Clinical Outcomes Following Total Ankle Arthroplasty without Concomitant Osteotomy in Ankles with Severe Preoperative Varus Deformity: Comparison to Ankles with Preoperative Neutral Alignment

Takahiro Yamashita Department of Orthopaedic Surgery, Kobe University Graduate School of Medicine https://orcid.org/0000-0002-4987-0440 Kanto Nagai Department of Orthopaedic Surgery, Kobe University Graduate School of Medicine Noriyuki Kanzaki (  kanzaki@med.kobe-u.ac.jp ) Tetsuya Yamamoto Department of Orthopaedic Surgery, Kobe University Graduate School of Medicine Kazuyuki Ibaraki Department of Orthopaedic Surgery, Kobe University Graduate School of Medicine Daisuke Araki Department of Orthopaedic Surgery, Kobe University Graduate School of Medicine Yuichi Hoshino Department of Orthopaedic Surgery, Kobe University Graduate School of Medicine Takehiko Matsushita Department of Orthopaedic Surgery, Kobe University Graduate School of Medicine Ryosuke Kuroda Department of Orthopaedic Surgery, Kobe University Graduate School of Medicine


Background
Modern total ankle arthroplasty (TAA) has been shown to have satisfactory outcomes with better function and equivalent pain relief compared to ankle arthrodesis in patients with ankle osteoarthritis (OA) or rheumatoid arthritis (RA) (1)(2)(3)(4). This could be attributable to many factors, including improvement of prosthetic design, development of surgical technique, achievement of soft tissue balancing, careful patient selection, and severity of preoperative deformity (5,6). However, experts are yet to arrive at a consensus on whether preoperative coronal deformity is associated with postoperative poor outcomes. Some authors reported that preoperative coronal malalignment was associated with poor outcomes (7,8) and others reported no association between preoperative coronal varus deformity and postoperative clinical outcomes (2,(9)(10)(11)(12). As a limitation, most previous studies performed simultaneous osteotomy at the time of TAA for ankles with severe varus coronal deformity.
The TNK ankle prosthesis, which was originally developed by Takakura et al. (13), has a medial wall in its tibial insert (Fig. 1). This shape of the tibial insert is unique and this may stabilize the medial part of the ankle more than other prostheses; however, to date, only a few studies have reported the clinical outcomes following TAA using the TNK ankle prosthesis (13)(14)(15). Thus, it remains unclear whether preoperative coronal varus malalignment is associated with postoperative poor outcomes following TAA using the TNK ankle prosthesis. Therefore, the purpose of the present study was to investigate the postoperative outcomes following TAA using the TNK ankle prothesis, and to compare the outcomes between a neutral group (preoperative coronal alignment < 9°) and a varus group (preoperative coronal alignment > 10°). It was hypothesized that both groups would show signi cant improvement following TAA, and that the neutral group would have better clinical outcomes than the varus group.

Methods
The present study was a retrospective analysis of prospectively collected data of TAA performed in our institution. Fifty-three arthritis ankles of 51 patients (age at the surgery: 71.4 ± 5.6 years; 16 males, 35 females) who underwent TAA were included in this study. The average follow-up period after the surgery was 36.8 ± 17.8 months. This analysis included 48 ankles with osteoarthritis and 5 ankles with rheumatoid arthritis. Inclusion criteria were patients who underwent TAA from May 2012 to November 2018 by a single surgeon. Exclusion criteria were patients who received less than 1-year follow-up after the surgery.
The study was conducted in accordance with the Declaration of Helsinki and the Guidelines for Good Clinical Practice. The institutional review board approved this study (No.180053), and informed consent was obtained from all the patients. The ankles were allocated into two groups according to the preoperative talar tilt (TT) angle (°) (16): the control group (< 9°; 37 ankles) and the varus group (> 10°; 16 ankles). The demographic data of the two groups are shown in Table 1. Two patients underwent bilateral TAA at different times. One patient had neutral alignment in both sides of the ankles, and one patient had varus alignment of > 10° in both sides. These 4 ankles were allocated to the control and varus groups accordingly and were evaluated independently. The third-generation TNK ankle prosthesis (Kyocera medical, Kyoto, Japan) was used for all patients. The TNK ankle was originally developed by Takakura et al (13,14). The third-generation TNK ankle is made of alumina-ceramic and comprises tibial and talar components with a polyethylene insert, which has a medial wall (Fig. 1). This insert is attached to the tibial component; thus, the TNK ankle is so-called a "2component type." All procedures were performed through an anterior approach between the extensor hallucis longus and tibialis anterior tendon to the ankle joint. Capsulectomy was performed and all osteophytes were removed piece by piece. In cases where the inner tightness was marked at the medial part of the ankle, the deep layer of the deltoid ligament was dissected and elongation of the super cial layer was performed.
Thereafter, tibial cutting was performed based on the preoperative planning, following talar cutting. The talar component was xed with cement, and the tibial component was xed with a small cancellous screw (DePuy Synthes, West Chester, PA, USA). Small beads of cement were initially placed on the surface of the tibial component facing the tibia and bone marrow was placed on its surface after the application of a calcium phosphate paste to induce osseous bonding (HOYA, Tokyo, Japan).
Simultaneous concomitant procedures during TAA were documented.

Postoperative rehabilitation
The same postoperative rehabilitation protocol was utilized in both groups. The short leg splint was applied and weight-bearing was not permitted for the rst 2 weeks after the surgery. Range of motion (ROM) exercises started 3 days after the surgery only during physical therapy. From 2 weeks after the surgery, a soft brace with a stay was applied, and full weight-bearing was allowed. ROM exercises were performed without limitations 2 weeks after the surgery. The stay of the brace was removed 6 weeks after the surgery and the brace was applied without the stay until 3 months after the surgery.

Clinical outcomes
The Japanese Society for Surgery of the Foot (JSSF) ankle-hindfoot scale (17,18) was used for objective evaluation, and the Self-Administered Foot Evaluation Questionnaire (SAFE-Q) (19) were used for subjective assessment prior to surgery and at the nal follow-up. The dorsi-exion and plantar-exion angles of the ankle were measured using a goniometer prior to the surgery and at the nal follow-up. Concomitant procedures and complications related to the surgery were documented.

Radiographic evaluation
Prior to the surgery and at the nal follow-up, the tibial anterior surface (TAS) angle, the tibial axis-medial malleolus (TMM) angle, the TT angle, and the tibial lateral surface (TLS) angle were measured using plain radiographs of the ankle under weight-bearing condition, as previously described (20).

Statistical analysis
All measurements were expressed as mean ± standard deviation (SD). Data analyses were performed using Excel Statistics version 7 (ESUMI Co., Ltd, Tokyo, Japan). A paired t test was performed to explore pre-and postoperative differences in clinical scores and radiographic parameters in each group. Mann-Whitney U test was performed to explore the differences in clinical scores and radiographic parameters between the neutral group and the varus group. Chi-square test was used to explore the differences in concomitant procedures and complications between the groups. Signi cance level was set as P < 0.05.
A prior power analysis showed that at least 31 cases and 15 cases were required to determine signi cant difference between two independent means with two-tails, an effect size of 0.9, a power of 0.80, and an alpha error of 0.5.

Results
Postoperative JSSF and SAFE-Q scores were signi cantly greater than the preoperative scores in each group ( Table 2, 3). In terms of the JSSF scale, there were no signi cant differences between the groups, except in preoperative alignment (Table 2). Regarding the SAFE-Q, there were no signi cant differences in subscale scores of pain, shoes, and preoperative health between the groups, while subscale scores of function, social, and postoperative health were greater in the varus group than in the neutral group (Table 3). No signi cant differences were noted in the improvement of these subscales between the groups.  Regarding the ROM of the ankle, there were no signi cant differences in dorsal and plantar exions between the groups pre-vs postoperatively (Table 4) Concerning the radiographic parameters, there were no signi cant differences between the groups at the nal follow-up (Table 5).   Table 6, and deltoid ligament release was more often performed in the varus group than in the neutral group. Complications are summarized in Table 7; no signi cant differences in the rate of complications were observed between the groups. Moderate talar component subsidence occurred in two ankles in each group, and total talar replacement was performed in those cases using total talar prostheses.

Discussion
The main ndings of the present study were 1) signi cant improvements in clinical outcomes and radiographic outcomes occurred in each group following TAA using the TNK ankle prosthesis and 2) the outcomes in the varus group were similar to those of the neutral group. These ndings suggest that the postoperative outcomes following TAA were favorable even in the patients who had preoperative varus coronal malalignment greater than 10°. Importantly, neutral coronal alignment was achieved only by performing soft tissue balancing without osteotomies even in the varus group. The tibial insert of the TNK ankle prosthesis has a medial wall and this could stabilize the mortise. Therefore, it could be assumed that the satisfactory outcomes with the TNK ankle prosthesis are partially attributable to this prosthetic design and proper soft tissue balancing. Thus, TAA using the TNK ankle prosthesis can be indicated for ankles with varus coronal malalignment.
Previous studies have shown that preoperative coronal malalignment is considered a risk factor for early prosthetic failure (6-8). Therefore, it is considered essential to establish a neutrally aligned ankle after TAA for successful long-term outcomes and to prevent complications (4,11,21). Recently, several studies have shown that satisfactory outcomes can be obtained following TAA even in ankles with substantial preoperative coronal deformity (4,(9)(10)(11)(12). The key to achieving postoperative neutral alignment seems to be associated with additional procedures at the time of TAA. Similar to previous studies, in the present study, additional procedures were performed at the time of TAA to obtain appropriate soft tissue balancing and neutral alignment, resulting in comparable outcomes with those in previous reports. The present ndings support the suggestion that preoperative deformity does not diminish clinical outcomes if a well-aligned stable ankle is obtained after TAA (4,11).
No concomitant osteotomies were performed in the present study even in the varus group; however, the postoperative TT and TMM angles were improved in both groups and were not signi cantly different between the neutral and coronal groups. This nding suggests that postoperative neutral coronal alignment was achieved by performing proper soft tissue balancing, such as deltoid ligament release, alone. The advantages of not performing concomitant osteotomy include less invasive surgery for patients and weight-bearing can be allowed soon after the surgery.
With regard to SAFE-Q, the varus group had slightly better postoperative outcomes in the subscales of social-functioning and health than the neutral group, contrary to the hypothesis (Table 3). However, the level of improvement was not signi cantly different between the groups. Additionally, other possible reasons may be that the number of patients was lower in the varus group than in the neutral group, and that the follow-up period was slightly shorter in the varus group.
Outcomes after TAA using the TNK ankle prosthesis have been shown to be satisfactory in several reports (13)(14)(15); however, these reports did not investigate the relationship between preoperative coronal deformity and postoperative outcomes. In 1990, Takakura et al. reported that postoperativeoutcomes were satisfactory in 67% of ceramic prostheses without cement (13). In 2004, they reported satisfactory outcomes with 70 ankles (39 OA ankles and 31 RA ankles) using the current third-generation TNK ankle prosthesis (14). The outcomes are in agreement with the report by Nagashima et al. (15) in which favorable outcomes were noted in 21 RA ankles using the same prosthesis. These previous studies support the present ndings that outcomes following TAA using the TNK ankle prosthesis were satisfactory independent of preoperative coronal deformity.
Several limitations of the present study should be noted. First, the number of patients was relatively small; however, a prior power analysis showed that the number of patients in the present study could be powered to detect signi cant differences. Second, the follow-up period in the current study was an average of 3 years after the surgery, which is relatively short. Thus, long-term outcomes of the present cases should be investigated in the future.

Conclusions
Outcomes after TAA using the TNK ankle prosthesis were favorable in patients with severe varus ankle alignment and in those with neutral ankle alignment without concomitant osteotomy. Satisfactory outcomes could be achieved in patients with severe varus ankle alignment after TAA using the TNK ankle prosthesis without concomitant osteotomy.  Figure 1 TheTNK ankle prosthesis. Total ankle arthroplasty using the TNK ankle prosthesis were performed via anterior approach. The insert has a medial wall (white circle).