**1. Introduction**

Valgus deformity of the ankle is a common manifestation of hereditary multiple exostoses (HME), affecting approximately half of the patients [1–4]. It may result in pain, restricted range of motion, and gait disturbances [5]. In a study investigating the natural history of untreated ankle joints in HME patients, 14 out of 75 ankles (19%) with an average tibiotalar tilt of valgus 9◦ showed early radiographic signs of osteoarthritis [6]. Appropriate surgical correction for excessive tibiotalar tilting has been recommended to preserve ankle function and prevent early arthritic changes in adulthood [7–9].

Past studies have suggested that the ankle valgus deformity is caused by a disproportionate shortening of the fibula relative to the tibia [2,4,6,10]. Compared to valgus deformity of the knee, which is primarily caused by changes in the distal femur and proximal tibia, ankle valgus deformity is directly affected by the disparity between the distal tibial and fibular length [1,6]. For this reason, predicting fibular shortening in HME patients may guide the indications for appropriate surgical treatment of ankle valgus.

Characteristic factors that have been associated with higher rates of ankle valgus deformities include an ankle valgus over 10◦ in males and exostoses on the proximal and distal tibia or fibula [10,11]. However, there is a lack of research analyzing the tibiofibular relationship, especially longitudinal observation studies investigating tibial and fibular growth. Furthermore, the literature is sparse in identifying factors related to the progression of ankle valgus and relative fibular shortening during growth periods.

The purpose of this study was to: (1) analyze longitudinal changes in the tibiofibular relationship of HME patients during their growing period; (2) identify factors associated with the progression of ankle valgus deformities; and (3) determine the effects of predisposing factors on the growth of the fibula and tibia over time.

#### **2. Materials and Methods**

#### *2.1. Patient Recruitment*

This retrospective study was approval by the institutional review board of the Severance Hospital, Seoul, Korea (28 November 2013; 4-2013-0700). Medical records and plain radiographs of patients who were skeletally immature at their first visit and underwent two or more full-length standing anteroposterior (AP) radiographs of the lower extremities during their follow-up were reviewed. Patients with a previous surgical history of the lower extremities and a limb-length discrepancy of over 2.5 cm, measured in any of the radiographs, were excluded. Among 53 patients who were initially enrolled, a total of 20 patients were excluded due to surgical history of their lower extremities (16 patients) and limb-length discrepancy of over 2.5 cm (4 patients). A total of 265 full-length radiographs of 63 affected legs in 33 patients were finally included in the study. The cases were divided into two groups of Group V and Group N, according to whether the tibiotalar angle increased into a more valgus position (Group V), or decreased into a more varus position (Group N) at the final evaluation when compared to the initial visit.

#### *2.2. Radiographic Measurements*

Full-length AP radiographs were taken in the standardized manner with the patients in a bipedal stance with equal weight bearing of both feet [12]. All radiographic parameters were measured using Picture Archiving and Communication Systems (PACS) (Centricity PACS 2.0; GE Medical Systems Information Technologies, Milwaukee, WI, USA). Radiographic parameters [1,2,4,10] were as follows: lateral distal femoral angles (LDFA), medial proximal tibial angles (MPTA), proximal and distal epiphyseal and physeal gaps in the tibia and fibula, fibula and tibia length, fibula/tibia length ratio, tibiotalar angles, locations of osteochondromas in the tibia or fibula, and the presence of tibiofibular synostosis.

Anatomical LDFA was defined as the lateral angle between the longitudinal axis of the femoral shaft and a line across the surface of the distal femoral epiphysis. Mechanical MPTA was defined as the medial angle between the longitudinal axis of the tibial shaft and a line across the tibial plateaus. The fibular length was measured as the straight-line distance from the tip of the proximal epiphysis to the tip of the lateral malleolus; the tibial length was measured as the straight-line distance from the tips of the tibial eminences to the tibial plafond. We also measured the distance between the tips of the tibial and fibular epiphyses, respectively. Proximally, due to the complex anatomy of the proximal tibial epiphysis, the distance was defined as the gap between a parallel line past the apex of the proximal

fibula and a parallel line through the midpoint between the line along the apex of the tibial spine and a line on the bottom-most portion of the condyle. Additionally, a parallel line through the midpoints between a line passing the top of the physis and a line crossing the bottom of physis were set as a reference line in order to measure the distance between the proximal and distal fibular and tibial physes (Figure 1A). The tibiotalar angle was defined as the medial angle between a perpendicular line to the axis of the tibia and the extended line that touches the articular surface of the talus (Figure 1B) [6]. In order to categorize the initial ankle alignment, the cases were divided into three positions: neutral (tibiotalar angle = 0–5◦ ), valgus (tibiotalar angle > 5 ◦ ), and varus (tibiotalar angle < 0 ◦ ) [13].

**Figure 1.** (**A**) The distance of proximal epiphyses (D) and physes (d) between tibia and fibula. The distance of proximal epiphyses was defined as the gap between the parallel line past the apex of the proximal fibula and the parallel line past the center of the distance between the line across the apex of the tibial spine and the most bottom of the condyle. The distances of proximal physes was defined as the gap between the parallel line passing through the midpoint of the distance between the line across the top of physis; the line crossing the bottom of physis was set as the reference line. (**B**) The measurement of the ankle joint. The tibiotalar angle was defined as the extension line of the tibial anatomical axis and a line perpendicular to the extension line that touches the dome of the talus. The left lower extremity indicated by the white arrow demonstrates a decreased distal tibiofibular distance compared to the right side, and a prominent ankle valgus deformity is also observed. (**C**) The tibiofibular synostosis. The white arrow indicates a definite bridge of an exostosis lesion connecting the proximal fibula and tibia without the overlapping cortex.

The presence of tibiofibular synostosis was determined when definite bridging by an exostosis lesion of the tibia and fibula, without overlapping cortices, was identified (Figure 1C). The location of osteochondromas were determined by dividing the entire length of the fibula and tibia, including both ends of the proximal and distal epiphysis, into two equal parts [10].

Reliability testing of radiographic measurements was performed. Two orthopedic surgeons independently measured the radiographic parameters, and inter-observer reliability of the two surgeons was determined using the intraclass correlation coefficient (ICC). Both surgeons also repeated measurements of the same subject at 2-week intervals to assure intra-observer reliability.

#### *2.3. Statistical Analyses*

The intra- and inter-observer reliabilities of the radiographic measurements were analyzed using the ICC. ICC was calculated using a two-way mixed effect for an absolute agreement between both values of each observer's measurements. An ICC value of 0.75 or greater was considered to reflect excellent reliability [14]. All statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA) and R statistics 3.6.2. *p* value < 0.05 was indicated as statistically significant.

Numerical variables were presented as mean ± standard deviation and frequency (percentage). Independent two-sample t tests were used to compare patient demographics and radiographic measurements between the groups of N and V. Categorical variables such as sex, the presence of tibiofibular synostosis, and location of osteochondroma were compared using a chi-square test (Fisher's

exact test). Mann–Whitney U-test was used to compare the follow-up duration between the groups. The linear mixed model, for a repeated-measures random intercept model, was used for LDFA, MPTA, proximal and distal epiphyseal and physeal gaps between the tibia and fibula, tibia and fibula length, fibula/tibia length ratio, and tibiotalar angle. The changes in the fibular length, tibial length, fibula/tibia ratio, and tibiotalar angle according to time were considered as covariate x time interactions. The fixed effects of associated factors on tibial and fibular lengths, and the tibiotalar angle per month which can affect the growth of lower extremities including patient's age, sex, synostosis and initial ankle alignment were analyzed.

#### **3. Results**

#### *3.1. Inter-and Intra-Observer Reliability*

The interobserver reliability for the radiographic measurements of the lower-limb alignment were found to be excellent (ICC; LDFA, 0.973; MPTA, 0.947; fibular length, 0.999; tibia; length, 0.999; proximal physis gap, 0.989; proximal epiphysis gap, 0.979; distal physis gap, 0.964; distal epiphysis, 0.953; tibiofibular angle, 0.993). The intraobserver reliability for the observer's repeated measurements also resulted in excellent results (ICC; LDFA, 0.946; MPTA, 0.968; fibular length, 0.999; tibia; length, 0.999; proximal physis gap, 0.995; proximal epiphysis gap, 0.990; distal physis gap, 0.985; distal epiphysis, 0.968; tibiofibular angle, 0.999).

#### *3.2. Demographic Distribution*

The demographic data of patients at their first visit are summarized in Table 1.


#### **Table 1.** Summary of the patients.

Values are presented as mean ± standard deviation or as numbers only. \* Statistical significance was noted. † The valgus angle is expressed as a positive value, and the varus angle as a negative value.

Age at the initial visit did not differ between the groups; however, the follow-up period for group V was significantly longer than that of group N (group N, 27.1 ± 14.5 months; group V, 40.5 ± 19.8 months, *p* = 0.004). Group V also had a significantly greater number of radiographs taken for each leg than group N (*p* = 0.003). The distribution of initial ankle alignment in both groups was similar. Neutral alignment was most commonly seen, followed by valgus and varus alignments. There were no significant differences in the location of exostosis (proximal tibia, *p* = 0.250; distal tibia, *p* = 0.176; proximal fibula, *p* = 0.262, and distal fibula, *p* > 0.999, respectively) and the presence of tibiofibular synostosis (*p* = 0.906) between the groups. The mean tibiofibular alignments at the initial and last

evaluation are displayed in Table 2. Based on the initial evaluation, no significant differences in any of the parameters were noted between the groups. However, the final evaluation revealed significant differences between the groups in all parameters, except for the distal physes gap.


**Table 2.** Measurements of lower limb alignment at initial visit and last follow-up.

Values are presented as mean ± standard deviation or as numbers only.\* Statistical significance was noted. † The valgus angle is expressed as a positive value, and the varus angle as a negative value.

#### *3.3. Change in Alignment During the Observation Period*

The estimated slopes of change between the initial and final measurements were analyzed using a linear mixed model. Group N had a significant constant increase in the proximal physes gap (0.062, *p* < 0.001) and the fibular (0.977, *p* < 0.001) and tibial (0.941; *p* < 0.001) lengths. Group V had a significant increase in the tibia and fibular length, the proximal physes and epiphyses gap, and the tibiotalar angle (Figures 2 and 3), but a decrease in the LDFA, the fibula/tibia ratio, and the distal physes and epiphyses gap. In terms of the differences between the slopes of both groups, significant differences were identified in the tibial length, the fibular/tibial length ratio, the proximal and distal epiphyses gap, and the tibiotalar angle (Table 3).

**Figure 2.** The spaghetti plot and estimated slope in the tibial, fibular length and fibula/tibia ratio of groups N and V. The solid line indicates the estimated slope of group V. and triple lines indicate the estimated slope of group N. No significant difference was noted in fibular length, and there was a significant difference in slope between group N and group V in tibial length and fibula/tibia ratio.

*Diagnostics* **2020**, *10*, 752

**Figure 3.** The spaghetti plot and estimated slopes in proximal and distal tibiofibular gaps of groups N and V. The solid line indicates the estimated slope of group V; and triple lines indicate the estimated slope of group N. Progressive fibular shortening with a significant increase in the proximal gap and a decrease in the distal tibiofibular gap were identified over time.


**Table 3.**Estimated slope of change between the initial and final measurements.

Values are presented as mean (SE, standard error) or as numbers only. \* Statistical significance was noted. † The valgus angle is expressed as a positive value, and the varus angle as a negative value.
