This cross-sectional, observational study included patients who underwent ankle fracture surgery. The ethics committee of our institute approved this study, which was conducted in accordance with the principles of the Declaration of Helsinki. All participants provided written informed consent before participation.

The study was conducted between July 2022 and November 2024 and enrolled patients with ankle fractures admitted to our hospital. The inclusion criteria were ankle fractures treated with open surgery and physical therapy. All patients underwent open reduction and internal fixation (ORIF) and were immobilized with a splint for at least 1 week postoperatively. The exclusion criteria included multiple fractures, open fractures, postoperative complications, such as infection or deep vein thrombosis, history of neurological and orthopedic diseases, refusal to undergo postoperative measurements, and visits to another hospital. Participants were assessed 3 months after ORIF. All measurements were performed by the same physiotherapist. All participants used crutches for at least the first 3 weeks and continued a training program for joint mobility, muscle strength, and functional skills, such as walking and stair climbing, at least once a week for 3 months. The data on participants' age, sex, height, number of fractures (12, 13), and Lauge–Hansen classification (14) were collected from medical records. Participants' weights were measured using a digital scale, and their BMIs were calculated.

The SAFE-Q was used to measure ankle QOL. It consists of 43 questions across 6 subscales: 9 questions on pain and pain-related (Q1–Q7, Q10, and Q11), 11 on physical functioning and daily living (Q12–Q22), 6 on social functioning (Q23–Q28), 3 on shoe-related (Q8–Q9, Q34), 5 on general health and well-being (Q29–Q33), and 9 assessing sports activity (Q35–Q43). The patients answered each question on a Likert scale (4, 3, 2, 1, or 0). Questions 3 and 43 were scored using a visual analog scale, with the formula (10−value) × 0.4 used to calculate the score. The score for each subscale was calculated as follows (11, 15):(∑ofsubscale×25/numberofsubscalequestions)/100,with higher subscale scores indicating higher QOL (16).

Each participant completed the questionnaire, and the examiners collected the data. In this study, 4 subscales were tabulated and scored on a maximum 100-point scale: pain and pain-related, physical functioning and daily living, social functioning, and general health and well-being.

Non-weight-bearing ankle ROM was measured in 1° increments using a goniometer during passive movement. With the participants in the supine position, ankle dorsiflexion ROM was measured with the knee in extended and flexed positions, and plantarflexion ROM was measured with the knee in flexed position. To measure passive ROM, the examiner manually dorsiflexed and plantarflexed the participant's ankle maximally. The examiner set the axis perpendicular to the fibula as the cardinal axis and the plantar surface of the foot as the axis of movement, and the angle between these axes was measured as ankle dorsiflexion and plantarflexion ROMs.

Weight-bearing ankle dorsiflexion ROM was measured using four methods: measuring the rear ankle with the knee extended (17) and flexed during forward lunge (18), measuring the front ankle during forward lunge (19), and measuring the ankle during deep squat sitting (20) (Figure 1). For the rear ankle measurements, the participants were instructed to step forward with the leg on the non-measurement side and to lean their lower leg forward as far as possible with the knee on the measurement side extended or flexed (Figures 1A,B). For the front ankle measurement, the participants were instructed to step forward with the lower leg being measured and lean their lower leg as far forward as possible in a forward lunge position against the wall (Figure 1C). The examiner instructed the participants to place their upper limbs against the wall for balance. Deep squat sitting measurements were performed as follows: participants were instructed to squat to the deepest position they could maintain for 3 s and then lean their lower leg forward as far as possible. They were instructed to stand with their feet shoulder-width apart, eyes looking straight ahead, and arms extended in front, parallel to the floor (Figure 1D). The participants were instructed to keep their heels on the ground during the tasks. The weight-bearing ankle dorsiflexion angle was measured using a goniometer to determine the angle between a perpendicular line to the floor and a line connecting the fibular head and lateral malleolus with a minimum value of 1°. The measurements were repeated twice for each method to verify reliability.

Ankle strength was measured using a Biodex 3 dynamometer (Biodex Medical Systems, Shirley, NY, USA) to determine the ankle plantar/dorsiflexion muscles. The participants' knees were bent 30° while seated. Straps were used to stabilize the lower trunk, thigh, and ankle muscles. Ankle plantar/dorsiflexion measurements were made bilaterally isokinetically (concentric/concentric), and 2 sets of 5 maximal dynamic repeats were performed at an angular velocity of 60 °/s, separated by 30 s (20). Participants were positioned with their feet parallel to the floor to prevent hamstring strain. Finally, the peak torque/body weight ratio was computed after the torque was measured at a minimum of 1 Nm.

Gait parameters were measured using a motion analyzer (MA-3000; Anima Corporation, Tokyo, Japan). Ten cameras were used with the motion analyzer, and data were acquired at 100 Hz. Each data point was low-pass filtered at a frequency of 10 Hz. The reflex marker sticking sites were the bilateral anterior superior iliac spine, greater trochanter, lateral femoral epicondyle, lateral malleolus, head of the fifth metatarsal bone, and midpoints of the left and right superior posterior iliac spines. Participants walked barefoot on a 6-m gait path at a self-selected speed. During walking, none of the participants required any walking aid. The gait parameters measured were the gait speed, cadence, step length, and plantar dorsiflexion angle of the ankle during the stance phase. In addition, sagittal plane knee angles were analyzed, with particular attention to the knee extension angle during the terminal stance phase, in order to evaluate potential compensatory movement patterns caused by limited ankle dorsiflexion. Gait was performed eight times, and the gait parameters were averaged using the motion analyzer's built-in software and analyzed.

G*Power 3.1 (Heinrich Hain University, Dusseldorf, Germany) (21) was used to calculate the required sample size (effect size f² = 0.35, alpha error = 0.05, power = 0.80), and the result was 31. This study included 32 participants. The Shapiro–Wilk test was used to assess the data distribution. Means and standard deviations were calculated for normally distributed data, and medians and interquartile ranges were calculated for non-normally distributed data. The intraclass correlation coefficient (ICC) and Bland–Altman plot were used to evaluate the reliability of the weight-bearing ankle dorsiflexion ROM measurements. Using the first and second measurements, we calculated the ICC (1,1) and performed a Bland–Altman analysis. To evaluate the intrarater reliability, the ICC (1,1), standard error of measurement (SEM) (22), 95% confidence interval of the minimal detectable change (MDC₉₅) (23), and relative repeatability (RR) (24) were calculated. The SEM, MDC₉₅, and RR were calculated according to previous studies using the following formula: SEM = standard deviation × √(1−ICC), MDC₉₅ = 1.96 × SEM × √2, and RR = MDC₉₅/mean, respectively. Pearson's product–moment and Spearman's rank correlation coefficients were calculated to determine the correlation between the four SAFE-Q subscales and the measurement items after surgery, and a heat map was created. Finally, to clarify the factors related to postoperative QOL, a multiple regression analysis using the stepwise method was performed. The four subscales of the SAFE-Q (pain and pain-related, physical functioning and daily living, social functioning, and general health and well-being) were used as dependent variables, while the measurement items along with age, BMI, and the number of fractured malleoli were entered as independent variables.

Statistical significance was set at P < 0.05. All statistical analyses were performed using SPSS Statistics version 30 (IBM Corp., Armonk, NY, USA).

Overall, 43 patients underwent ORIF during the study period. Of these, 11 were excluded (multiple fractures, n = 2; open fractures, n = 1; postoperative infection, n = 1; history of neurological disease, n = 1; refused to undergo postoperative measurements, n = 2; went to other hospitals, n = 4). Finally, 32 participants met the eligibility criteria and were included in this study. The participants' characteristics are summarized in Table 1. The median age of the participants was 48.5 (interquartile range, 23.5–61.5) years. The mean duration of fixation and crutch use after ORIF were 19.0 ± 12.3 and 49.3 ± 14.3 days, respectively. The mean time from ORIF to measurement was 92.0 ± 5.6 days.

The measured values of the participants are summarized in Table 2. The SAFE-Q scores of the four subscales (pain and pain-related, physical functioning and daily living, social functioning, and general health and well-being) were 75.6 ± 14.2, 74.1 ± 19.4, 69.9 ± 28.8, and 77.2 ± 23.2 points, respectively. The mean ankle dorsiflexion ROMs measured on the rear ankle with the knee extended during forward lunge, rear ankle with the knee flexed, front ankle during a forward lunge, and ankle during deep squat sitting were 15.6 ± 2.4, 18.3 ± 2.9, 21.8 ± 4.4, and 23.7 ± 5.0 degrees, respectively. Additionally, the knee extension angle during terminal stance phase was −3.1 ± 4.4 degrees.

The intrarater reliability values for weight-bearing ankle dorsiflexion ROM are summarized in Table 3. None of the methods showed fixed or proportional bias, and there was good agreement between the first and second measurements. The ICC (1,1) values for ankle dorsiflexion ROM measured on the rear ankle with the knee extended during a forward lunge, rear ankle with the knee flexed during a forward lunge, front ankle during a forward lunge, and ankle during deep squat sitting were 0.961, 0.959, 0.984, and 0.986, respectively. The SEMs were 0.49, 0.61, 0.54, and 0.58 degrees; the MDC₉₅ values were 1.36, 1.68, 1.50, and 1.60 degrees; and the RRs were 0.09, 0.09, 0.07, and 0.07, respectively.

A heat map of the correlation coefficients between the SAFE-Q subscales and measured values is shown in Figure 2. Weight-bearing ankle dorsiflexion ROM during deep squat sitting positively correlated with the four subscales: pain and pain-related (r = 0.58; P < 0.001); physical functioning and daily living (r = 0.69; P < 0.001); social functioning (r = 0.61; P < 0.001); general health and well-being (r = 0.48; P = 0.005). Gait speed was also positively correlated with four subscales: pain and pain-related (r = 0.52; P = 0.002); physical functioning and daily living (r = 0.77; P < 0.001); social functioning (r = 0.69; P < 0.001); general health and well-being (r = 0.41; P = 0.002). In contrast, the knee extension angle during the terminal stance phase showed no significant correlation with any SAFE-Q subscales (P > 0.05).

The results of the multiple regression analysis are presented in Table 4. Weight-bearing ankle dorsiflexion ROM during deep squat sitting was selected as the independent variable for pain and pain-related [standardized partial regression coefficient (β) = 0.584, P < 0.001], physical functioning and daily living (β = 0.376; P = 0.006), social functioning (β = 0.317; P = 0.045), and general health and well-being (β = 0.483; P = 0.005). Furthermore, gait speed was selected as the independent variable for physical functioning and daily living (β = 0.555; P < 0.001) and social functioning (β = 0.514; P = 0.002). The adjusted coefficients of determination (R²) for pain and pain-related, physical functioning and daily living, social functioning, and general health and well-being as dependent variables were 0.319, 0.662, 0.517, and 0.207, respectively; the Durbin-Watson ratios were 1.666, 1.459, 2.143, and 1.640, respectively; the residuals were normally distributed, with P = 0.12, P = 0.10, P = 0.81, and P = 0.19, respectively.