Midterm Outcomes After Suture Button Fixation for Arthroscopically Confirmed Concomitant Syndesmotic Instability in Chronic Lateral Ankle Instability: A Retrospective Cohort Study

Abstract

Background:

Distal tibiofibular syndesmosis instability (DTSI) often coexists with chronic lateral ankle instability and may delay recovery. Suture button fixation is widely used, but its benefit in this setting is uncertain.

Purpose:

To evaluate midterm outcomes of suture button fixation for DTSI with concomitant lateral ankle instability.

Study Design:

Cohort study; Level of evidence, 3.

Methods:

Patients who underwent surgery for lateral ankle instability with arthroscopic syndesmosis assessment between 2017 and 2023 were retrospectively identified. Patients with an intact syndesmosis formed the intact group, whereas those with arthroscopically confirmed dynamic syndesmotic instability (≥2-mm diastasis) underwent additional suture button fixation and formed the instability group. All patients received standardized lateral ligament surgery and rehabilitation. Clinical assessments were performed preoperatively, at 12 months, and at final follow-up using Karlsson-Peterson, Foot and Ankle Ability Measure (FAAM), Tegner, Ankle Ligament Reconstruction–Return to Sport after Injury (ALR-RSI), visual analog scale (VAS) pain scores, and ankle range of motion. Group differences were analyzed with α = .05 and interpreted using minimal clinically important differences (MCIDs).

Results:

Of 216 eligible patients, 154 completed follow-up at a mean of 67.3 ± 24.8 months (range, 24-113 months). No bilateral cases were included; therefore, the number of ankles was identical to the number of patients: 80 patients (80 ankles) in the intact syndesmosis group and 74 patients (74 ankles) in the syndesmotic instability group. Preoperatively, the instability group had lower Karlsson-Peterson scores than the intact group, with a median between-group difference of −4.00 points (95% CI, −6.00 to −2.00; P < .001). At 12 months, Karlsson-Peterson scores were comparable between groups (median between-group difference, 0.00 points; 95% CI, −1.00 to 1.00; P = .63). At final follow-up, the instability group had a statistically higher Karlsson-Peterson score than the intact group, although the absolute difference was small (median between-group difference, 1.00 point; 95% CI, 0.00 to 2.00; P = .02). The between-group difference in Karlsson-Peterson improvement was 4.88 points (95% CI, 3.13 to 6.63; P < .001), which did not exceed the MCID threshold of 13.50 points. After adjustment for preoperative Karlsson-Peterson score, sex, body mass index, age, anterior talofibular ligament surgical modality, and follow-up duration, syndesmotic instability treated with suture button fixation remained modestly associated with the final Karlsson-Peterson score (β = 1.525; 95% CI, 0.340 to 2.709; P = .01). Final FAAM Activities of Daily Living, FAAM Sports, Tegner, ALR-RSI, and VAS scores showed no clinically meaningful between-group differences. The instability group had greater residual dorsiflexion limitation at final follow-up, with a median between-group difference of 2.00° (95% CI, 2.00° to 3.00°; P < .001), but this limitation was not significantly correlated with final patient-reported outcomes.

Conclusion:

Our study demonstrated that, for patients with chronic lateral ankle instability and arthroscopically confirmed concomitant DTSI, suture button fixation combined with lateral ligament surgery was associated with substantial midterm functional improvement, with outcomes approaching those of patients with an intact syndesmosis. However, because the comparison involved biologically distinct cohorts, these findings should be interpreted as supportive rather than definitive. Future matched or randomized studies are needed to clarify the independent contribution of suture button fixation.

Ankle sprains are among the most prevalent and burdensome musculoskeletal injuries in sports medicine. ³⁹ They are broadly categorized into low and high ankle sprains, with distal tibiofibular syndesmosis instability (DTSI) representing a distinct subtype of high ankle sprain, which accounts for 1% to 20% of all cases. ²⁸ DTSI is especially prevalent in collision and field sports.^9,13 While the isolated incidence of DTSI is lower than that of lateral ankle instability, it frequently occurs concomitantly, ²⁶ which typically leads to prolonged rehabilitation periods and inferior patient outcomes.^20,32,33 Although arthroscopic probing of the distal tibiofibular joint space is commonly used as an intraoperative reference method for assessing syndesmotic instability, the optimal threshold for defining dynamic instability, such as 2 to 3 mm of maximal separation, remains debated.^3,7,11,17,37 Furthermore, preoperative magnetic resonance imaging (MRI) parameters, including the lambda sign and the height of distal tibiofibular joint effusion, serve as valuable diagnostic indicators for DTSI. ³⁸

From a therapeutic standpoint, screw and suture button constructs represent the principal fixation strategies for managing DTSI.^12,18,20,31 Suture button fixation has become increasingly favored due to its capacity to permit physiological micromotion, circumvent the need for secondary hardware removal, and expedite early rehabilitation. ¹⁹ Nevertheless, a paucity of evidence remains regarding its long-term functional and psychological outcomes. ⁴³ The existing literature largely centers on comparing the clinical efficacy of suture button and screw fixation.^10,22,30,45 However, there is a notable lack of research investigating the multidimensional outcome differences between 2 biologically distinct populations: patients treated with surgical fixation for syndesmotic instability and those with an intact syndesmosis.

Given that prior studies have demonstrated broadly comparable clinical efficacy across various surgical techniques for lateral ankle ligament repair or reconstruction,^15,24 we retrospectively evaluated patients with chronic lateral ankle instability (CLAI) with or without concomitant DTSI. The purpose of this study was to evaluate whether patients with CLAI and arthroscopically confirmed concomitant DTSI treated with additional suture button fixation achieve midterm functional, psychological, pain, and range of motion (ROM) outcomes comparable with patients with CLAI and an intact syndesmosis. We hypothesized that despite worse preoperative function, patients with concomitant syndesmotic instability treated with additional suture button fixation would demonstrate substantial midterm improvement and would not show clinically meaningful inferior outcomes compared with patients with an intact syndesmosis.

Methods

Study Design and Ethics

This single-center retrospective cohort study was conducted at Chinese PLA General Hospital between January 2017 and December 2023. Patients who underwent ankle-related surgery during this period were retrospectively identified through the preoperative electronic medical records system. The study was conducted in accordance with the Declaration of Helsinki and was approved by the ethics committee of the institution. Written informed consent for inclusion in the study and use of anonymized clinical data were obtained at the time of final follow-up.

Patients

We retrospectively identified patients who underwent surgery for lateral ankle instability with arthroscopic syndesmosis assessment between 2017 and 2023. Depending on the stability of the DTS—as determined by intraoperative arthroscopic dynamic evaluation—patients were allocated into 1 of 2 groups: an intact syndesmosis group (group 1) and a syndesmotic instability group (group 2).

Eligibility Criteria and Grouping

Patients were included in this study if they met the following criteria: (1) age between 18 and 60 years; (2) a definitive diagnosis of CLAI meeting strict surgical indications; (3) arthroscopically assisted anterior talofibular ligament (ATFL) surgery, including repair, reconstruction, or repair with augmentation; (4) either an intact syndesmosis or dynamic syndesmotic instability, operationally defined in this study as a maximal dynamic gap ≥2 mm during intraoperative arthroscopic probing; and (5) availability of complete clinical records, standardized postoperative rehabilitation documentation, and clinical follow-up assessments at 12 months postoperatively and at final follow-up.

Patients were excluded from the study if they met any of the following criteria: (1) osteochondral lesions of the distal tibia or talus, as ruled out by a combination of preoperative MRI and intraoperative arthroscopic assessment; (2) bilateral ankle sprains; (3) active intra-articular infection of the ankle; (4) concomitant injury to other tendons around the ankle; (5) equivocal dynamic syndesmotic widening of >0 but <2 mm on arthroscopic probing, to avoid overlap between the intact and unstable syndesmosis groups; and (6) loss to follow-up or incomplete clinical data. Group 1 included patients with an arthroscopically intact syndesmosis, whereas group 2 included patients with dynamic syndesmotic diastasis of ≥2 mm who underwent additional suture button fixation.

The strict indications for lateral ankle surgery were determined through a comprehensive assessment of clinical and MRI findings, specifically requiring all of the following: (1) a history of recurrent ankle sprains or “giving way” episodes lasting >6 months; (2) failure of ≥3 months of systematic conservative treatment, including bracing, physical therapy, proprioceptive training, and peroneal strengthening; and (3) a positive anterior drawer or talar tilt test confirmed by a board-certified senior sports medicine specialist with >20 years of experience in foot and ankle surgery (M.W.).

Patient screening and data verification were performed using the preoperative electronic medical records system. Potentially eligible patients were screened according to the predefined inclusion and exclusion criteria described above. Data integrity was independently verified by 2 orthopaedic surgeons specializing in foot and ankle surgery (P.T.S. and Y.W.). Any disagreements regarding eligibility, grouping, or data extraction were resolved through discussion with a senior foot and ankle surgeon (M.W.).

Notably, stress radiographs were not used as a routine preoperative standard for evaluating DTSI in this cohort because standardized stress radiography equipment and protocols were not routinely available at Chinese PLA General Hospital during the study period. Additionally, radiographic assessment of syndesmotic diastasis on 2-dimensional radiographs is susceptible to measurement variability because these measurements are influenced by ankle rotation and the difficulty of reproducing standardized ankle positioning.^23,36 Intraoperative arthroscopic dynamic probing under anesthesia, which minimizes the influence of muscle guarding, was used as the intraoperative reference standard in this study for confirming lateral ligament injury and classifying syndesmotic stability. For the purposes of this study, dynamic syndesmotic instability was operationally defined as a maximal arthroscopically probed gap of ≥2 mm.^40,41

Exposure Definition and Group Allocation

The exposure of interest was the intraoperative arthroscopic status of the DTS. Syndesmotic stability was assessed dynamically under anesthesia using an arthroscopic probe. If the probe could not be inserted into the distal tibiofibular interval and no dynamic widening was observed, the syndesmosis was classified as intact, and the patient was assigned to the intact syndesmosis group (group 1). If the probe could be inserted between the tibia and fibula and dynamic diastasis of ≥2 mm was observed during probing, the syndesmosis was classified as unstable according to the study definition, and the patient was assigned to the syndesmotic instability group (group 2). Patients with dynamic widening <2 mm were not included in the present analysis. Patients in group 2 underwent additional suture button fixation as part of the operative procedure.

Surgical Procedures

All procedures were performed with the patient in the supine position under general or spinal anesthesia. A thigh tourniquet was applied, and the operative limb was positioned at the end of the operating table to allow free ankle motion. Noninvasive ankle distraction was applied when necessary to facilitate visualization of the tibiotalar joint. Before portal creation, the superficial peroneal nerve and anterior tibial tendon were identified and marked. Standard anteromedial and anterolateral ankle arthroscopy portals were then established. The anteromedial portal was created first, just medial to the anterior tibial tendon at the level of the ankle joint line. The anterolateral portal was subsequently established under direct visualization, lateral to the peroneus tertius tendon, while avoiding the superficial peroneal nerve. Skin incisions were made sharply, and blunt dissection was used to enter the joint to minimize the risk of neurovascular injury.

Diagnostic ankle arthroscopy was then performed in a systematic manner. Synovectomy and debridement of impinging soft tissue or scar tissue were carried out as necessary. The DTS was inspected and dynamically assessed with an arthroscopic probe. The ATFL was treated according to intraoperative tissue quality, generalized ligamentous laxity, revision status, and repairability. When the residual ATFL and capsular tissue were of adequate quality, an arthroscopic anatomic Broström-Gould repair was performed using suture anchors (Arthrex). When the ATFL remnant was repairable but attenuated, repair was reinforced with a high-strength suture tape augmentation construct (Arthrex). In cases of poor tissue quality, generalized ligamentous laxity, or revision surgery, anatomic ATFL reconstruction was performed using tendon autograft or allograft according to graft availability and patient-specific factors.

Preoperative MRI scans were reviewed for the presence of the lambda sign, and the height of fluid accumulation in the distal tibiofibular joint was measured. During arthroscopic evaluation, the DTS was probed for dynamic instability. In cases of dynamic instability with a probed gap ≥2 mm, a commercially available suture button fixation device (TightRope; Arthrex) was inserted percutaneously approximately 3 cm above the tibiotalar joint line to stabilize the syndesmosis. Fluoroscopy was used to confirm anatomic reduction and appropriate tensioning, with the construct tensioned under fluoroscopic guidance to avoid excessive internal rotation or overcompression of the syndesmosis. All surgeries were performed by the senior surgeon.

Postoperative Rehabilitation

During the first 2 postoperative weeks, the ankle was immobilized in a brace, and patients remained nonweightbearing with crutches. The goals during this period were wound protection, swelling control, pain management, and maintenance of knee, hip, and toe motion. Active ankle ROM exercises were not performed during the initial immobilization period.

After wound healing, usually beginning at 2 weeks postoperatively, the immobilization device was removed only during supervised or instructed exercises, and gentle active ankle ROM exercises were initiated within a pain-free and tolerable range, primarily including dorsiflexion, plantarflexion, and subtalar joint motion. Isometric strengthening of the ankle and surrounding muscle groups was also started during this phase, while patients otherwise continued protected immobilization and nonweightbearing ambulation.

Starting at 4 weeks postoperatively, weightbearing was gradually introduced, with an emphasis on improving ankle muscle strength and performing lower limb closed-chain exercises, in addition to proprioceptive and balance training. At between 8 and 12 weeks, sport-specific training was initiated, including dynamic exercises such as jumping, cutting, acceleration, and deceleration, focusing on improving neuromuscular control and joint stability. Patients were allowed to gradually return to running, jumping, and contact sports once they met functional recovery criteria and had adequate pain and swelling control. Functional recovery was assessed using standardized measures such as the FAAM and Tegner activity scale. Return to sport was generally recommended at approximately 6 months postoperatively, depending on the achievement of functional milestones. A full return to normal activities was expected by approximately 1 year postoperatively. Throughout the rehabilitation process, training intensity and progression were adjusted dynamically based on the individual's recovery, with regular assessments of functional capacity.

Preoperative Assessment

Preoperative MRI was reviewed to evaluate DTS injury and the lambda sign. The lambda sign was defined as a high-signal fluid extension from the ankle mortise into the DTS on coronal fluid-sensitive MRI sequences, creating a configuration resembling the Greek letter lambda (Figure 1A). To quantify distal tibiofibular joint effusion height, a reference line was drawn along the tibial plafond on the coronal MRI plane, and a perpendicular line was drawn from this reference line to the most proximal point of the distal tibiofibular joint effusion. The maximal measured distance was recorded as the effusion height. When no distal tibiofibular joint effusion was present, the effusion height was recorded as 0 mm, and the lambda sign was considered negative.^34,38,46

Figure 1.

Imaging and arthroscopic assessment of distal tibiofibular syndesmosis instability (DTSI). (A) Preoperative coronal T2 fat-suppressed magnetic resonance imaging (MRI) showing synovial fluid extension into the distal tibiofibular syndesmosis within 1 month before surgery. (B) Intraoperative arthroscopic view from the anteromedial portal showing probing of the distal tibiofibular joint space through the anterolateral portal to assess dynamic instability and record the maximal gap. (C) Postoperative anteroposterior radiograph showing suture button fixation for DTSI.

Outcome Measures and Follow-up

The primary outcome was the Karlsson-Peterson score. Secondary outcomes included the visual analog scale (VAS) pain score, Foot and Ankle Ability Measure (FAAM) Activities of Daily Living subscale (ADL), FAAM Sports subscale, Ankle Ligament Reconstruction–Return to Sport after Injury (ALR-RSI) score, Tegner activity scale, active ankle dorsiflexion and plantarflexion ROM, complications, and reoperations. Active ankle ROM was measured with a goniometer and reported as the difference between the operative ankle and the contralateral ankle.

Clinical outcomes were assessed preoperatively, at 12 months postoperatively, and at the final follow-up. Preinjury functional status was collected as a retrospective self-assessment and was used to describe the patient's recalled functional level before the index injury. The 12-month follow-up was used to evaluate early postoperative recovery. The final follow-up was not a fixed 1-year assessment; rather, it was defined as the most recent available postoperative clinical assessment performed after the 12-month visit. In the final analysis cohort, the mean final follow-up duration was 67.3 ± 24.8 months (range, 24-113 months), and this assessment was used to evaluate midterm outcomes.

Postoperative follow-up was performed during routine outpatient visits. Patients included in the final analysis were required to have ≥1 postoperative imaging evaluation and 2 clinical outcome assessments.

Postoperative MRI was not mandated at predefined time points for all patients. MRI examinations after surgery were obtained during routine follow-up or when clinically indicated, depending on patient availability, symptoms, and willingness to undergo additional imaging. Therefore, serial postoperative MRI findings were available only in a limited subset of patients and were used descriptively to illustrate postoperative imaging changes rather than as a formal longitudinal imaging outcome measure. No quantitative correlation analysis was performed between postoperative syndesmotic effusion or bone marrow edema and clinical outcomes.

Sample Size Considerations

Given the retrospective cohort design, no formal a priori sample-size calculation was performed. The final sample size was determined by the number of eligible patients with complete follow-up data. A post hoc precision assessment was performed for the primary outcome, and the magnitude and precision of between-group differences were interpreted using 95% CIs and minimal clinically important difference (MCID) thresholds.

Statistical Analysis

Statistical analyses were performed using R Version 4.3.1 (R Foundation for Statistical Computing). The normality of continuous variables was evaluated using the Shapiro-Wilk test. Normally distributed continuous variables are presented as mean ± SD, whereas nonnormally distributed continuous variables were expressed as median with interquartile ranges (Q1-Q3). Categorical data are presented as frequency (percentage). Baseline comparisons between groups were conducted using independent-samples t tests or Mann-Whitney U tests for continuous variables and chi-square tests or Fisher exact tests for categorical variables.

The MCID was defined as the smallest change in an outcome score that patients perceived as clinically meaningful. For outcomes with available preoperative baseline and final follow-up data, including the Karlsson-Peterson score, FAAM Sports, FAAM ADL, Tegner activity scale, and VAS pain score, MCID thresholds were calculated based on the change from preoperative baseline to final follow-up.

Because preoperative ALR-RSI scores and ankle ROM limitation values were not available, these outcomes were not analyzed as baseline-to-final improvement outcomes. Instead, they were evaluated separately as exploratory late postoperative change outcomes using paired data from the 12-month postoperative visit to the final follow-up. For ALR-RSI, improvement was defined as an increase in score. For plantarflexion and dorsiflexion limitation, improvement was defined as a reduction in the side-to-side limitation compared with the contralateral ankle.

Within-group comparisons were performed according to the available time points for each outcome. For outcomes with available preoperative baseline data, including the Karlsson-Peterson score, FAAM Sports, FAAM ADL, Tegner activity scale, and VAS pain score, changes from preoperative baseline to 12 months postoperatively and to the final follow-up were evaluated. For ALR-RSI and ankle ROM limitations, changes were evaluated from the 12-month postoperative visit to the final follow-up because preoperative values were not available. Given the nonnormal distribution of most repeated outcome measures, within-group longitudinal changes were assessed using the Friedman test and reported as χ² statistics with corresponding P values. For outcomes with only 2 postoperative time points, including ALR-RSI and ROM limitations, the Friedman test was also used to evaluate exploratory late postoperative changes from 12 months to final follow-up, allowing within-group statistics to be reported consistently across outcomes.

For both baseline-to-final outcomes and exploratory late postoperative change outcomes, anchor-based receiver operating characteristic (ROC) curve analysis was performed using patients’ self-reported global improvement at the final follow-up as the external anchor. Patients reporting substantial improvement were compared with those reporting minimal improvement. The cutoff point maximizing the Youden index was selected as the MCID threshold. The Youden index was calculated as sensitivity + specificity − 1. The discriminatory ability of each MCID threshold was evaluated using the area under the ROC curve (AUC), sensitivity, specificity, and Youden index. An AUC >0.70 was interpreted as good discriminatory ability, 0.60 to 0.70 as fair discriminatory ability, and <0.60 as poor discriminatory ability.

Responder analysis was performed by calculating the proportion of patients whose change met or exceeded the corresponding MCID threshold. For Karlsson-Peterson, FAAM Sports, FAAM ADL, Tegner activity scale, and VAS pain, responder status was based on change from preoperative baseline to final follow-up. For ALR-RSI and ankle ROM limitations, responder status was based on exploratory change from 12 months postoperatively to final follow-up. These exploratory postoperative change analyses were used only to contextualize late recovery and were not interpreted as evidence of baseline-to-final treatment benefit. The clinical significance of treatment effects was assessed using 3 complementary approaches: comparison of between-group mean differences with corresponding MCID thresholds, responder analysis based on the proportion of patients achieving the MCID, and combined interpretation of statistical significance and clinical relevance. All statistical analyses were 2-sided, with the significance level set at α = .05.

To address potential confounding from baseline imbalances, an expanded analysis of covariance (ANCOVA) model was used for the primary outcome. The Karlsson-Peterson score at final follow-up was entered as the dependent variable, and group allocation was entered as the primary exposure. The model adjusted for preoperative Karlsson-Peterson score, sex, body mass index (BMI), age, ATFL surgical modality, and follow-up duration. Weight was not included because of collinearity with BMI. ATFL surgical modality was treated as a categorical covariate, with arthroscopic Broström-Gould repair as the reference category. Adjusted effects are reported as unstandardized regression coefficients (β) with 95% CIs and P values. Model collinearity was assessed using variance inflation factors. To examine the robustness of the group effect, a heteroscedasticity-consistent HC3 sensitivity analysis was also performed.

Results

Recruitment Process

A total of 752 patients who underwent ankle-related surgery at Chinese PLA General Hospital between January 2017 and December 2023 were initially screened for eligibility. After exclusion of 536 patients who did not meet the eligibility criteria, including 389 patients with concomitant talar or tibial osteochondral lesions, 21 undergoing bilateral ankle ligament repair or reconstruction, 22 with isolated DTS injury without lateral ligament damage, 28 with accessory navicular deformities, 41 aged <18 years or >60 years, and 35 with distal tibiofibular diastasis <2 mm and no suture button fixation, 216 patients were eligible for follow-up analysis. Of these, 62 patients were excluded because of missing or incomplete follow-up data, leaving 154 patients in the final analysis (Figure 2).

Figure 2.

Flow diagram of patient inclusion and analysis.

A total of 80 patients were assigned to the intact DTS group, and 74 patients were assigned to the DTSI group. In the instability group, 63 patients demonstrated a diastasis of ≥4 mm on arthroscopic evaluation (Figure 2).

Baseline Characteristics

As shown in Table 1, a total of 154 patients were included in the final analysis, including 80 patients in the intact DTS group (group 1) and 74 patients in the DTSI group treated with additional suture button fixation (group 2). There were no statistically significant between-group differences in age, height, follow-up duration, surgical side, or ATFL surgical modality. However, group 1 had significantly greater body weight than group 2 (74.94 ± 11.43 kg vs 69.01 ± 12.36 kg; 95% CI, −9.73 to −2.13 kg; P = .002) and higher BMI (25.17 ± 3.11 kg/m² vs 23.77 ± 3.48 kg/m²; 95% CI, −2.45 to −0.35; P = .01). Sex distribution also differed significantly between groups, with a lower proportion of female patients in group 1 than in group 2 (23.8% vs 55.4%; P < .001). Preoperative distal tibiofibular joint effusion height was significantly greater in group 2 than in group 1 (11.30 ± 4.79 mm vs 5.65 ± 5.67 mm; between-group difference, 5.65 mm; 95% CI, 3.98 to 7.32 mm; P < .001).

Table 1.

Comparison of Baseline Characteristics Between the 2 Groups ^a

Variable	Level	Total	Group 1	Group 2	95% CI	P
n		154	80	74
Age, y		32.35 ± 9.81	33.17 ± 9.91	31.46 ± 9.69	−4.83 to 1.41	.28
Height, cm		171.30 ± 8.10	172.39 ± 8.52	170.13 ± 7.51	−4.81 to 0.29	.08
Weight, kg		72.09 ± 12.21	74.94 ± 11.43	69.01 ± 12.36	−9.73 to −2.13	.002
BMI, kg/m2		24.50 ± 3.36	25.17 ± 3.11	23.77 ± 3.48	−2.45 to −0.35	.01
Sex	Female	60 (39.0)	19 (23.8)	41 (55.4)		<.001
	Male	94 (61.0)	61 (76.3)	33 (44.6)
Follow-up time, mo		67.27 ± 24.76	67.60 ± 25.26	66.91 ± 24.37	−8.59 to 7.21	.86
Surgical side	Left	75 (48.7)	42 (52.5)	33 (44.6)		.41
	Right	79 (51.3)	38 (47.5)	41 (55.4)
ATFL surgical modality	Arthroscopic Broström-Gould repair	55 (35.7)	26 (32.5)	29 (39.2)		.56
	Broström-Gould repair with suture tape augmentation	56 (36.4)	31 (38.8)	25 (33.8)
	Isolated suture tape augmentation	19 (12.3)	12 (15.0)	7 (9.5)
	Anatomic ATFL reconstruction	24 (15.6)	11 (13.8)	13 (17.6)
DTJE height, mm		8.36 ± 5.96	5.65 ± 5.67	11.30 ± 4.79	3.98 to 7.32	<.001

^aData are presented as mean ± SD or n (%). The 95% CI refers to the between-group difference, calculated as group 2 minus group 1. Group 1 indicates the intact distal tibiofibular syndesmosis group, and group 2 indicates the syndesmotic instability group treated with additional suture button fixation. ATFL surgical modality refers to the lateral ligament procedure: Broström-Gould repair indicates anatomic ATFL repair with suture anchors; repair with suture tape augmentation indicates Broström-Gould repair reinforced with a high-strength suture tape/internal brace construct; isolated suture tape augmentation indicates suture tape reinforcement without graft reconstruction; and anatomic ATFL reconstruction indicates reconstruction using tendon autograft or allograft. ATFL, anterior talofibular ligament; BMI, body mass index; DTJE, distal tibiofibular joint effusion.

Adjusted and Sensitivity Analyses

An expanded ANCOVA model was used to evaluate whether group allocation remained associated with the final follow-up Karlsson-Peterson score after accounting for baseline imbalance. After adjustment for preoperative Karlsson-Peterson score, sex, BMI, age, ATFL surgical modality, and follow-up duration, group 2 remained associated with a higher final Karlsson-Peterson score than group 1 (β = 1.525; 95% CI, 0.340 to 2.709; P = .01). No other covariate was significantly associated with the final Karlsson-Peterson score in the expanded model. Variance inflation factors were <1.40 for all predictors, indicating no meaningful multicollinearity. A heteroscedasticity-robust HC3 sensitivity analysis yielded a similar group effect (β = 1.525; 95% CI, 0.450 to 2.599; P = .006), supporting the stability of the adjusted association (Table 2). Because the comparison involved biologically distinct cohorts rather than randomized treatment groups, this adjusted association should be interpreted as supportive but not causal.

Table 2.

Expanded ANCOVA Model for the Karlsson-Peterson Score at Final Follow-up ^a

Variable	Unstandardized Coefficient (β)	95% CI	P
Group 2 vs group 1	1.525	0.340 to 2.709	.01
Karlsson-Peterson score, preoperative	0.076	−0.029 to 0.181	.16
Sex, male vs female	0.979	−0.205 to 2.163	.10
BMI, kg/m2	−0.012	−0.168 to 0.144	.88
Age, y	−0.027	−0.083 to 0.029	.34
Follow-up duration, mo	0.000	−0.021 to 0.021	.98
Broström-Gould repair with suture tape augmentation vs Broström-Gould repair	0.048	−1.150 to 1.246	.94
Isolated suture tape augmentation vs Broström-Gould repair	−0.298	−1.973 to 1.376	.73
Anatomic ATFL reconstruction vs Broström-Gould repair	0.464	−1.098 to 2.025	.56

^aThe dependent variable was the Karlsson-Peterson score at final follow-up. ATFL surgical modality was treated as categorical, with arthroscopic Broström-Gould repair as the reference category. Variance inflation factors were <1.40 for all predictors. A heteroscedasticity-consistent HC3 sensitivity analysis yielded a similar group effect (β = 1.525; 95% CI, 0.450 to 2.599; P = .006). ANCOVA, analysis of covariance; ATFL, anterior talofibular ligament; BMI, body mass index.

Comparison of Functional Scores

The comparison of functional scores between the 2 groups is shown in Table 3. Overall, patient-reported outcomes with available preoperative baseline data improved significantly over time in both groups. ALR-RSI and ROM limitations were analyzed separately as exploratory late postoperative change outcomes from the 12-month visit to final follow-up.

Table 3.

Comparison of Primary Functional Scores Between the 2 Groups ^a

Karlsson-Peterson	Group 1	Group 2	HL Shift, G2−G1 (95% CI)	P
Preinjury	98.00 [95.00 to 100.00]	99.00 [96.00 to 100.00]	0.00 (0.00 to 2.00)	.08
Preoperative	65.00 [60.00 to 68.00]	58.50 [57.00 to 63.00]	−4.00 (–6.00 to −2.00)	<.001
Postoperative, 12 mo	80.00 [78.00 to 81.00]	80.00 [78.00 to 82.00]	0.00 (–1.00 to 1.00)	.63
Final follow-up	81.00 [79.00 to 83.00]	82.00 [80.00 to 85.00]	1.00 (0.00 to 2.00)	.02
Within-group longitudinal change	χ2 = 233.104; P < .001	χ2 = 209.786; P < .001

^aData are presented as median [IQR]. The HL (Hodges-Lehmann) shift estimator is calculated as the median of all pairwise differences (group 2–group 1), with the corresponding 95% CI. Between-group comparisons were performed using the Mann-Whitney U test. Within-group longitudinal changes were assessed using the Friedman test; χ² values indicate Friedman test statistics. Group 1 indicates the intact distal tibiofibular syndesmosis group, and group 2 indicates the distal tibiofibular syndesmosis instability group treated with additional suture button fixation. “Final follow-up” refers to the most recent postoperative assessment after the 12-month visit. HL, Hodges-Lehmann.

Primary Outcome

As shown in Table 3, the Karlsson-Peterson score improved significantly over time in both groups (group 1: χ² = 233.104; P < .001; group 2: χ² = 209.786; P < .001).

Preinjury Karlsson-Peterson scores were comparable between groups. Preoperatively, group 2 had significantly lower Karlsson-Peterson scores than group 1. At 1 year postoperatively, the scores were comparable between groups. At final follow-up, group 2 showed a statistically higher Karlsson-Peterson score than group 1, although the absolute between-group difference was small.

Secondary Outcomes

The comparison of secondary functional outcomes between the 2 groups is shown in Table 4.

Table 4.

Comparison of Secondary Functional Outcome Scores Between the 2 Groups ^a

Item	Group 1	Group 2	HL Shift, G2−G1 (95% CI)	P
FAAM Sports
Preinjury	96.95 [95.62 to 100.00]	99.00 [96.75 to 100.00]	1.00 (0.00 to 2.00)	.009
Preoperative	62.00 [59.75 to 66.25]	59.17 [55.00 to 63.00]	−3.00 (–5.00 to −1.00)	.005
Postoperative, 12 mo	85.00 [81.00 to 86.00]	83.00 [80.00 to 85.00]	−1.00 (–2.00 to 0.00)	.03
Final follow-up	85.00 [82.00 to 86.25]	86.00 [83.00 to 87.75]	1.00 (0.00 to 2.00)	.08
	χ2 = 234.512; P < .001	χ2 = 205.833; P < .001
FAAM ADL
Preinjury	100.00 [100.00 to 100.00]	100.00 [100.00 to 100.00]	0.00 (0.00 to 0.00)	.002
Preoperative	70.00 [65.00 to 73.25]	67.00 [62.00 to 71.00]	−3.00 (–5.00 to −1.00)	.01
Postoperative, 12 mo	91.00 [88.00 to 94.00]	89.00 [86.00 to 93.00]	−2.00 (–3.00 to 0.00)	.02
Final follow-up	92.00 [88.00 to 94.00]	92.00 [89.25 to 95.00]	0.00 (–1.00 to 2.00)	.53
	χ2 = 213.872; P < .001	χ2 = 206.747; P < .001
Tegner
Preinjury	7.00 [6.00 to 7.00]	7.00 [7.00 to 7.00]	0.00 (0.00 to 1.00)	<.001
Preoperative	2.00 [2.00 to 3.00]	2.00 [2.00 to 3.00]	0.00 (–0.33 to 0.00)	.18
Postoperative, 12 mo	4.00 [3.00 to 4.00]	4.00 [3.00 to 4.00]	0.00 (0.00 to 0.00)	.91
Final follow-up	5.00 [4.00 to 5.00]	5.00 [4.00 to 5.00]	0.00 (0.00 to 0.00)	.41
	χ2 = 227.675; P < .001	χ2 = 212.141; P < .001
ALR-RSI
Postoperative, 12 mo	80.00 [75.00 to 83.00]	80.00 [75.00 to 82.00]	0.00 (–2.00 to 1.00)	.53
Final follow-up	82.00 [79.50 to 86.00]	82.00 [78.00 to 86.00]	0.00 (–2.00 to 3.00)	.83
	χ2 = 13.235; P < .001	χ2 = 15.783; P < .001
VAS
Preoperative	4.00 [3.00 to 5.00]	4.00 [3.00 to 5.00]	0.00 (0.00 to 0.00)	.92
Postoperative, 12 mo	1.00 [0.00 to 2.00]	1.00 [0.00 to 2.00]	0.00 (0.00 to 0.00)	.39
Final follow-up	0.50 [0.00 to 2.00]	0.00 [0.00 to 2.00]	0.00 (0.00 to 0.00)	.47
	χ2 = 142.247; P < .001	χ2 = 116.308; P < .001
Plantarflexion limitation, deg
Postoperative, 12 mo	3.00 [0.00 to 5.00]	0.00 [0.00 to 3.00]	0.00 (–2.00 to 0.00)	.009
Final follow-up	0.00 [0.00 to 3.00]	0.00 [0.00 to 3.00]	0.00 (0.00 to 0.00)	.66
	χ2 = 29; P < .001	χ2 = 14; P < .001
Dorsiflexion limitation, deg
Postoperative, 12 mo	0.00 [0.00 to 0.00]	3.00 [0.00 to 5.00]	2.00 (2.00 to 3.00)	<.001
Final follow-up	0.00 [0.00 to 0.00]	2.00 [0.00 to 3.00]	2.00 (2.00 to 3.00)	<.001
	χ2 = 10; P = .002	χ2 = 23; P < .001

^aData are presented as median [IQR]. The HL shift estimator is calculated as the median of all pairwise differences between group 2 and group 1, with the corresponding 95% CI. Between-group comparisons at each time point were performed using the Mann-Whitney U test. Within-group longitudinal changes were assessed using the Friedman test and are reported as χ² statistics with corresponding P values. For outcomes with only 2 postoperative time points, including ALR-RSI and range of motion limitations, the Friedman test was also used to evaluate exploratory late postoperative changes from 12 months postoperative to final follow-up, allowing within-group statistics to be reported consistently across outcomes. Group 1 indicates the intact distal tibiofibular syndesmosis group, and group 2 indicates the distal tibiofibular syndesmosis instability group treated with additional suture button fixation. “Final follow-up” refers to the most recent postoperative assessment after the 12-month visit. ADL, Activities of Daily Living; ALR-RSI, Ankle Ligament Reconstruction–Return to Sport after Injury; FAAM, Foot and Ankle Ability Measure; HL, Hodges-Lehmann; VAS, visual analog scale.

FAAM Sports scores improved significantly over time in both groups. Group 2 had higher preinjury FAAM Sports scores than group 1, but lower scores preoperatively and at 1 year. At final follow-up, there was no statistically significant between-group difference in FAAM Sports scores.

FAAM ADL scores also improved significantly over time in both groups. Although both groups had identical preinjury median scores of 100.00, the between-group comparison was statistically significant. Group 2 had lower FAAM ADL scores than group 1 preoperatively and at 1 year. At final follow-up, FAAM ADL scores were comparable between groups.

Tegner activity scores improved significantly over time in both groups. Group 2 had higher preinjury Tegner scores than group 1. However, there were no statistically significant between-group differences preoperatively, at 1 year, or at final follow-up.

ALR-RSI scores improved significantly from 1 year to final follow-up in both groups. No statistically significant between-group differences were observed at 1 year or at final follow-up.

VAS pain scores decreased significantly over time in both groups. There were no statistically significant between-group differences in VAS scores preoperatively (group 2 vs group 1), at 1 year, or at final follow-up.

For ankle ROM, plantarflexion limitation improved significantly from 1 year to final follow-up in both groups. At 1 year, group 2 had less plantarflexion limitation than group 1. At final follow-up, plantarflexion limitation was comparable between groups.

Dorsiflexion limitation also improved significantly from 1 year to final follow-up in both groups (group 1: χ² = 10.000, P = .002; group 2: χ² = 23.000, P < .001). However, group 2 had significantly greater dorsiflexion limitation than group 1 at 1 year: median 3.00 (IQR, 0.00-5.00) vs 0.00 (IQR, 0.00-0.00), Hodges-Lehmann shift, 2.00° (95% CI, 2.00°-3.00°; P < .001) as well as at final follow-up: median, 2.00 (IQR, 0.00-3.00) vs 0.00 (0.00-0.00), Hodges-Lehmann shift, 2.00° (95% CI, 2.00°-3.00°; P < .001).

No other patients demonstrated positive anterior drawer or varus-valgus stress test findings.

Calculation of the MCID for Baseline-to-Final and Exploratory Late Postoperative Change Outcomes

Anchor-based MCID thresholds and corresponding AUC values are summarized in Table 5. The MCID thresholds were 13.50 points for the Karlsson-Peterson score, 24.50 points for FAAM Sports, 21.25 points for FAAM ADL, 1.58 points for the Tegner activity scale, 2.50 points for VAS pain, 1.50 points for ALR-RSI, 5.50° for plantarflexion limitation, and 0.50° for dorsiflexion limitation. Discriminatory ability was good for the Karlsson-Peterson score and Tegner activity scale, fair for FAAM Sports, FAAM ADL, and ALR-RSI, and poor for VAS pain and ROM limitation outcomes.

Table 5.

Anchor-Based MCID Analysis for Baseline-to-Final and Exploratory Postoperative Change Outcomes ^a

Item	MCID	AUC (95% CI)	Sensitivity	Specificity	Youden Index
Karlsson-Peterson	13.500	0.761 (0.659-0.852)	0.912	0.483	0.395
FAAM Sports	24.500	0.629 (0.510-0.739)	0.496	0.724	0.22
FAAM ADL	21.250	0.624 (0.502-0.743)	0.640	0.621	0.261
ALR-RSI	1.500	0.639 (0.519-0.756)	0.624	0.690	0.314
Tegner scale	1.583	0.781 (0.676-0.869)	0.920	0.552	0.472
VAS	2.500	0.518 (0.400-0.639)	0.672	0.448	0.12
Plantarflexion limitation, deg	5.500	0.496 (0.405-0.591)	0.024	1.000	0.024
Dorsiflexion limitation, deg	0.500	0.566 (0.493-0.629)	0.240	0.897	0.137

^aDefinitions of baseline-to-final outcomes and exploratory late postoperative change outcomes are described in the Statistical Analysis section. AUC >0.70 = good; AUC 0.60–0.70 = fair; AUC <0.60 = poor. ADL, Activities of Daily Living; ALR-RSI, Ankle Ligament Reconstruction–Return to Sport after Injury; AUC, area under the receiver operating characteristic curve; FAAM, Foot and Ankle Ability Measure; MCID, minimal clinically important difference; VAS, visual analog scale.

As shown in Table 6, the mean improvement in Karlsson-Peterson score from preoperative baseline to final follow-up was 17.51 points in group 1 and 22.39 points in group 2. The between-group mean difference was 4.88 points (95% CI, 3.13 to 6.63; P < .001), favoring group 2. This between-group difference was statistically significant but did not exceed the MCID threshold of 13.50 points.

Table 6.

Comparison of Between-Group Mean Changes With MCID Thresholds According to Change Interval ^a

Item	Group 1 Mean Change	Group 2 Mean Change	Mean Difference, G2−G1	95% CI	MCID Threshold	P
Karlsson-Peterson	17.51	22.39	4.88	3.13 to 6.63	13.50	<.001
FAAM Sports	22.20	26.32	4.12	1.10 to 7.14	24.50	.008
FAAM ADL	22.32	26.20	3.88	0.78 to 6.98	21.25	.02
Tegner activity scale	2.35	2.33	−0.02	−0.31 to 0.27	1.58	.89
VAS	3.09	3.18	0.09	−0.43 to 0.61	2.50	.74
ALR-RSI	1.73	2.14	0.41	−0.93 to 1.75	1.50	.55
Plantarflexion limitation, deg	1.24	0.32	−0.91	−1.37 to −0.46	5.50	<.001
Dorsiflexion limitation, deg	0.29	0.78	0.50	0.12 to 0.87	0.50	.01

^aFor Karlsson-Peterson, FAAM Sports, FAAM ADL, Tegner activity scale, and VAS pain, changes were calculated from preoperative baseline to final follow-up. For ALR-RSI and range of motion limitations, changes were calculated from the 12-month postoperative visit to final follow-up. For VAS pain and range of motion limitations, positive change indicates reduction in pain or limitation. MCID thresholds were used to contextualize clinical relevance rather than to establish treatment superiority. Group 1 indicates the intact syndesmosis group; group 2 indicates the syndesmotic instability group treated with additional suture button fixation. ADL, Activities of Daily Living; ALR-RSI, Ankle Ligament Reconstruction–Return to Sport after Injury; FAAM, Foot and Ankle Ability Measure; MCID, minimal clinically important difference; VAS, visual analog scale.

For FAAM Sports, the mean improvement was 22.20 points in group 1 and 26.32 points in group 2, with a between-group mean difference of 4.12 points (95% CI, 1.10 to 7.14; P = .008). For FAAM ADL, the mean improvement was 22.32 points in group 1 and 26.20 points in group 2, with a between-group mean difference of 3.88 points (95% CI, 0.78 to 6.98; P = .02). Although both differences were statistically significant, neither exceeded the corresponding MCID thresholds of 24.50 points for FAAM Sports and 21.25 points for FAAM ADL.

For the Tegner activity scale, the mean improvement was similar between groups (2.35 vs 2.33 points), with a between-group mean difference of −0.02 points (95% CI, −0.31 to 0.27; P = .89), which did not exceed the MCID threshold of 1.58 points. For VAS pain, the between-group mean difference in improvement was 0.09 points (95% CI, −0.43 to 0.61; P = .74), which did not exceed the MCID threshold of 2.50 points. For ALR-RSI, the between-group mean difference was 0.41 points (95% CI, −0.93 to 1.75; P = .55), which did not exceed the MCID threshold of 1.50 points.

For ROM outcomes, the between-group mean difference in plantarflexion limitation improvement was −0.91° (95% CI, −1.37° to −0.46°; P < .001), which was statistically significant but did not exceed the MCID threshold of 5.50°. The between-group mean difference in dorsiflexion limitation improvement was 0.50° (95% CI, 0.12° to 0.87°; P = .01), matching but not exceeding the MCID threshold of 0.50°. Given the low discriminatory performance of the dorsiflexion MCID threshold in Table 5, this finding should be interpreted cautiously.

Overall, Table 6 demonstrates that several between-group differences reached statistical significance; however, none clearly exceeded the corresponding MCID thresholds. These findings suggest that the observed statistical differences were small and may not represent clinically meaningful between-group superiority.

The proportion of patients achieving the MCID for FAAM Sports was significantly higher in group 2 (n = 74) than in group 1 (n = 80): 41 (55.4%) vs 29 (36.3%); risk difference, 19.2 percentage points (95% CI, 3.4 to 33.6; P = .03). Similarly, MCID achievement for FAAM ADL was more frequent in group 2 than in group 1: 53 (71.6%) vs 38 (47.5%); risk difference, 24.1 percentage points (95% CI, 8.6 to 38.0; P = .004). The most pronounced between-group difference was observed for the Karlsson-Peterson score, for which 73 of 74 patients in group 2 achieved the MCID compared with 56 of 80 patients in group 1: 98.6% vs 70.0%; risk difference, 28.6 percentage points (95% CI, 17.9 to 39.5; P < .001).

In contrast, there were no statistically significant between-group differences in MCID achievement for the Tegner activity scale: 85.1% vs 81.3%; risk difference, 3.9 percentage points (95% CI, −8.2 to 15.7; P = .67); VAS pain: 66.2% vs 63.8%; risk difference, 2.5 percentage points (95% CI, −12.4 to 17.1; P = .88), or ALR-RSI: 62.2% vs 51.3%; risk difference, 10.9 percentage points (95% CI, −4.7 to 25.7; P = .23). For ROM outcomes, MCID achievement for plantarflexion limitation was low in both groups and did not differ significantly: 0.0% vs 3.8%; risk difference, −3.8 percentage points; 95% CI, −10.5 to 1.8; P = .27). However, MCID achievement for dorsiflexion limitation was significantly higher in group 2 than in group 1 (31.1% vs 12.5%; risk difference, 18.6 percentage points; 95% CI, 5.6 to 31.1; P = .009). For ALR-RSI and ROM outcomes, responder status was based on exploratory late postoperative change from the 12-month visit to final follow-up rather than on baseline-to-final improvement (Table 7).

Table 7.

Responder Analysis (Proportion of Patients Achieving the MCID) ^a

Item	Group 1 (n = 80)	Group 2 (n = 74)	Risk Difference, Group 2−Group 1	95% CI	P
Karlsson-Peterson	56 (70.0)	73 (98.6)	28.6	17.9 to 39.5	<.001
FAAM Sports	29 (36.3)	41 (55.4)	19.2	3.4 to 33.6	.03
FAAM ADL	38 (47.5)	53 (71.6)	24.1	8.6 to 38.0	.004
Tegner activity scale	65 (81.3)	63 (85.1)	3.9	−8.2 to 15.7	.67
VAS	51 (63.8)	49 (66.2)	2.5	−12.4 to 17.1	.88
ALR-RSI	41 (51.3)	46 (62.2)	10.9	−4.7 to 25.7	.23
Plantarflexion limitation, deg	3 (3.8)	0 (0.0)	−3.8	−10.5 to 1.8	.27
Dorsiflexion limitation, deg	10 (12.5)	23 (31.1)	18.6	5.6 to 31.1	.009

^aData are presented as n (%). For Karlsson-Peterson, FAAM Sports, FAAM ADL, Tegner activity scale, and VAS pain, MCID achievement was defined as improvement from preoperative baseline to final follow-up meeting or exceeding of the corresponding MCID threshold. For ALR-RSI, plantarflexion limitation, and dorsiflexion limitation, MCID achievement was defined as exploratory late postoperative improvement from the 12-month visit to final follow-up meeting or exceeding of the corresponding MCID threshold. Improvement was defined as an increase in ALR-RSI and as a reduction in VAS pain or range of motion limitation. Risk differences are presented as percentage point differences, calculated as group 2–group 1, with 95% CIs estimated using the Newcombe-Wilson method. These exploratory ALR-RSI and range of motion responder analyses should be interpreted as late postoperative recovery analyses rather than baseline-to-final treatment effects. ADL, Activities of Daily Living; ALR-RSI, Ankle Ligament Reconstruction–Return to Sport after Injury; FAAM, Foot and Ankle Ability Measure; MCID, minimal clinically important difference; VAS, visual analog scale.

Correlation Between Dorsiflexion and Clinical Outcomes at Final Follow-up

As shown in Table 8, Spearman correlation analysis was performed to determine whether residual dorsiflexion limitation at final follow-up was associated with final clinical outcomes. No statistically significant correlations were observed between final dorsiflexion limitation and final patient-reported functional, psychological, activity-related, or pain outcomes.

Table 8.

Correlation Analysis Between Final Dorsiflexion Limitation and Clinical Outcomes at Final Follow-up in the Full Cohort (N = 154) ^a

Clinical Outcome	Spearman rho (rs)	95% CI	P
Final FAAM Sports	0.055	−0.104 to 0.211	.50
Final FAAM ADL	0.003	−0.155 to 0.161	.97
Final Karlsson-Peterson score	0.099	−0.060 to 0.253	.22
Final Tegner activity level	0.056	−0.103 to 0.213	.49
Final ALR-RSI	−0.048	−0.205 to 0.111	.55
Final VAS score	0.052	−0.107 to 0.208	.53

^aData are presented as Spearman rank correlation coefficients (r_s) with 95% CIs. Correlations were assessed between dorsiflexion limitation at final follow-up and final clinical outcome scores. A positive rs indicates that greater dorsiflexion limitation was associated with a higher outcome score, whereas a negative rs indicates an inverse association. Because higher VAS scores indicate greater pain, the direction of correlation for VAS should be interpreted accordingly. P values are 2-sided. ADL, Activities of Daily Living; ALR-RSI, Ankle Ligament Reconstruction–Return to Sport after Injury; FAAM, Foot and Ankle Ability Measure; VAS, visual analog scale.

Imaging Outcomes

Postoperative MRI was not obtained systematically at predefined time points in all patients. At least 1 postoperative MRI examination was available in 63 of 80 patients in group 1 (78.8%) and 65 of 74 patients in group 2 (87.8%), corresponding to 128 of 154 patients (83.1%) in the full cohort. Because the timing and frequency of postoperative MRI examinations were not standardized, imaging findings were analyzed descriptively only and were not used as formal longitudinal imaging outcomes. Figure 3 shows serial MRI findings from 1 representative patient after suture button fixation and is intended to illustrate postoperative imaging evolution rather than to establish a quantitative association between imaging changes and clinical outcomes.

Figure 3.

Representative serial postoperative coronal T2 fat-suppressed magnetic resonance imaging (MRI) findings after suture button fixation in a patient with distal tibiofibular syndesmosis instability concomitant with chronic lateral ankle instability. Images were obtained at 3 months, 6 months, 12 months, 2 years, 3 years, and 5 years postoperatively. Arrows indicate the suture button tunnel or attachment site, the distal tibiofibular syndesmotic joint effusion corresponding to the lambda sign, and the anterior talofibular ligament repair anchor or adjacent repair site, when visible on the corresponding image. The suture button label is shown only on images in which the implant tunnel or attachment site is visible. These images are presented as a representative case example only. Serial postoperative MRI was not obtained systematically in all patients at each time point, and these imaging findings were not used for formal quantitative correlation analysis with clinical outcomes. ATFL, anterior talofibular ligament.

In the subset of patients with available postoperative MRI after syndesmotic elastic fixation, localized bone marrow edema was noted around the tibiofibular button attachment sites. This observation may be related to stress concentration where the suture button construct passes through the medullary canal. The lambda sign appeared to persist after elastic fixation, with only minor fluctuations in the maximal height of syndesmotic joint effusion during imaging follow-up. Because postoperative MRI was not obtained systematically in all patients, these findings should be interpreted as descriptive imaging observations rather than definitive longitudinal imaging outcomes. Representative postoperative MRI findings were included for illustrative purposes only.

Complications

In the elastic fixation group, 32 patients reported discomfort due to palpable buttons over the thin medial and lateral ankle skin, primarily presenting as mild to moderate irritation during shoe or sock wear. Two of these patients expressed a desire for button removal because of more pronounced symptoms. In group 2, 1 patient experienced suture knot irritation with superficial exposure at the suture button incision site. Another patient, who underwent ATFL repair combined with syndesmotic elastic fixation, sustained a recurrent ankle sprain after a new injury and subsequently underwent ATFL reconstruction at an outside institution 3 years postoperatively.

Discussion

The major findings of this study were that patients with CLAI and concomitant DTSI achieved substantial midterm functional improvement after suture button fixation combined with lateral ligament surgery, with outcomes approaching those of patients with an intact syndesmosis. Compared with group 1, group 2 had significantly worse preoperative ankle function, as reflected by lower Karlsson-Peterson, FAAM Sports, and FAAM ADL scores (all P < .05). Despite this poorer baseline status, both groups improved significantly after surgery (all longitudinal P < .001). At 1 year, Karlsson-Peterson scores were comparable between groups (P > .05), and at final follow-up, group 2 showed a slightly higher Karlsson-Peterson score than group 1 (P = .02). Similarly, postoperative FAAM ADL, FAAM Sports, Tegner activity level, ALR-RSI, and VAS pain scores showed no clinically important disadvantage in the DTSI group at final follow-up. MCID responder analysis further showed higher proportions of clinically meaningful improvement in group 2 for FAAM Sports, FAAM ADL, and Karlsson-Peterson score (all P < .05). However, group 2 had a persistent but small dorsiflexion limitation at final follow-up (P < .001). Taken together, these findings indicate that patients with arthroscopically confirmed concomitant DTSI who were treated with additional suture button fixation achieved midterm patient-reported outcomes approaching those of patients with an intact syndesmosis, although a mild direction-specific dorsiflexion limitation may persist.

These findings suggest that concomitant DTSI imposes an additional functional burden before surgery. Concurrent injury to the lateral ligament complex and the syndesmosis may expose the ankle to a more complex multiplanar and rotational instability pattern, which can clinically manifest as persistent pain, swelling, subjective instability, and reduced athletic performance before surgical stabilization. The postoperative recovery observed in group 2 suggests that, when dynamic syndesmotic instability is identified arthroscopically and treated with additional suture button fixation during surgery for CLAI, midterm functional recovery can approach that observed in patients with an intact syndesmosis. At the same time, the residual dorsiflexion limitation highlights the need for careful fixation tensioning, postoperative ROM monitoring, and individualized rehabilitation.

Because lateral ankle instability can confound the evaluation of syndesmotic instability,^4,29,35 all patients underwent standardized lateral ligament surgery. However, because the study did not include an untreated DTSI group, the present design cannot isolate the independent therapeutic effect of suture button fixation. It should be emphasized that suture button fixation represents nonanatomic stabilization rather than true biological repair and therefore may not fully replicate the native biomechanics of the syndesmosis.^21,27 Postoperative outcomes were evaluated across 2 dimensions: immediate and long-term fixation effects. Clinically, the immediate stabilization was universally satisfactory without notable patient discomfort. More importantly, our results indicate that the long-term efficacy of retained suture button implants yielded no significant differences compared with the control group regarding postoperative activity levels, pain severity, and psychological readiness to return to sport. Collectively, these findings suggest that elastic fixation adequately restores syndesmotic stability, achieving functional outcomes that closely approximate those of patients with an intact syndesmosis. ¹

Notably, baseline analysis revealed that syndesmotic instability occurred more frequently in the female patient cohort, despite comparably high preinjury activity levels between groups. One possible explanation is that high-level athletic loading may interact with sex-related anatomical or biomechanical characteristics, including greater generalized joint laxity.^5,44 Differences in neuromuscular control during high-demand cutting, jumping, or pivoting activities may also increase susceptibility to dynamic syndesmotic instability. ⁸ However, the present study was not designed to determine the mechanism underlying this sex-related difference. This explanation remains speculative and should be interpreted as a hypothesis for future biomechanical studies rather than a confirmed mechanism.

To address potential confounding from baseline imbalances, we performed an expanded ANCOVA model for the final follow-up Karlsson-Peterson score. The group effect remained statistically significant after adjustment for preoperative Karlsson-Peterson score, sex, BMI, age, ATFL surgical modality, and follow-up duration, and this finding was consistent in a heteroscedasticity-robust HC3 sensitivity analysis. These adjusted findings support the robustness of the primary association but do not eliminate residual confounding. Because the 2 groups represented biologically distinct cohorts and were not randomized treatment arms, the adjusted association should not be interpreted as definitive evidence of an independent therapeutic effect of suture button fixation.

In assessing clinical significance, although several long-term functional outcomes demonstrated statistically significant between-group differences, the absolute mean differences did not consistently exceed the MCID thresholds. This finding cautions against interpreting statistical significance alone as evidence of clinically meaningful superiority at the cohort level. Responder analysis may provide an additional patient-centered perspective by evaluating the proportion of patients who achieved a clinically meaningful improvement. In the present study, a higher proportion of patients in group 2 attained the MCID for the Karlsson-Peterson score and FAAM ADL. This result is clinically reassuring but not unexpected, because these patients had arthroscopically confirmed syndesmotic instability that was specifically treated with suture button fixation, and they also had worse preoperative baseline function, allowing greater potential for measurable improvement. Importantly, because this study did not include a third group of patients with syndesmotic instability who were not treated with suture button fixation, the responder analysis should not be interpreted as definitive evidence of the independent superiority of suture button fixation. Rather, these findings suggest that when concomitant DTSI is recognized and stabilized during surgery for CLAI, patients with initially worse function can still achieve clinically meaningful improvement and midterm outcomes approaching those of patients with an intact syndesmosis.

Furthermore, our data suggest that concomitant DTSI may not be fully explained by the apparent severity of ATFL injury alone. Baseline comparisons demonstrated no significant difference between the 2 cohorts in the distribution of ATFL surgical modalities, including arthroscopic Broström-Gould repair, Broström-Gould repair with suture tape augmentation, isolated suture tape augmentation, and anatomic ATFL reconstruction (P = .56). This approach is supported by previous studies showing broadly comparable clinical outcomes among commonly used ATFL repair procedures, including arthroscopic or open repair and repair with or without suture tape augmentation.^6,24,25,42 This finding suggests that in patients with CLAI, concomitant syndesmotic instability may not be fully explained by the severity of lateral ligament injury alone. Even patients with repairable ATFL tissue may demonstrate dynamic syndesmotic instability during arthroscopic evaluation. Although this observation is consistent with the possibility that DTSI may reflect a partially distinct injury mechanism related to external rotation or dorsiflexion stress, our retrospective design does not allow definitive conclusions regarding injury causation. Therefore, rather than relying solely on preoperative imaging findings or the apparent severity of ATFL injury, surgeons should consider careful intraoperative assessment of syndesmotic stability when clinical suspicion for concomitant DTSI exists. ³⁵

In the present study, we observed that the elastic fixation group experienced a mild limitation in dorsiflexion both postoperatively and at the final follow-up. Existing literature suggests that variations in the insertion angle and tensioning of the suture button construct can alter the stress distribution and micromotion characteristics of the syndesmosis. However, the precise combination of angle and tension required to optimize physiological micromotion remains undefined.^14,16

We hypothesize that this nonanatomic elastic constraint could potentially limit the physiological widening of the syndesmotic clear space during ankle dorsiflexion, particularly as the wider anterior portion of the talar dome engages the ankle mortise. This restriction might also be influenced by overcompression applied during intraoperative tensioning, which could transiently limit native biomechanical micromotion within the syndesmosis. ² To explore whether this potential limitation has clinical relevance, we conducted a Spearman correlation analysis evaluating its relationship with patient-reported outcome measures and activity levels. Our quantitative data suggest that, although the construct may impose a mild dorsiflexion deficit, this effect falls within a clinically tolerable range. Importantly, this biomechanical rationale remains a plausible hypothesis rather than a proven mechanism. Future investigations using 3-dimensional gait analysis or intra-articular contact pressure testing are warranted to test these hypotheses and clarify their clinical relevance.

From a clinical perspective, implant-related local soft tissue irritation remains an important consideration after syndesmotic elastic fixation. In the present cohort, 32 patients in the elastic fixation group reported discomfort related to palpable buttons over the thin medial or lateral ankle soft tissue envelope, and 2 patients expressed a desire for implant removal because of more pronounced symptoms. These findings indicate that even when functional outcomes are satisfactory, local implant awareness may remain clinically relevant in a subset of patients.

In patients with available postoperative MRI, localized bone marrow or soft tissue signal changes were observed around the suture button attachment sites. These representative imaging findings suggest that local postoperative signal changes may persist around the suture button construct in some patients. However, because serial MRI was available only in a limited subset of patients and was not obtained according to a standardized imaging protocol, these observations should be interpreted cautiously. They should be considered hypothesis generating and cannot establish whether postoperative syndesmotic effusion, local bone marrow edema, or soft tissue signal changes contributed to pain, dorsiflexion limitation, psychological readiness, or implant-related discomfort. For patients whose function is persistently impaired by implant-related symptoms, local management or implant removal may be considered after confirming syndesmotic healing.

Limitations

Several limitations should be acknowledged. First, this was a retrospective, single-center cohort study, which may limit generalizability and introduce selection bias. The intact syndesmosis group provided a practical physiological benchmark rather than a randomized or untreated DTSI control group; therefore, the independent effect of suture button fixation cannot be definitively isolated. Second, the cohort consisted exclusively of patients with CLAI, and the findings may not be directly applicable to patients with isolated syndesmotic instability or acute syndesmotic injuries. In addition, the relatively high proportion of concomitant DTSI in the final analytic cohort should not be interpreted as reflecting population-level prevalence, because the study was conducted at a tertiary sports medicine referral center and included only eligible patients with complete follow-up data. Third, procedural heterogeneity remained because ATFL treatment was individualized according to tissue quality, generalized ligamentous laxity, revision status, and repairability. Although ATFL surgical modality did not differ significantly between groups, all procedures were performed by a single senior surgeon, and all patients followed a standardized rehabilitation protocol, the study was not powered for subgroup analyses according to lateral ligament technique. Therefore, the findings should be interpreted as reflecting a combined treatment strategy rather than the isolated effect of suture button fixation or any single ATFL surgical technique.

Conclusion

Our study demonstrated that, for patients with CLAI and arthroscopically confirmed concomitant DTSI, suture button fixation combined with lateral ligament surgery was associated with substantial midterm functional improvement, with outcomes approaching those of patients with an intact syndesmosis. However, because the comparison involved biologically distinct cohorts, these findings should be interpreted as supportive rather than definitive. Future matched or randomized studies are needed to clarify the independent contribution of suture button fixation.