Background: There are several types of prostheses available to surgeons when performing a total ankle arthroplasty (TAA). The main objective of this study was to summarize the clinical and functional outcomes of 4 TAA prostheses: the Hintegra implant (Integra LifeSciences), the Agility implant (DePuy), the Mobility implant (DePuy), and the Scandinavian Total Ankle Replacement (STAR) implant (Small Bone Innovations [SBi]).
Methods: Patients were prospectively recruited. A total of 451 TAAs with a mean follow-up (and standard deviation) of 4.5 ± 2.0 years were included. Patients were assessed annually and completed self-reported outcome measures at these visits. Complications and revisions were reported at the time of incident. Mean improvements are reported by prosthesis. Linear mixed-effects models were used to obtain adjusted comparisons of scores across prostheses. Survivorship curves were generated by prosthesis and type of complication.
Results: Mean improvement in the Ankle Osteoarthritis Scale (AOS) total score was less among patients with the Mobility implant (19.5; 95% confidence interval [CI], 15 to 24) than it was among patients with the Agility implant (29.1; 95% CI, 24 to 34), Hintegra implant (29.7; 95% CI, 27 to 33), and STAR implant (28.5; 95% CI, 23 to 34). Patients in the Mobility group also had less mean improvement in the AOS pain score (21.3; 95% CI, 17 to 26) compared with patients in the Hintegra (29.0; 95% CI, 26 to 32), Agility (29.8; 95% CI, 25 to 35), and STAR (29.1; 95% CI, 23 to 35) groups. The Mobility group also had less mean improvement in the AOS disability score (17.3; 95% CI, 12 to 23) compared with the Hintegra (30.4; 95% CI, 27 to 34), Agility (28.8; 95% CI, 23 to 34), and STAR (27.8; 95% CI, 21 to 34) groups. Survival results among the 4 prostheses are reported.
Conclusions: This study demonstrated acceptable outcomes of 4 modern TAA prostheses. Outcome results from patient-reported scores were comparable between at least 3 of the 4 prostheses (the Hintegra, STAR, and Agility implants). The rates of complications and revisions found in this study are within the limits reported in the literature for similar prostheses and methods of reporting.
Level of Evidence: Therapeutic Level II. See Instructions for Authors for a complete description of levels of evidence.
Total ankle arthroplasty (TAA) is a frequently selected treatment for end-stage ankle arthritis. Modern implant design changes feature the development of an uncemented, unconstrained prosthesis with modern bearing surfaces. Outcome studies of modern TAA prostheses have demonstrated good patient satisfaction and clinical outcomes1,2. Surgeons who perform total ankle arthroplasty now have several options of implants to use for the procedure. There were 4 commonly used TAA implants available in Canada at the time of review: the Scandinavian Total Ankle Replacement (STAR) implant (Small Bone Innovations [SBi]), the Agility implant (DePuy), the Hintegra implant (Integra LifeSciences), and the Mobility implant (DePuy).
The current STAR implant is the third-generation design. The tibial component has a flat articular side and a double-barrel design for fixation into the distal aspect of the tibia. The talar component is side-specific and has a central ridge that matches the polyethylene liner. Mann et al. published a long-term outcome study of 84 TAA procedures involving the use of the STAR prosthesis with an average follow-up of 9 years; the 5-year survivorship rate was 96% and the 10-year survivorship rate was 90%3.
Similar to the STAR implant, the Hintegra prosthesis is a 3-component, unconstrained design. The flat articular surface on the tibial component has a porous coating for bone ingrowth as well as an anterior flange that allows 2 screws to be placed into the distal aspect of the tibia for additional fixation. The implant has anatomically contoured components that reportedly can mimic physiologic flexion and extension as well as provide axial rotation4. The largest study to date that we are aware of regarding the Hintegra implant, published in 2006 by Hintermann et al.5, reported outcomes of 271 TAA procedures with an average follow-up of 36 months. A total of 39 revision surgeries were required during the follow-up period.
The Mobility implant was designed to maintain bone stock while attempting to preserve ankle motion. The prosthesis has a long conical stem on the tibial side to maximize bone ingrowth and stability. The talar component rests on the preserved medial and lateral cortices of the talus. Sproule et al. reported the early-term results of TAA using the Mobility implant in 88 patients. Their results showed good pain relief and functional outcome in 82% of patients but bone-implant interface abnormalities in 43% of the retained prostheses6. The Mobility implant is no longer commercially available.
The Agility implant has a wide tibial base plate that bridges the fused tibiofibular syndesmosis to distribute forces. The medial and lateral sides of the tibial component sit against the medial and lateral malleoli. It differs from the other implants mentioned above in that it has a semiconstrained design. It had been approved for cemented use but was often used in an uncemented manner. The Agility implant is no longer commercially available.
Three systematic reviews of the literature addressed the question of a comparison of outcomes of TAA1,7,8. Each of these articles concluded that high-quality prospective studies were needed to further evaluate this question.
The main objective of the current study was to summarize the outcomes of 4 TAA prostheses (Hintegra, Agility, Mobility, and STAR). The main outcome measures were scores of the Short Form (SF)-36 mental component summary (MCS) and physical component summary (PCS) and Ankle Osteoarthritis Scale (AOS)9-11. Secondary outcomes included complications and revision rates. We hypothesized that patients recruited prospectively and followed to a mid-term outcome at a minimum of 2 years would demonstrate a difference in patient-reported outcomes among the total ankle arthroplasty implants used.
Materials and Methods
Patients with end-stage ankle arthritis were prospectively recruited and their data were collected in the Canadian Orthopaedic Foot and Ankle Society (COFAS) Ankle Reconstruction Database. Patient recruitment took place in 4 tertiary-care academic teaching hospitals in Halifax, Nova Scotia; Toronto, Ontario; and Vancouver and Victoria, British Columbia. Research ethics board approval was obtained from all participating centers, and this study was registered with www.ClinicalTrials.gov (NCT00552136).
All enrolled patients were seen initially for consultation with 1 of the senior authors (T.D., M.G., A.Y., P.D., K.W., and M.J.P.), who are subspecialty-trained foot and ankle surgeons. All patients with end-stage ankle arthritis were treated initially with nonsurgical modalities. Patients in whom nonsurgical treatments failed elected to proceed with ankle arthroplasty in consultation with of 1 of the senior authors. Implant selection was left to the discretion of the treating surgeon. The usage of different implants by the surgeons and the changes in such decisions that took place over the study period represent the varying factors that contribute to a surgeon’s selection of a particular implant, i.e., implant availability (cost, hospital influence, and market considerations), best available evidence in favor of an implant at the given time (and the surgeon’s interpretation of the evidence), and previous favorable or nonfavorable experiences with a particular implant.
Included in this study were patients who underwent a primary TAA between November 2001 and August 2010, who had baseline data available for analysis and a minimum of 2 years of follow-up, and who provided informed consent to participate. Patients were excluded from the database if they had any of the following: osteonecrosis of the talus or tibia, severe foot or ankle deformity, active infection in the ankle within 12 months prior to surgery, medical conditions precluding safe surgery, muscle or nerve disease, or severe osteoporosis. Also excluded were patients who were <40 years of age, pregnant, or Workers’ Compensation Board patients, and those with mental illness preventing them from following the study schedule.
Patients who underwent TAA provided baseline demographic data and were assessed using patient-reported outcome measures. Follow-up scores for each of the patient-reported measures were obtained at annual visits. The treating surgeon reported all reoperations.
Data were analyzed by comparing the outcomes for each implant. Patient demographics at baseline were summarized by implant type. Improvement in the AOS, PCS, and MCS scores was assessed by subtracting the baseline scores from the scores obtained at the time of the last follow-up. Outcomes were compared across implant types using a linear mixed-effects model with clustering by surgeon and adjusted for patient age, sex, the presence of diabetes, the presence of inflammation, body mass index (BMI), smoking status, and baseline score.
All operations after the index procedure were assigned a code according to the COFAS Ankle Reconstruction Reoperations Classification System (Table I). In cases in which multiple additional procedures were performed, all reoperation codes were collected. Implant failure was analyzed both as failure of the metal components (reoperation code 9 or 10) and as failure of the metal components or isolated polyethylene exchange (reoperation code 5). Amputation was also considered to be a failure of the implant (reoperation code 11). The time to (first) implant failure was evaluated using the Kaplan-Meier curve and compared across implant types using a Cox frailty survival model with clustering by surgeon and adjusted for patient age, sex, the presence of diabetes, the presence of inflammation, BMI, operatively treated side, and smoking status.
A total of 451 TAAs were included in this comparison (75 Agility, 209 Hintegra, 92 Mobility, and 75 STAR implants). The mean follow-up (and standard deviation) was 4.5 ± 2.0 years and by prosthesis type was 3.5 years for the Hintegra, 4.2 years for the Mobility, 6.1 years for the Agility, and 6.2 years for the STAR group. Patient demographics were similar with respect to age, sex, BMI, smoking status, and presence or absence of inflammatory arthritis as well as some comorbidities (Table II).
The mean preoperative AOS total score (and standard deviation) by implant was as follows: Hintegra, 51.6 ± 17.8; Mobility, 49.6 ± 15.1; Agility, 60.6 ± 14.5; and STAR, 50.3 ± 19.0. The mean preoperative AOS pain score was higher for the Agility group (55.3 ± 16.8) compared with that for the Hintegra group (46.4 ± 18.8), the Mobility group (45.3 ± 16.3), and the STAR group (46.4 ± 20.8). In addition, the mean preoperative expectation score was lower for those with the Agility implant (13.4 ± 15.0) compared with those with the Hintegra implant (20.7 ± 21.1), the Mobility implant (24.3 ± 18.1), and the STAR implant (26.1 ± 21.7) (Table III). The mean preoperative MCS and PCS and the mean satisfaction scores were comparable (Table III).
The patients with a Mobility implant demonstrated less mean improvement in the AOS total score (19.5) than did those with the Hintegra implant (29.7), Agility implant (29.1), and STAR implant (28.5) (Table IV). Less mean improvement in the AOS pain score (21.3 points) was noted for patients with the Mobility implant compared with the Hintegra (29.0), Agility (29.8), and STAR (29.1) implants. Finally, less improvement in the disability score (17.3) was also shown for those with the Mobility prosthesis compared with the Hintegra (30.4), Agility (28.8) and STAR (27.8). Although the patients with the Mobility implant also had the lowest mean improvement in PCS and MCS scores from baseline to the latest follow-up, there was a notable overlap of the confidence intervals among all implant types (Table IV).
In the mixed-effects analysis, the adjusted mean AOS total score at the time of follow-up was significantly better for the Hintegra implant compared with the Mobility implant (a 9.2-point difference; p = 0.002) as well as for the STAR implant compared with the Mobility implant (a 10.0-point difference; p = 0.005). The same was true of the AOS disability score, with an adjusted mean difference of −10.9 between the Hintegra and Mobility implants (p = 0.001) and −11.8 between the STAR and Mobility implants (p = 0.002). There was also a significant difference in the mean AOS pain score between the Hintegra and Mobility implants (adjusted difference of −6.9; p = 0.02) and between the STAR and Mobility implants (adjusted difference of −8.0; p = 0.02). Finally, we found a significant difference in the SF-36 PCS when comparing the Hintegra with the Mobility implant (adjusted difference of 4.1; p = 0.01) and when comparing the STAR with the Mobility implant (adjusted difference of 3.6; p = 0.04); the Mobility implant had worse outcome scores in all comparisons.
The total number of complete revisions, whereby the metal components were not retained (reoperation codes 9 and 10), was 17 in the Agility group (23%), 17 in the Mobility group (19%), 6 in the STAR group (8%), and 16 in the Hintegra group (8%) (Tables I and V). All but 2 of the Mobility implants that were not retained underwent at least 1 revision procedure; the other 2 patients underwent conversion to ankle fusion. The only implant in this series that required reoperation for an isolated polyethylene liner exchange due to polyethylene liner failure (code 5) was the STAR implant, with a rate of 11% (8 of 75 implants).
Kaplan-Meier curves for time to implant failure are shown in Figure 1 (metal component revision) and Figure 2 (metal component revision or isolated polyethylene exchange). For both definitions of implant failure, the time to implant failure differed significantly across implant types (p < 0.005) in the Cox frailty model.
When comparing the Hintegra implant with the Mobility implant, the adjusted odds ratio [OR] was 0.31 (95% confidence interval [CI], 0.15 to 0.66; p = 0.002); when comparing the STAR with the Mobility implant, the adjusted OR was 0.37 (95% CI, 0.17 to 0.82; p = 0.016); and finally, when comparing the Agility with the Mobility implant, the adjusted OR was 0.48 (95% CI, 0.23 to 1.01; p = 0.052) (Table VI).
In this study, we noted subtle but important differences among 4 TAA prostheses. This adds to the complexity of decisions around the surgical treatment of end-stage ankle arthritis. First, a surgeon needs to determine if a patient is a suitable candidate for a TAA and then determine which implant to employ.
The DePuy Agility and Mobility implants both demonstrated higher rates of metal component revisions. While the cause of this was outside the scope of this study, other authors have found similar results. Specific to the Agility implant, a twofold (6.6% to 12.2%) increased rate of failure (metal component revisions or conversion to fusion) was identified when data from the designer’s practice was excluded12. Data registry trends have also identified increasing failure of the Mobility prosthesis with longer-term follow-up and identified the most common failure mechanisms as aseptic osteolysis and talar component subsidence13. A radiostereometric study of the Mobility implant demonstrated excess early micromotion, which may contribute to aseptic loosening and implant subsidence14. A cadaveric biomechanical investigation of the Agility prosthesis demonstrated increased relative motion that may contribute to higher rates of aseptic loosening15.
Patients with the Agility prosthesis were found to have higher AOS scores and lower expectation scores preoperatively than the other patient cohorts, despite similar baseline demographics. While there was not a clear cause for this in our data, the Agility patients were many of the earliest patients enrolled in the study. It is possible that there were subtle differences not captured in our baseline demographic information. We can speculate that some of these patients may have experienced their disability longer than the other cohorts of patients or simply underwent TAA surgery when outcomes were less well known and, therefore, expectations were set lower.
The STAR implant demonstrated good survival of the metal components in comparison with the Agility and Mobility implants. However, this was the only type of prosthesis in our series that required exchange of the polyethylene for failure of the polyethylene liner. Mann et al. reported on 84 STAR implants with an average follow-up of 9.1 years3. The complication rates were similar to those in our study, but that series had no reported cases of isolated polyethylene failure. Other studies have reported similar rates of isolated polyethylene failure due to fracture, incomplete fracture, and severe wear16,17 and such failure likely relates to implant design.
The Hintegra prosthesis demonstrated good patient outcomes and good survival of the components. This is consistent with a study by Choi et al.18, in which a small series of patients with the Hintegra prosthesis was compared with patients with the Mobility prosthesis and is also consistent with the short-term follow-up report by Hintermann in 20065. In our study, the Hintegra prostheses were placed in the later part of the study period; surgeons were likely more experienced and the follow-up was of shorter duration.
This study provides high-quality evidence to assist surgeons in selecting an appropriate TAA prosthesis. Patients were recruited from 4 centers, and the procedures were performed by 5 different surgeons. This adds to the reproducibility of the results. Patients were recruited preoperatively and followed in a prospective manner that minimized bias. A small loss to follow-up also minimized bias.
The primary weakness of this study was that of implant selection. While most of the participating surgeons used multiple types of implants throughout the study period, there was not a perfect distribution among the surgeons. The frequency with which a particular implant was selected changed throughout the study period and reflects the dynamic practice pattern of each of the surgeons. This is summarized in Figure 3. Furthermore, the implants that were inserted earlier in the study period had longer follow-up. We attempted to address this bias with an adjusted analysis of the implant survival. This bias would best be addressed with a randomized controlled design. However, there would be substantial challenges to executing a randomized controlled trial with comparable recruitment numbers and follow-up to those reported in this study.
In summary, TAA is a viable and reproducible procedure to offer to patients with end-stage ankle arthritis. The type of prosthesis selected impacts both clinical and patient-reported outcome as well as survival and rates of complications.
Investigation performed at St. Michael’s Hospital, Toronto, Ontario, St. Paul’s Hospital, Vancouver, British Columbia, Vancouver Island Health Authority, Victoria, British Columbia, and Queen Elizabeth II Health Sciences Center, Halifax, Nova Scotia, Canada
Disclosure: Direct or indirect research funding for this study was received from Integra LifeSciences and DePuy. An unrestricted research grant from DePuy supported data collection involving the Mobility prosthesis for each patient entered in the COFAS database. Some patients receiving a Mobility total ankle replacement at the Dalhousie site were also part of an independent radiostereometric analysis study supported by an unrestricted research grant from DePuy. On the Disclosure of Potential Conflicts of Interest forms, which are provided with the online version of the article, one or more of the authors checked “yes” to indicate that the author had a relevant financial relationship in the biomedical arena outside the submitted work and “yes” to indicate that the author had a patent and/or copyright, planned, pending, or issued, broadly relevant to this work.
- Copyright © 2017 by The Journal of Bone and Joint Surgery, Incorporated