Patient Demographics and Study Characteristics
Between November 2003 and October 2007, two of the authors (P.F.R. and M.H.) performed 240 consecutive primary total ankle arthroplasties with the Mobility implant in 233 patients (115 women and 118 men) with a mean age (and standard deviation) of 61.6 ± 13.1 years (range, twenty-four to eighty-six years). The diagnosis was posttraumatic ankle arthritis in 123 ankles (51.3%), primary ankle osteoarthritis in seventy-four (30.8%), rheumatoid arthritis in thirty-six (15.0%), and hemochromatosis in seven ankles (2.9%). Patients were not considered for total ankle arthroplasty if they were diagnosed with insulin-dependent diabetes mellitus, peripheral artery disease, or severe hindfoot deformity. Clinical and radiographic follow-up visits were scheduled at six weeks, six months, and one year postoperatively, and yearly thereafter. Two hundred and thirty-three (97.1%) of the 240 ankles were available for follow-up at one year. One patient died four months after surgery (of a cause unrelated to the procedure), and the remaining six patients were living in foreign countries and could not return for evaluation. The latter six patients (all with unilateral arthroplasties) were contacted by telephone, and all six reported that the ankle was functioning well and had not undergone a reoperation. The data on these six cases were not included in the results. The mean duration of follow-up was 32.8 ± 15.3 months (range, twelve to sixty-three months). The study was approved by the local ethical committee, and all patients provided written informed consent.
Design Rationale
The Mobility prosthesis is an unconstrained three-component system consisting of a tibial component, a talar component, and a mobile-bearing, highly cross-linked polyethylene inlay. The backsides of the tibial and the talar component are porous-coated titanium surfaces designed to provide osseous ingrowth after press-fit implantation. The talar component is designed to leave the malleolar surfaces intact. One advantage of retaining the articular surfaces of the malleoli, in addition to providing more physiological ankle biomechanics, is that the intact medial and lateral cortices of the talar dome provide improved support for the talar component, reducing the risk of secondary subsidence or migration. In the case of a failure, revision to an ankle fusion is facilitated, the amount of limb shortening resulting from the fusion is minimized, and the bone graft used to fill the central defect is not required to provide as much support. The doubly curved shape of the articular surface of the talar component stabilizes the polyethylene insert and enables articular congruency during eversion and inversion. This shape also increases stability if the talar component has not been implanted perfectly parallel to the ground. Two pegs on the backside of the talar component increase the initial stability and create additional surface area for bone ingrowth. The interface between the implant and the bone can be visualized from the side during implantation, allowing intraoperative confirmation that the component is fully seated on the talar dome. The talar component is available in six sizes. Sizes 1 through 4 have the same "footprint" and use the same talar bone cuts, and the same is true for sizes 5 and 6. Thus, except in the case of a switch between size 4 and size 5, no additional bone cut is necessary if it is determined intraoperatively that the planned talar component should be increased or decreased by one size in order to provide optimal talar support.
The tibial component has a flat articular surface and a stem on the tibial side. The stem allows for final rotational adjustment after placement of the tibial component through an anterior cortical window; additional stability can then be achieved by impacting cancellous bone around the stem if desired. The plafond is relatively long in the anteroposterior direction to allow for optimal support of the distal tibial cortex. The posterior aspect of the plafond is narrower and rounded in order to minimize posterior impingement with the fibula or the medial soft tissues. The surface of the tibial side of the polyethylene insert is smaller than the articular surface of the tibial component in order to avoid protrusion of the polyethylene insert medially or laterally. The instruments used for implantation allow centering of the tibial and the talar component with respect to each other in both the sagittal and the frontal plane.
Surgical Technique
The surgical procedure was performed in an operating room with laminar air flow. The patient was placed in a supine position, and a pneumatic tourniquet was applied. Regional anesthesia and a single preoperative dose of a second-generation cephalosporin were administered. An ischiofemoral nerve block was routinely performed for postoperative pain control. An anterior midline incision allowed a direct approach between the tibialis anterior and the extensor hallucis longus tendon to visualize the joint from the medial malleolus to the lateral malleolus. Anterior osteophytes were resected. The distal tibial resection guide was then installed to facilitate a distal tibial resection that was oriented perpendicular to the weight-bearing axis in the frontal plane and had a posterior slope of 6° in the sagittal plane. Sufficient bone was resected to accommodate a 5-mm inlay. The appropriate size of the tibial component was identified, and the corresponding guide jig was used to mark and open the anterior tibial window. The guide was centered exactly over the talus in the frontal plane and followed the orientation (rotation) of the talar body. The bone block was extracted, and cancellous bone was impacted to the correct depth to complete the tibial preparation.
The stepwise preparation of the talus was begun by transferring the center of the tibial component to the talus; the position of the talar component (frontal position and rotational orientation) is determined by a linked system that allows a central drill hole to be placed in the talus. The talar centering device was used to identify the correct position of the central drill hole in the sagittal plane. Since the talar cuts for component sizes 1 through 4 and the cuts for sizes 5 and 6 would be performed with use of different sets of drill guides and jigs, the size of the talar component had to be determined at this stage. The talar preparation was completed by performing a posterior and an anterior oblique resection and then contouring the resected talar surface to match the two pegs of the talar component. Trial components were inserted, and the correct fit and alignment were verified with use of fluoroscopy before the definitive components were implanted. The porous-coated surfaces of the components usually provided excellent primary press-fit stability. Stability of the tibial component could be increased, if required, by impacting cancellous bone chips around the tibial stem before reinserting the bone block to close the anterior tibial window. We did not use bone cement. The wound was closed in layers, and a well-padded cast was applied with 5° of ankle dorsiflexion.
A total of 171 concomitant procedures were performed in 124 ankles (51.7%) (Table I). Most of the concomitant procedures, including osteotomies to achieve precise correction of hindfoot alignment, were planned preoperatively. We did not obtain hindfoot-view radiographs routinely. Alignment was assessed clinically both before surgery and intraoperatively. We accepted slight valgus alignment of <5° if the deformity was localized below the ankle joint line. Varus alignment or a valgus deformity located at or above the ankle joint line was not accepted. Percutaneous gastrocnemius or Achilles tendon lengthening was performed after implantation if ankle dorsiflexion was <5°.
Postoperative Management
The cast was changed, if required, before the patient was discharged. The length of the hospital stay was influenced primarily by the health care system, averaging five days for the series as a whole but longer in Switzerland than in the United States. The cast and skin sutures were removed two to three weeks postoperatively, and use of a removable walker boot and passive and active mobilization of the ankle were initiated under the supervision of a proficient physiotherapist. Patients were limited to partial weight-bearing with use of crutches for six weeks postoperatively, during which time all patients received low-molecular-weight heparin for thromboembolic prophylaxis. Full weight-bearing, initially in the removable walker, was allowed after the clinical and radiographic evaluation six weeks after surgery, and therapy was adapted to include strengthening exercises and training involving proprioception and coordination. Patients with a concomitant corrective osteotomy or fusion underwent the same postoperative protocol except that return to full weight-bearing was carried out in a stepwise fashion during the two weeks following removal of the walker.
Clinical Evaluation
Clinical results were assessed with use of the American Orthopaedic Foot & Ankle Society (AOFAS) hindfoot score12 and a visual analog scale (VAS) score for pain (0 = no pain, 10 = the worst pain imaginable). We asked the patients to specify the "overall average pain" in or around the ankle in order to include intermittent or activity-related pain. (Thus, a pain value greater than 0 would be recorded if a patient with intermittent pain had a pain-free interval that coincided with the time of the assessment.) Patients were asked whether they were satisfied with the outcome of surgery, whether they would undergo the same surgery again, and whether they experienced difficulty climbing stairs. All intraoperative and postoperative complications, reoperations, and implant failures were noted. Failure was defined as exchange or pending exchange of the tibial and/or the talar component or implant removal and arthrodesis.
Radiographic Evaluation
Weight-bearing anteroposterior and lateral radiographs of the ankle, including functional radiographs (weight-bearing lateral views with the ankle in maximum plantar flexion and in maximum dorsiflexion), were made preoperatively and at each follow-up visit. The range of motion of the prosthetic ankle joint and the combined range of motion of the hindfoot and midfoot (including the ankle joint and the other foot joints) were assessed on the functional radiographs. Since hindfoot dorsiflexion or plantar flexion is a combined motion of the ankle, the subtalar joint, and the talonavicular joint, clinical measurement of the range of ankle motion is not possible because the contributions of the individual joints to the hindfoot range of motion cannot be distinguished clinically. We therefore considered the measurement on the functional radiographs to represent the best estimate of the range of motion of the ankle. The patient was instructed to bear weight on the foot and to perform maximal dorsiflexion starting from the neutral position, then maximal plantar flexion starting from the neutral position. Ankle alignment was determined on the anteroposterior radiograph by measuring the angle between the talar and the tibial component. An angle of 0° is ideal, and >5° was defined as varus or valgus ankle alignment.
The angle between the axis of the tibial component stem and the axis of the tibia (α) was measured in order to assess the alignment of the tibial component in the frontal plane. An angle of >5° was defined as varus or valgus alignment of the tibial component. The posterior slope of the tibial component with respect to the tibial axis (β) was measured on the lateral radiograph. The centering of the talar and the tibial component with respect to each other was analyzed on both the anteroposterior and the lateral radiograph. The components were considered to be improperly aligned if the medial or the lateral border of the talar component extended ≥2 mm beyond the border of the tibial component, as seen on the anteroposterior radiograph (Fig. 1). The components were also considered to be improperly aligned if either the anterior or the posterior metallic marker in the polyethylene inlay was located partially or completely within the zone created by extending the lines marking the anterior and posterior borders of the stem of the tibial component, as seen on the lateral radiograph (Fig. 2). Tibial radiolucencies >1 mm in width were assessed in five zones around the implant on the anteroposterior radiograph (zones 1 through 5; Fig. 3, A) and in five zones on the lateral radiograph (zones 6 through 10; Fig. 3, B). Talar radiolucencies were assessed in three zones around the implant on the lateral radiograph (Fig. 3, B). Migration of the tibial component was defined as a change in the α angle or the β angle of >3°. Migration of the talar component was defined as >2 mm of subsidence into the talar bone. We also noted the presence and location of periprosthetic cysts and osteophytes.
Statistical Methods
Kaplan-Meier survival curves with 95% confidence intervals were calculated, with censoring of the ankles at their latest follow-up. Preoperative and postoperative data were compared with use of paired t tests after the Shapiro-Wilk W test demonstrated that the data followed a normal distribution. Comparisons between groups were performed with use of unpaired t tests, also after verifying that the data followed a normal distribution. All data are presented as the mean and the standard deviation. All preoperative data (including percentages) were based on the entire series of 240 ankles, and all postoperative data were based on the 233 ankles with at least one year of follow-up.
Source of Funding
We received no external funding for this study.
Complications, Reoperations, and Failures
There were ten intraoperative complications (4.2%) and 20 postoperative complications (8.6%) (Table II). Eight of the ten intraoperative complications occurred within the first 100 cases. Reoperations were necessary in eighteen ankles (7.7%), with the most common reason being the removal of a painful osteophyte at the tip of the medial or the lateral malleolus (Table III). Postoperative complications and reoperations were significantly less frequent in patients with rheumatoid arthritis (see Appendix). One deep infection was successfully treated with open revision including synovectomy, irrigation, and exchange of the polyethylene insert followed by administration of intravenous antibiotics for fourteen days and oral antibiotics for six months. A second deep infection was successfully treated with arthroscopic lavage followed by intravenous antibiotics for fourteen days and oral antibiotics for six months. Five arthroplasties (2.1%; in two women and three men with a mean age of 59.6 years at the time of surgery) failed at a mean of twenty-seven months (range, twelve to fifty-three months). There were four cases of tibial loosening, which were treated with tibial component exchange in three ankles and with implant removal and tibiotalar arthrodesis in one ankle. One talar component was oversized, resulting in persistent pain, and was replaced. A structural allograft was inserted during the one tibiotalar arthrodesis, and the arthrodesis healed uneventfully. A low-grade infection was implicated as the reason for loosening in two of the tibial failures, as bacterial cultures of tissue samples taken during the revision surgery were positive. The estimated survival rate was 97.7% (95% confidence interval, 96.5% to 98.9%) at four years, with exchange or pending exchange of the tibial and/or the talar component or implant removal and arthrodesis as the end point (Fig. 4).
Clinical Outcomes
The AOFAS hindfoot score improved significantly, from a mean (and standard deviation) of 48.2 ± 17.5 points (range, 6 to 85 points) preoperatively to 84.3 ± 12.1 points (range, 44 to 100 points) at one year postoperatively (p < 0.001). The mean AOFAS score did not continue to improve after the first postoperative year, remaining at 84.1 ± 13.6 points (range, 31 to 100 points) at the time of the latest follow-up (p < 0.001 compared with the preoperative value). The mean preoperative and postoperative AOFAS scores were lower in patients with rheumatoid arthritis than in patients with primary osteoarthritis and in patients with posttraumatic arthritis (see Appendix), but the mean improvement in the AOFAS score was significantly greater in patients with rheumatoid arthritis than in patients with posttraumatic arthritis (44.9 compared with 29.3 points; p = 0.014). The mean VAS pain score decreased significantly from 7.7 ± 1.4 points (range, 2 to 10 points) preoperatively to 1.7 ± 2.0 points (range, 0 to 7 points) at the time of the latest follow-up (p < 0.001). The mean pain level one year postoperatively was significantly lower than the preoperative level (p < 0.001), and did not change further after the first postoperative year. The mean pain level at the time of the latest follow-up was significantly lower in patients with rheumatoid arthritis than in patients with primary or posttraumatic ankle arthritis (see Appendix). Pain, if present, was localized primarily in the anteromedial soft tissues. Most patients (73.6%) reported that they were very satisfied with the surgery, 22.2% were satisfied, and 4.2% were unsatisfied. Almost all patients (97.3%) stated that they would undergo the same procedure again, whereas only 2.7% stated that they would not. Patients typically had little or no difficulty climbing stairs, with 90.4% reporting no difficulty, 8.2% reporting slight difficulty, and 1.4% reporting severe difficulty.
Radiographic Outcomes
The mean ankle plantar flexion (i.e., plantar flexion of the total ankle replacement), as determined on the functional radiographs, improved slightly from 12.5° ± 7.0° preoperatively to 13.6° ± 6.4° at the time of the latest follow-up (p = 0.049), and the mean ankle dorsiflexion also improved slightly from 7.3° ± 5.9° to 8.3° ± 5.3° (p = 0.021). Thus, the mean total range of ankle motion improved from 19.8° ± 9.8° to 21.9° ± 8.7° (p < 0.001). The preoperative and postoperative range of ankle motion were moderately correlated (r = 0.41, p < 0.001). The total range of motion increased by >10° in forty-eight (21.1%) of the 228 ankles with complete radiographic follow-up and decreased by >10° in twenty-two ankles (9.7%). The plantar flexion, dorsiflexion, and total range of motion of the ankle did not improve further after the first postoperative year.
The mean combined range of motion improved significantly from 29.3° ± 14.1° to 34.6° ± 11.0° (p < 0.001). Forty-five (18.8%) of the ankles had a varus deformity preoperatively, and thirty (12.5%) had a valgus deformity. Varus or valgus malalignment between the tibial and the talar component was observed at the time of the latest follow-up in nine (3.9%) of 228 ankles, with varus deformity in five ankles and valgus deformity in four. Eight of the malalignments were residual, and one represented a new valgus malalignment that developed in an ankle with normal preoperative alignment. Seven of these nine postoperative ankle deformities were in patients with rheumatoid arthritis. One of these nine patients reported severe pain (VAS score, 7 points), whereas the remaining eight patients had a mean VAS pain score of 0.8 point (range, 0 to 2.5 points).
Ninety-three percent of the arthroplasty components were correctly centered in the frontal plane, and 97.4% were correctly centered in the sagittal plane. Five talar components were centered too far anteriorly and one was centered too far posteriorly. The tibial components were placed in a mean of 2.1° ± 2.9° (range, —5.5° to 10.2°) of varus relative to the tibial axis in the frontal plane (α angle). The mean posterior slope was 6.0° ± 3.8° (range, −5.8° to 17.1°) relative to the tibia in the sagittal plane (β angle).
Angulation of the ankle often resulted in difficulty in accurately assessing periprosthetic radiolucency when radiographs were obtained with a projection that was not perfectly parallel to the tibial plafond. Tibial zones 1 and 5 on the anteroposterior radiograph and zones 6 and 10 on the lateral radiograph were difficult to evaluate in 44% and 8% of the cases, respectively. The higher proportion of lateral radiographs allowing assessment can be explained by the availability of three different radiographs (functional radiographs in the neutral position, plantar flexion, and dorsiflexion), of which at least one usually represented a true lateral projection. Zones 2 through 4 could be assessed on all radiographs. The prevalence of tibial radiolucency ranged from 1.8% in zone 7 to 37.3% in zone 6 (see Appendix). The three talar zones were all easily assessed. The prevalence of talar radiolucency was 0% in zone 2 and 2.2% in zones 1 and 3 (see Appendix). The radiolucencies were nonprogressive, and none was >2 mm in width. There were no cases of tibial component migration. Nonprogressive subsidence of the talar component was detected in eight ankles (3.5%). In all eight cases, the subsidence occurred within the first postoperative year, and there was no evidence of further change or of loosening at the latest follow-up.
Periprosthetic cysts were detected in twenty ankles (8.8% of 228). The cysts were located in the medial malleolus in ten ankles, in the talus in four, in the lateral malleolus in two, in the distal aspect of the tibia in two, and in more than one location in two ankles. Eleven of the patients with bone cysts had primary ankle osteoarthritis, four had rheumatoid arthritis, and two had hemochromatosis. Cysts were significantly less common in patients with posttraumatic ankle arthritis (three ankles) (see Appendix). We observed a posterior osteophyte overhanging distally from the tibial plafond in sixty-one (26.8%) of the 228 ankles (Fig. 5). Patients with a posterior osteophyte had significantly less ankle plantar flexion than patients without a posterior osteophyte at the time of the latest follow-up (11.8° compared with 14.9°, p = 0.0014), but all other clinical and radiographic outcomes were similar. The prevalence of posterior osteophytes was highest in patients with primary ankle arthritis and lowest in patients with rheumatoid arthritis (see Appendix).
The current study involved a large, consecutive series of total ankle arthroplasties performed with the Mobility prosthesis. The short-term clinical and radiographic outcomes, including an estimated arthroplasty survival rate of 97.7% at four years, are encouraging and are at least comparable with the short-term results associated with other modern three-component total ankle arthroplasty designs.
Hintermann et al. reported in 2004 that component loosening occurred in 3.3% of HINTEGRA total ankle arthroplasties at 18.9 months postoperatively13. A more recent study by the same group reported that loosening or migration of the HINTEGRA talar component occurred in 6% of ankles after 2.8 years and was responsible for most of the implant-related reoperations14. The talar component of the Mobility total ankle prosthesis is apparently not at high risk for loosening, although we observed subsidence in 3.5% of the ankles. All occurrences of subsidence were noted at the one-year follow-up, with no evidence of further progression at later time points, consistent with the settling of a talar component that was not fully seated during surgery. None of the talar components in the present series failed due to loosening, and radiolucency surrounding this component was observed in only 2.2% of the ankles. Intraoperative complications occurred in ten ankles (4.2%). Nine of these had a very low impact on the patient because they were resolved intraoperatively (by screw fixation of a fracture of the medial malleolus or plating of the lateral malleolus), and we did not change the routine postoperative protocol in these patients. Only one intraoperative complication, an incomplete laceration of the tibial nerve, negatively impacted the patient outcome. There were twenty postoperative complications (8.6%) and eighteen reoperations (7.7%). Of the eighteen reoperations, only four were major procedures (two for the treatment of a deep infection, one muscle flap for wound coverage, and one arthrolysis for the treatment of arthrofibrosis). Eight minor procedures involved cheilectomy to remove a painful osteophyte at the tip of either the medial or the lateral malleolus; these were performed on an outpatient basis under ankle-block anesthesia and provided complete pain relief. These cheilectomies could have been avoided since the osteophyte was present, but not treated, at the time of the total ankle arthroplasty.
Higher complication and reoperation rates have been reported in several other studies. Valderrabano et al. reported a 34% reoperation rate at a follow-up of 3.7 years after implantation of the STAR prosthesis; 13% of the reoperations were implant-related15. Schutte and Louwerens reported a 32.7% rate of intraoperative complications and an 8.2% failure rate twenty-eight months after implantation of the same prosthesis16. However, Wood et al. reported a 95% survival rate of the STAR prosthesis after six years of follow-up17. A 14.8% complication rate and a 14.4% reoperation rate were reported three years after implantation of the HINTEGRA ankle prosthesis14. A 38% rate of intraoperative complications was reported in a series of fifty Agility ankle replacements, with a 16% rate of major reoperation and a 6% failure rate at 24.2 months18. In contrast, a low 2% reoperation rate at thirty-five months was reported for the Salto prosthesis19. A survivorship of 93.5% at ten years was reported for the Buechel-Pappas implant by its developers6, although another study of the same implant performed at an independent institution reported a six-year survivorship of only 79%17. Consequently, the failure rate, reoperation rate, and complication rate in the present study can all be considered relatively low. This might support the use of the Mobility implant, but the low rates might also be due to our previous experience with other total ankle arthroplasty designs20. Despite our previous experience, a learning curve was also observed in the present study, since eight of the ten intraoperative complications occurred among the first 100 cases. A steep learning curve with the HINTEGRA total ankle prosthesis was also reported recently by Lee et al.21.
The proper measurement of the range of ankle motion is an important consideration in studies of ankle arthroplasty. First, the true range of ankle motion (i.e., that of the total ankle arthroplasty) must be differentiated from the combined range of motion that involves all of the joints of the hindfoot and midfoot. The combined range of motion is greater than the range of ankle motion, and we strongly advise that this distinction be made clearly in future studies. The range of motion following total ankle arthroplasty has varied widely in previous reports22, and it is not clear that these differences can be explained entirely by the fact that a variety of implants were used. Second, the method used to evaluate the range of ankle motion is important. We do not consider clinical assessment to be an adequate method, and we strongly recommend determining the range of motion with use of functional radiographs. Although radiographic measurements may be reliable and somewhat more objective than clinical assessments23,24, the "true" range of ankle motion may still be underestimated, as there is great dependence on the ability of the technician and on patient compliance to maximally flex and extend the ankle during radiography. Nevertheless, the postoperative range of motion in our series was somewhat disappointing. We have concluded that the range of motion after total ankle arthroplasty is largely determined within the first weeks after surgery. Concerns regarding wound-healing complications and the development of ankle equinus during this period initially led us to immobilize the ankle in a cast for several weeks postoperatively. However, because of the limited improvement in the range of motion after this period, we have since implemented changes in our postoperative protocol to shorten the period of immobilization.
The clinical results, as assessed with use of the VAS pain score and the AOFAS hindfoot score, compare well with those in previous reports13-15,19 as well as with values presented in a recent meta-analysis of total ankle arthroplasty outcomes25. Patients with rheumatoid arthritis experienced the greatest improvement and patients with posttraumatic ankle arthritis experienced the least improvement. This observation is in contrast to the mixed results of total ankle arthroplasty in patients with rheumatoid arthritis reported in several previous studies26-28, but it is consistent with the results presented by Bonnin et al., who reported a 100% survival of the Salto prosthesis in patients with rheumatoid arthritis after sixty-eight months19. The difference in outcomes between patients with rheumatoid arthritis and those with posttraumatic arthritis may be related to the differences in functional demands and expectations between these two groups, as highlighted in a previous study analyzing the sports and activity habits of patients after total ankle arthroplasty29. Alternatively, the difference may be due to the preoperative AOFAS scores, which were lower in the patients with rheumatoid arthritis than in the other patients in our series. Most patients who reported continuing pain indicated that the pain was localized in the anteromedial aspect of the ankle, and the rate of stress fractures of the medial malleolus was 4.3%. Future studies with longer follow-up and biomechanical investigations will be needed to evaluate these problems on the medial side of the ankle following modern three-component total ankle arthroplasty.
The prevalence of nonprogressive radiolucency (<2 mm in width) ranged from 1.8% to 37.3% in the ten zones surrounding the tibial component. The rate of periprosthetic radiolucency reported in previous studies is highly variable, ranging from 0% to almost 80%26,30,31. However, the clinical importance of these radiolucencies remains unclear. The formation of new bone around the posterior aspect of the tibial component (the posterior osteophyte, Fig. 5) has been reported previously32. Patients with a posterior osteophyte had significantly less plantar flexion of the ankle than patients without an osteophyte; however, the slight difference of 3° does not appear to be clinically relevant.
This study has some limitations. The length of follow-up (particularly the one-year minimum follow-up) was short, although it should be noted that we identified no further change in clinical outcomes after the first postoperative year. Also, some of the patients could not be followed clinically at our institution. We tried to reduce this potential bias by contacting these patients by telephone in order to verify that the total ankle arthroplasty had not failed and that there had been no reoperation. Finally, the subjective patient satisfaction questionnaire has not been validated.
In conclusion, the short-term results of total ankle arthroplasty with the Mobility prosthesis are encouraging and are at least comparable with those of other modern three-component implants. Total ankle arthroplasty must be considered a technically demanding surgical procedure. Successful total ankle arthroplasty often requires a full understanding and treatment of associated pathologic conditions, which are common (as highlighted by the fact that concomitant procedures were performed in more than half of the feet in the present series). Continuing efforts are warranted to further reduce complication and reoperation rates in order to permit even more successful patient outcomes.