle, knee, and hip OA. BMLs have also been associated with OA progression. Recently in the context of PTOA, a randomized controlled trial tracked the progression of BMLs as an outcome following knee meniscal injury. This study found more severe BMLs detected in young patients with isolated traumatic meniscal tears without radiographic OA compared with controls and worsening BMLs 24 months following injury in some patients who underwent surgery compared to those who underwent physical therapy with delayed surgery. ²⁹ Further, our previous research in the knee has also demonstrated bone loss at the site of acute BMLs (bone bruises) following ACL injury. ¹⁹ Although PTOA has defined clinical criteria in the knee, there are important differences between the knee and ankle that may limit the extent to which research findings from the knee are applicable to the ankle. ¹⁰ Most importantly, the etiology of hip and knee OA is largely associated with primary OA, whereas the majority of ankle OA is related to injury leading to PTOA. ²⁸ The combination of the rates of ankle injuries in adolescent sports and earlier onset and progression of PTOA highlight the need to study SAS ankle PTOA. Additionally, PTOA is further complicated by the changes in an individual’s lifestyle, typically resulting in an increase in BMI, which puts individuals who have suffered an SAS at further risk of developing primary OA. ³²

A strength of this study was that the HxAI group was able to act as their own internal control, and the control group was age- and sex-matched to the HxAI group to reduce variability. The multimodal imaging including CT and MRI provided a comprehensive imaging to assess the participants’ ankle structure. However, recruiting age- and sex-matched controls without any history of ankle injuries posed challenges. The original cohort study used similarly active controls ²³ ; however, because of the timing of this study and COVID-related delays, we needed to recruit a new set of controls. This ultimately led to a small sample size for this study, perhaps limiting the application and generalizability of these findings. The lower BMD of the control group suggests that they may have been less physically active than their HxAI counterparts, although activity levels were not quantified in the present study. A notable strength of the study design was the combination of MRI and CT scans, enabling a comprehensive analysis of the ankle joint structure and BMD. Although joint space narrowing is a feature of PTOA, we did not perform weightbearing imaging and therefore did not perform this analysis. Examining joint space narrowing in this population should be included in future research. Additionally, another key limitation is that the time after injury varied from 3 to 15 years post-SAS among the HxAI group which could have a significant influence on the imaging findings and disease progression.

It remains unknown if bone degradation is an early indicator that precedes cartilage degradation in the ankle in relation to PTOA. Additionally, the contribution of activity levels, participation in various sports, and ankle stability in relation to imaging outcomes is unclear. This study demonstrates that SAS sustained during youth sport may be associated with osteochondral changes following a 3-15-year follow-up as seen on MRI and CT BMD analysis. In conclusion, our findings underscore the need for investigation focus on a better understanding of the potential lasting effects of SAS injuries, particularly in the context of development of PTOA.