Abstract
Aims
Revision rates for ankle arthroplasties are higher than hip or knee arthroplasties. When a total ankle arthroplasty (TAA) fails, it can either undergo revision to another ankle replacement, revision of the TAA to ankle arthrodesis (fusion), or amputation. Currently there is a paucity of literature on the outcomes of these revisions. The aim of this meta-analysis is to assess the outcomes of revision TAA with respect to surgery type, functional outcomes, and reoperations.
Methods
A systematic review was conducted using Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. PubMed, Medline, Embase, Cinahl, and Cochrane reviews were searched for relevant papers. Papers analyzing surgical treatment for failed ankle arthroplasties were included. All papers were reviewed by two authors. Overall, 34 papers met the inclusion criteria. A meta-analysis of proportions was performed.
Results
Six papers analyzed all-cause reoperations of revision ankle arthroplasties, and 14 papers analyzed failures of conversion of a TAA to fusion. It was found that 26.9% (95% confidence interval (CI) 15.4% to 40.1%) of revision ankle arthroplasties required further surgical intervention and 13.0% (95% CI 4.9% to 23.4%) of conversion to fusions; 14.4% (95% CI 8.4% to 21.4%) of revision ankle arthroplasties failed and 8% (95% CI 4% to 13%) of conversion to fusions failed.
Conclusion
Revision of primary TAA can be an effective procedure with improved functional outcomes, but has considerable risks of failure and reoperation, especially in those with periprosthetic joint infection. In those who undergo conversion of TAA to fusion, there are high rates of nonunion. Further comparative studies are required to compare both operative techniques.
Cite this article: Bone Jt Open 2022;3(7):596–606.
Take home message
Revision of primary total ankle arthroplasty (TAA) can be an effective procedure with improved functional outcomes, but has considerable risks of failure and reoperation, especially in those with periprosthetic joint infection.
Conversion of TAA to fusion has high rates of nonunion.
Introduction
Ankle arthritis has been estimated to effect 47.7 per 100,000 people in the UK, and 29,000 cases are referred to specialists each year.1 The surgical treatment of ankle arthritis is either an ankle fusion or total ankle arthroplasty (TAA). Over 1,000 TAAs are performed annually in the UK, and it is thought a much larger number of ankle arthrodeses (fusions) are undertaken.2
When a TAA fails it can either undergo a revision TAA, a conversion to fusion, or below-knee amputation. A revision TAR is defined as any procedure with removal of a component of the ankle arthroplasty.3
According to the National Joint Registry for England and Wales (NJR), the five-year revision rates for TAA are 6.86% compared to 2.29% for total hip arthroplasties and 2.66% for total knee arthroplasties.2 The number of revisions of TAA is increasing year on year.2 Unfortunately, it is thought that this number underestimates the true burden of failed ankle arthroplasties due to under reporting of conversions of arthroplasty to fusion.2
As the number of ankle arthroplasties increases, so too will the total number of patients requiring further surgery for failure. The most common indications for ankle arthroplasty failure are aseptic loosening, lysis, pain, malalignment, and infection.2
There is a scarcity of literature on the surgical management of the failed TAA, and the published evidence is controversial.4,5 Therefore, the aim of this systematic review is to assess the outcomes of revision TAA and conversion to fusion following failed TAA, with respect to functional outcomes, complications, and reoperation.
Methods
Data sources, search strategy, and screening
A systematic review was undertaken following Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. PubMed, Medline, Embase, Cinahl, and Cochrane reviews were searched for relevant papers. The search terms used were a combination of (ankle AND (arthroplasty or arthroplasty)) AND (ankle AND (salvage OR arthrodesis OR fusion OR reconstruction)) AND ((revision ankle arthroplasty) OR (revision ankle arthroplasty)).
All references identified were cross-referenced for further papers for inclusion. This resulted in 511 papers identified. Following this, 359 abstracts were reviewed, which resulted in 84 full papers. Each of these were reviewed by two authors (TJ, CSD) independently. There were a total of 33 papers that met the inclusion criteria, with 15 analyzing revision TAA and 23 analyzing conversion of a failed TAA to an ankle fusion, of which five analyzed both revision and conversion (Figure 1).
Fig. 1
Eligibility criteria
Any papers that related to the surgical treatment of a failed TAA were included with outcomes of failure and further surgery. Papers were excluded if they 1) had less than a minimum 12 months’ follow-up, 2) any paper that grouped revision and primary ankle arthroplasties together, 3) any paper that grouped revision TAA and conversion to fusion together, 4) papers not in English language, 5) case reports, and 6) outcomes of further surgery.
Data extraction and statistical analysis
Two reviewers (TJ, CSD) independently reviewed all included papers. Data recorded included the number of patients, demographics, details of primary procedure, details of revision procedure, and outcomes including further surgical procedures and outcome scores. Analyzing indication for primary ankle arthroplasty, all different inflammatory arthritis were grouped together, and post-traumatic arthritis and primary osteoarthritis were grouped together.
Analyzing the reason for ankle arthroplasty failure, all known causes were grouped together into either aseptic or septic failure due to differences in reporting between studies. In both of these, there was considerable variation in reporting between studies and this classification prevented ambiguity. Not all studies were included in all analysis due to differences in reporting.
Definitions
The overall reoperation rate for revision ankle arthroplasty or conversion to fusion was defined as all-cause surgical interventions.
A revision procedure for a failure of a revision ankle arthroplasty was defined as any procedure where one or more of the components were removed. This included re-revision to another arthroplasty, conversion to fusion, or amputation. For those that underwent conversion to fusion, the revision procedure was defined as a further attempt at fusion at the same level, an extension of the fusion to adjacent joints, or an amputation.
Union following conversion to fusion was classified based on the authors’ definition, and defined as union following a single surgical procedure. If secondary procedures were required prior to union then this was classified as a nonunion.
If there was any ambiguity or uncertainty about the results, then these were discussed among the authors. Where the data were considered unreliable, these were excluded from that specific analysis. Therefore, in different analyses it was accepted that there may be differing numbers of patients included in each analysis.
Study bias was assessed using the Methodological Index for Non-Randomized Studies (MINORS) criteria. This is designed with eight items, each of which are scored as 2 (reported or adequate), 1 (reported but inadequate), or 0 (not reported). This gives a total score of 16 for non-comparative studies.
Statistical analysis
Descriptive statistics were calculated. Statistical analysis was undertaken using Stata version 15 (Stata Corp, USA). The total number of patients undergoing the surgical procedure was calculated. The number of failures, non-failure reoperations, and union was calculated based on the above definitions. Proportions with 95% confidence intervals (CIs) for each study were calculated and weighting based on study size. Using these proportions a meta-analysis was performed. The metaprop command in Stata was used to perform a random effects meta-analysis pooling percentages using the Freeman-Tukey arscine transformation of the percentage. This produced a pooled percentage for these with 95% CIs.
Results
A total of 15 papers that analyzed revision ankle arthroplasties met the inclusion criteria, and these covered 397 patients; 23 papers with 480 ankles in which a failed TAA was converted to fusion were included (Tables I and III). Five papers included patients from both procedures. All papers were Level III or IV evidence. Overall, there were 14 studies from the USA and 20 from Europe. For those studies on revision ankle arthroplasties, ten out of 15 were from the USA, but only seven of 23 for conversion to fusion (p = 0.0281, chi-squared test).
Table I.
Author | Year | Country | TAAs, n | Mean age, yrs | Female, n (%) | Mean follow-up, yrs | Aetiology | Mean time since primary, yrs (range) | Primary implant removed | Indication for revision | Revision Implants |
---|---|---|---|---|---|---|---|---|---|---|---|
Lachman et al6 | 2018 | USA | 29 | 62.4 | 44.8 | 3.3 | 82.8% arthritis, 17.2% inflammatory | 3.9 (0.2 to 7.3) | INBONE I 15, Salto 8, STAR 5, Infinity 1 | 100% aseptic | INBONE II 18, INBONE I 5, Salto XT 3, Infinity 2, STAR 1 |
Wagener et al7 | 2017 | Switzerland | 12 | 53 | 41.7 | 6.9 | 83.3% arthritis, 16.7% inflammatory | 7.8 (2 to 37) | 8 STAR, 2 Hintegra, 1 Mobility, 1 Irvine. second revision in 4 | 100% aseptic | Hintegra with custom made talus |
Kamrad et al8 | 2015 | Sweden | 73 | 55 | 60.3 | Not stated | 78.1% arthritis, 21.9% inflammatory | 1.8 (0 to 9.2) | STAR 39, CCI 10, BP 8, AES 4, Hintegra 5, Mobility 1, Rebalance 2 | 97.3% aseptic, 2.7% septic | Not stated |
Roukis and Simonson9 | 2015 | USA | 32 | 64.6 | 34.4 | 2.1 | Not stated | 6.4 (1.6 to 12.4) | Agility and Agility LP | 93.7% aseptic, 6.3% septic | 23 Agility or Agility LP, 8 INBONE II, 1 Salto Talaris XT |
Horisberger et al10 | 2015 | USA | 10 | 52 | 60 | 4 | Not stated | 6 (2 to 11) | 2 Agility, 4 Hintegra, 2 STAR, 1 BP, 1 Salto | 100% aseptic | Hintegra |
Patton et al11 | 2015 | USA | 14 | 61.9 | 42.9 | 4.6* | 85.7% arthritic, 14.3% inflammatory | Not stated | 11 Agility, 3 Salto | 100% septic | 11 Agility, 1 Salto 2 Inbone, 13 2 stage, 1 1 stage |
Ellington et al12 | 2013 | USA | 41 | 59.5 | 71 | 4.1 | 85.4% arthritic, 14.6% inflammatory | Not stated | 52 Agility | 100% aseptic | Agility (15 talar only, 26 combined) 19 custom talus |
Hintermann et al4 | 2013 | Switzerland | 117 | 55 | 47.9 | 6.2 | Not stated | 4.3 | Not stated | 92% aseptic, 8% septic | Hintegra |
DeVries et al13 | 2013 | USA | 14 | 65.2 | 42.9 | 2.4 | 92.9% arthritic, 7.1% inflammatory | 7.8 (3.5 to 23) | Agility | 100% aseptic | Inbone |
Schuberth et al14 | 2011 | USA | 17 | Not stated | Not stated | 1 | Not stated | Not stated | Not stated | 100% aseptic | Inbone+ metal-reinforced bone cement augmentation |
-
*
Includes all in the paper, not just revision procedures.
-
TAA, total ankle arthroplasty.
Table II.
Author | Year | Country | TAAs, n | Mean age, yrs | Female, n | Follow-up, yrs | Primary indication | Time since primary, yrs | Primary implant | Indication | Procedure |
---|---|---|---|---|---|---|---|---|---|---|---|
Halverson et al15 | 2019 | USA | 5 | 63.2 | 40.0 | 5.2 | Not stated | 6.1 | 1 STAR, 2 Agility, 1 Salto Talaris, 1 InBone | 80% aseptic, 20% septic | IM nail |
Kruidenier et al16 | 2019 | Netherlands | 47 | 63 | 60.9 | 6.6 | Not stated | Not stated | 10 Beuchel–Pappas, 29 Cobalt Coated Implant, 4 Low contact stress, 1 STAR, 1 Salto Talaris, 1 AES, 1 Hintegra | 78.7% aseptic, 21.3% septic | 33 plating, 8 internal screws, 5 IM nail, 1 external fixation |
Ali et al17 | 2018 | UK | 23 | 67 | 18.2 | 1.2 | Not stated | Not stated | AES | 100% aseptic | IM nail |
Aubret et al18 | 2017 | France | 10 | Not stated | Not stated | 1.6 | 90% arthritis, 10% inflammatory arthritis | 6.9 | 7 AES, 2 Integra, 1 Ramses, 1 Salto | 100% aseptic | Trabecular Metal Implant, 10 IM nail, 1 plates |
Kamrad et al5 | 2016 | Sweden | 118 | 61 | 59.3 | 2 | 60% arthritis, 40% inflammatory | Not stated | 61% STAR, 12% AES, 11% Mobility, 8% BP, 5% CCI, 3% Hintegra | 88% aseptic, 12% septic | 49% IM nail, 13% plate fixation 8% metal spacer with plate or nail, 6% ex fix, 5% screw, 19% not recorded |
Rahm et al19 | 2015 | Switzerland | 23 | 62 | 65.2 | 3.2 | 100% arthritis | 4.67 | 16 Agility, 3 STAR, 2 Hintegra, 1 BP, 1 SALTO | 73.9% aseptic, 26.1% septic | Mixture |
Paul et al20 | 2014 | Switzerland | 6 | 55 | 50 | 2.2 | Not stated | Not stated | Not stated | 83.3% aseptic, 16.7% septic | IM nail |
McCoy et al21 | 2012 | USA | 7 | 52 | 42.9 | 4.8 | 100% arthritis | 5.9 | 5 prior revisions | 57.1% aseptic, 42.9% septic | External fixator |
Berkowitz et al22 | 2011 | USA | 24 | 61.7 | 45.8 | 3.7 | 79.2% arthritis, 20.8% inflammatory arthritis | 4.4 | 15 Agility, 3 Agility long stemmed talus, 7 STAR, 2 BP | 91.7% aseptic, 8.3% septic | 12 plate, 12 IM nail |
Doets and Zürcher23 | 2010 | Netherlands | 18 | 55 | 77.8 | 7.3 | 16.7% arthritis, 83.3% inflammatory arthritis | 4 | 6 New Jersey, 11 BP, 1 CCI | 94.4% aseptic, 5.6% septic | 7 plate, 6 IM nail, 1 k wire 4 screws |
Henricson and Rydholm24 | 2010 | Sweden | 13 | Not stated | Not stated | 1.4 | 53.7% arthritis, 46.2% inflammatory arthritis | 7 | 9 STAR, 2 AES, 1 Mobility, 1 BP | 100% aseptic | TM tibial cone and IM nail |
Plaass et al25 | 2009 | Switzerland | 9 | 59.9 | 44.4 | Not stated | Not stated | Not stated | Not stated | 100% aseptic | anterior double plate |
Culpan et al26 | 2007 | France | 16 | 54 | 68.8 | 3.75 | 81.3% arthritis, 18.7% inflammatory | 3.4 | 1 New Jersey, 3 BP, 1 Mendolia, 1 Custom, 8 SALTO, 2 STAR | 93.7% aseptic, 6.3% septic | Screws |
Schill27 | 2007 | Germany | 15 | 56 | 20 | 1.9 | Not stated | 6.73 | 6 Thompson-Richards, 8 STAR, 1 Salto | 100% aseptic | IM nail |
Hopgood et al28 | 2006 | UK | 23 | 62 | 40.9 | 2.4 | 52.2% arthritis, 47.8% inflammatory arthritis | 3.42 | 15 STAR, 6 BP, 2 others | Not stated | 13 screws, 10 IM nail |
Anderson et al29 | 2005 | Sweden | 16 | 62 | 93.3 | 2.8 | 100% inflammatory arthritis | Not stated | 10 STAR, 6 cemented (3 B + W, 1 ICLH, BP) | Not stated | IM nail |
Carlsson et al30 | 1998 | Sweden | 21 | 59 | 85.7 | Not stated | 14.3% arthritis, 85.7% inflammatory arthritis | 3.33 | 8 Bath & Wessex, 5 custom, 3 ICLH, 2 STAR, 2 St George, 1 New Jersey | 81.0% aseptic, 19.0% septic | External fixator |
Kitaoka31 | 1992 | USA | 38 | 56.8 | 61.1 | 8.3 | 73.7% arthritis, 26.3% inflammatory arthritis | 3.5 | Mayo 30, others 8 | 84.2% aseptic, 15.8% septic | Exfix 36, internal 2 |
-
IM, intramedullary; TAA, total ankle arthroplasty.
Table III.
Author | Year | Country | Fusion or revision | TAAs, n | Mean age, yrs | Female, n | Follow-up, yrs | 1 n indication | Time since primary | 1 n implant | Indication | Procedure |
---|---|---|---|---|---|---|---|---|---|---|---|---|
Myerson et al32 | 2014 | USA | F | 6 | 63.7* | 50* | 1.6* | 66.7% arthritis,* 33.3% inflammatory arthritis | 6 Agility | 100% septic | IM nail | |
R | 7 | Not stated | Not stated | 6 Agility, 1 Salto | 100% septic | Not stated | ||||||
Kotnis et al33 | 2006 | UK | F | 9 | 60.7 | 55.6 | > 12* | 77.8% arthritis, 22.2% inflammatory arthritis | Not stated | 8 STAR, 1 BP | 100% aseptic | IM nail |
R | 16 | 62.7 | 50 | 81.3% arthritic, 18.7% inflammatory | Not stated | 14 STAR, 1 Agility, 1 BP | 87.5% aseptic, 12.5% septic | Not stated | ||||
Makwana et al34 | 1995 | UK | F | 5 | 60.2 | 80 | 5.4 | 18.2% arthritis, 81.8% inflammatory arthritis | 5 | Bath and Wessex | 100% aseptic | 2 IM nail, 3 Charnley arthrodesis |
R | 4 | 63.3 | 100 | 6.6 | 3.4 | Bath and Wessex | 100% aseptic | Not stated | ||||
Groth and Fitch35 | 1987 | USA | F | 11 | 56.5 | 45.5 | 6.5* | 100% arthritis | 2.4 | Not stated | 50* | 1.6* |
R | 5 | 53.2 | 80 | 80% arthritic, 20% inflammatory | 1.8 | Not stated | 100% aseptic | Semiconstrained Oregon | ||||
Stauffer36 | 1982 | USA | F | 17 | Not stated | Not stated | 2.1* | Not stated | Not stated | Not stated | 70.6% aseptic, 29.4% septic | Exfix |
R | 6 | Not stated | Not stated | Not stated | Not stated | Not stated | 100% aseptic | Not stated |
-
*
Includes all patients in the study, not just those included in this analysis.
-
IM, intramedullary; TAA, total ankle arthroplasty.
Further surgical interventions
Six papers analyzed reoperations of revision TAAs and 14 papers analyzed failures of conversion to fusion. Overall, 26.9% (95% confidence interval (CI) 15.4% to 40.1%) of revision TAAs required further surgical intervention (Figure 2); 13.0% (95% CI 4.9% to 23.4%) of conversion to fusions failed, requiring further surgical intervention (Figure 3).
Fig. 2
Fig. 3
Surgery for failure
A total of 15 studies analyzed the requirement for re-revision surgery for failure following revision TAA and 23 following conversion of a failed TAA to ankle fusion.
The pooled percentage requiring re-revision procedures following a revision TAA was 14.4% (95% CI 8.4% to 21.4%) with 2.7% (95% CI 0.8% to 5.5%) being converted to a further TAA, 8.1% (95% CI 2.6% to 15.4%) being converted to a fusion and 0.0% (95% CI 0.0% to 0.2%) undergoing amputation (Figure 4).
Fig. 4
The pooled percentage requiring revision surgery for a failure of a conversion of primary TAA to fusion was 8% (95% CI 4% to 13%) with 5.8% (95% CI 2.5% to 10.1%) undergoing a further attempt at fusion and 0.1% (95% CI 0.0% to 1.1%) undergoing amputation (Figure 5).
Fig. 5
Outcome scores
Five studies with a total of 16 scores reported pre- and postoperative outcome scores for revision ankle arthroplasty; 12 demonstrated significant improvement, and four demonstrated a non-significant improvement (Table IV). Seven studies with a total of 22 individual outcome scores reported pre- and postoperative functional scores for conversion to an ankle fusion. Of these, four demonstrated a significant improvement, 13 did not demonstrate significant improvement, and in five significance was not calculated. (Table V)
Table IV.
Author | TAAs, n | Scores | AAOFAS preop | AAOFAS postop | Significant |
---|---|---|---|---|---|
Lachman et al6 | 29 | AOFAS | 40.6 | 64.6 | Significant |
Lachman et al6 | 29 | SF-36 Mental | 63.8 | 77.4 | Significant |
SF-36 Physical | 28.5 | 59.2 | |||
Lachman et al6 | 29 | VAS | 59.5 | 16.9 | Significant |
Lachman et al6 | 29 | SMFA | 44.3 | 24.2 | Significant |
Lachman et al6 | 29 | Bother | 37.8 | 25.5 | Significant |
Wagener et al37 | 12 | AOFAS | 41 (SD 15; 20 to 79) | 65 (SD 19; 31 to 89), p = 0.01 | Significant (p = 0.01) |
Kamrad et al8 | 7 | SEFAS | 19 | 22 | 0.2 |
7 | EQ-5D | 0.5 | 0.6 | 0.4 | |
7 | EQ-VAS | 51 | 56 | 0.6 | |
7 | SF-36 Physical | 46 | 48 | 0.9 | |
SF-36 bodily pain | 34 | 47 | Significant (0.04) | ||
SF-36 Physical | 31 | 35 | 0.2 | ||
SF-36 Mental | 48 | 49 | 0.8 | ||
Horisberger et al10 | 10 | AOFAS | 39 (18 to 56) | 84 (72 to 97) (p < 0.001) | p < 0.001 |
Horisberger et al10 | 10 | VAS | 6.2 | 0.9 (p < 0.001) | p < 0.001 |
Hintermann et al38 | 100 | AOFAS | 44 (SD 18; 3 to 80) | 72 (SD 19; 25 to 100) (p < 0.01) | p < 0.01 |
VAS | 6.2 (SD 2.4; 0 to 10) | 2.8 (SD 2.4; 0 to 9) | p < 0.01 |
-
AAOFAS, American Association of Orthopedic Foot and Ankle Surgeons; AOFAS, American Orthopedic Foot and Ankle Society; EQ-5D, EuroQol five-dimension questionnaire; SD, standard deviation; SEFAS, Self-reported Foot and Ankle Score; SF-36, 36-Item Short-Form Health Survey questionnaire; TAA, total ankle arthroplasty; VAS, visual analogue scale.
Table V.
Author | Number | Scores | Pre-treatment score | Post-treatment score | Significance | |
---|---|---|---|---|---|---|
Halverson et al15 | 5 preop (3 postop) | FFI | 77.06 (65.88 to 94.71) | 20.42 (0 to 35.38) | Not calculated | |
Aubret et al18 | 10 | AOFAS | 33.8 (12 to 72) | 56 (21 to 78) | Not calculated | |
Kamrad et al5 | 10 | SEFAS | 13 | 17 | p = 0.3 | |
10 | EQ-5D | 0.4 | 0.5 | p = 0.6 | ||
10 | EQ-VAS | 43 | 52 | p = 0.2 | ||
10 | SF-36 physical function | 35 | 32 | p = 0.4 | ||
SF-36 bodily pain | 33 | 37 | p = 1.0 | |||
SF-36 physical | 33 | 29 | p = 0.4 | |||
SF-36 mental | 45 | 47 | p = 0.7 | |||
Paul et al20 | 6 | AOFAS Hindfoot score | 29 (SD 11.1; 12 to 40) | 65 (SD 8.68; 49 to 73) | Significant (p = 0.026) |
|
Wagener et al7 | 6 | VAS | 7.5 +(SD 0.55; 7 to 8) | 2 (SD 1.1; 1 to 4) | Significant (p = 0.0277) | |
Berkowitz et al22 | Pre 12, 9 post | AOFAS | TT | 43.0 +(SD 13) | 67.0 (SD 12) | Significant (p < 0.05) |
Pre 12, 10 post | TTC | 48.4 (SD 14) | 51.2 (SD 17) | Not significant | ||
Berkowitz et al22 | SF-36 PCS | TT | 32.5 (SD 4) | 41.6 (SD 13) | Not significant | |
TTC | 35.6 (SD 6) | 34.1 (SD 7) | Not significant | |||
Berkowitz et al22 | SF-36 MCS | TT | 45 (SD 25) | 48.4 (SD 7) | Not significant | |
TTC | 45.8 (SD 11) | 46.4 (SD 11) | Not significant | |||
Berkowitz et al22 | Maryland | TT | 56.7 (SD 14) | 71.2 (SD 16) | Significant (p < 0.05) | |
TTC | 58.3 (SD 14) | 64.5 (SD 14) | Not significant | |||
Plaass et al25 | 29 | AOFAS | 37 (20 to 63) | 68 (50 to 92) | Not calculated | |
Plaass et al25 | 29 | AOFAS Pain | 8 (0 to 30) | 29 (20 to 40) | Not calculated | |
Culpan et al26 | 12 preop, 16 postop | AOFAS | 31 (12 to 56) | 70 (41 to 87) | Not calculated |
-
AOFAS, American Orthopedic Foot and Ankle Society; EQ-5D, EuroQol five-dimension questionnaire; FFI, Foot Function Index; MCS, mental component summary; PCS, physical component summary; SD, standard deviation; SF-36, Short-Form Health Survey questionnaire; TT, tibiotalar; TTC, tibiotalocalcaneal; VAS, visual analogue scale.
Conversion of primary TAA to fusion
Of 480 patients in 23 papers, the pooled percentage of patients who went onto union at the first surgery was 87% (95% CI 80% to 93%, range 33.3% to 100%) (Figure 6). Some papers reported that union occurred after second or third surgery, and many patients were asymptomatic despite nonunion and did not undergo further surgery.
Fig. 6
Study bias
Bias was assessed using the MINORS criteria. The mean score for conversion to fusion was 7.8261 (95% CI 6.8581 to 8.7941; standard deviation (SD) 2.367). For revision to arthroplasty the mean score was 7.5238 (95% CI 6.34 to 8.71; SD 2.77). There was no significant difference between the scores (p = 0.749, Mann-Whitney U test).
Discussion
This is the largest systematic review of surgery for failed primary ankle arthroplasties. This systematic review and meta-analysis demonstrates no significant differences in the rates of failure and further surgery between either revision ankle arthroplasties or conversion of an ankle arthroplasty to ankle fusion. The rates of below-knee amputation were low.
Revision TAA has a higher rate of failure defined by all reoperations of 26.9%, compared to 13.0% for conversion of TAA to ankle fusion, but this difference was not statistically significant.
A conversion to fusion can either be of the tibiotalar joint alone or also include the subtalar joint. The latter has the advantage of performing a single definitive surgery, but has downsides including leg length discrepancy, nonunion and ongoing symptoms.39,40 Conversion of a failed TAA to fusion also has a high nonunion rate of 13%. The decision on fusion technique will be dependent on many factors, including remaining bone stock in the talus following removal of the ankle arthroplasty and the presence of arthritis in the subtalar joint. Unfortunately, many papers did not differentiate the results between techniques, and it is therefore impossible to draw conclusions as to the relative outcomes.
There were low rates of amputations with 0.1% of conversion to fusion undergoing amputation. These are considerably lower than found in Haddad et al’s41 previous systematic review, which found in primary ankle arthroplasties 1% required an amputation and 5% in primary arthrodesis.
Revision TAA to another ankle arthroplasty historically involved using primary ankle arthroplasties. In recent years, new revision implants have been introduced to the market with increased modularity. This allows for larger deformities and bone loss to be corrected.42 The studies in this review used a mixture of implants.
In our study, 14% of the revision TAAs needed revising again. The largest study by Hintermann et al38 reported a re-revision rate of 14.5%. The studies with the highest risk of failure were those where surgery was performed for infection, which was also true for conversion to fusion.32 This highlights the difficulties in treating periprosthetic joint infection, which are well known.
This study found failure rates for conversion of TAA to fusion of 8%, but nonunion rates were 13% suggesting that some patients live with their nonunion and do not choose to undergo further surgery. A previous systematic review demonstrated fusion rates of 81%,40 which is consistent with our findings. There is a large amount of variation in surgical techniques and indication for arthrodesis following a failed ankle arthroplasty.
It is important to be cognizant of the many variables that dictate choice of salvage surgery following failure of a primary TAA, such as patient variables, bone loss, soft-tissue condition, and the suspicion of infection that may affect the findings, which were invariably not reported.
The patient reported outcome scores in this paper were promising with all studies reporting improved scores. All AOFAS scores improved above the minimally clinical important difference of 7.9. Hintermann et al38 reported 81 of 100 had good or excellent AOFAS scores, and found those with custom components did slightly worse. It should be noted that both Lachman et al6 and the Swedish Arthroplasty Registry demonstrated that functional scores do not improve as much with revision arthroplasty as they do with primary arthroplasty.41 The Swedish Arthroplasty Registry reports a mean SEFAS score of 22 after revision ankle arthroplasty compared to 31 after primary arthroplasties, and this was also found by Lachman et al6 across all scores.41,43 The only study that directly compares functional scores between revision arthroplasty and conversion to fusion demonstrates similar functional scores for both techniques.5,8 A greater proportion of outcome scores were significantly improved with revision ankle arthroplasty than conversion to fusion, but due to small numbers it was impossible to calculate if this was statistically significant. A meta-analysis of functional scores was not undertaken, as only two papers for both revision ankle arthroplasties and conversion to fusion included sufficient data for this to be performed.
Limitations to this systematic review and meta-analysis include the fact that there were few studies that directly compared revision TAA with conversion to fusion. There was considerable heterogeneity between the studies. This includes indication for surgery, surgical technique, and a wide range of outcome scores and complications. The majority of studies were small single-centre case series, which introduces potential selection and reporting bias. A further limitation is the lack of long-term outcomes. The majority of these studies have follow-up of less than five years, or have incomplete data. While all the papers could be included for the general outcomes, many were excluded on some specific analysis as it was impossible to differentiate between surgical techniques and individual outcomes. It was also impossible to include other complications such as deep vein thrombosis and pulmonary embolism, and it was unable to distinguish outcomes between inflammatory and noninflammatory arthritis.
The strengths of this systematic review are that it includes the largest number of studies and is the most comprehensive review of surgery for a failed ankle arthroplasty. This study also attempts to critically analyze all the patients in the papers to draw conclusions on outcomes and differences between surgical techniques.
In summary, revision of primary TAA can be an effective procedure with improved functional outcomes, but has considerable risks of failure and reoperation, especially in those with periprosthetic joint infection. In those who undergo conversion of TAA to fusion there are high rates of nonunion. There is a need for comparative studies using validated outcome scores to assess outcomes following revision of a failed primary ankle arthroplasty.
References
1. Goldberg AJ , MacGregor A , Dawson J , et al. The demand incidence of symptomatic ankle osteoarthritis presenting to foot & ankle surgeons in the United Kingdom . Foot (Edinb) . 2012 ; 22 ( 3 ): 163 – 166 . Crossref PubMed Google Scholar
2. National Joint Registry Steering Committee . 16th Annual Report 2019: National Joint Registry for England, Wales, Northern Ireland and the Isle of Man . 2019 . https://reports.njrcentre.org.uk/Portals/0/PDFdownloads/NJR%2016th%20Annual%20Report%202019.pdf ( date last accessed 12 May 2022 ). Google Scholar
3. Henricson A , Carlsson A , Rydholm U . What is a revision of total ankle replacement? Foot Ankle Surg . 2011 ; 17 ( 3 ): 99 – 102 . Crossref Google Scholar
4. Hintermann B , Zwicky L , Knupp M , Henninger HB , Barg A . HINTEGRA revision arthroplasty for failed total ankle prostheses: surgical technique . JBJS Essent Surg Tech . 2014 ; 3 ( 2 ): e12 . Crossref PubMed Google Scholar
5. Kamrad I , Henricson A , Magnusson H , Carlsson Å , Rosengren BE . Outcome after salvage arthrodesis for failed total ankle replacement . Foot Ankle Int . 2016 ; 37 ( 3 ): 255 – 261 . Crossref PubMed Google Scholar
6. Lachman JR , Ramos JA , Adams SB , Nunley JA , Easley ME , DeOrio JK . Patient-reported outcomes before and after primary and revision total ankle arthroplasty . Foot Ankle Int . 2019 ; 40 ( 1 ): 34 – 41 . Crossref PubMed Google Scholar
7. Wagener J , Gross CE , Schweizer C , Lang TH , Hintermann B . Custom-made total ankle arthroplasty for the salvage of major talar bone loss . Bone Joint J . 2017 ; 99-B ( 2 ): 231 – 236 . Crossref PubMed Google Scholar
8. Kamrad I , Henricsson A , Karlsson MK , et al. Poor prosthesis survival and function after component exchange of total ankle prostheses . Acta Orthop . 2015 ; 86 ( 4 ): 407 – 411 . Crossref PubMed Google Scholar
9. Roukis TS , Simonson DC . Incidence of complications during initial experience with revision of the agility and agility LP Total Ankle Replacement Systems: a single surgeon’s learning curve experience . Clin Podiatr Med Surg . 2015 ; 32 ( 4 ): 569 – 593 . Crossref Google Scholar
10. Horisberger M , Henninger HB , Valderrabano V , Barg A . Bone augmentation for revision total ankle arthroplasty with large bone defects . Acta Orthop . 2015 ; 86 ( 4 ): 412 – 414 . Crossref PubMed Google Scholar
11. Patton D , Kiewiet N , Brage M . Infected total ankle arthroplasty: risk factors and treatment options . Foot Ankle Int . 2015 ; 36 ( 6 ): 626 – 634 . Crossref PubMed Google Scholar
12. Ellington JK , Gupta S , Myerson MS . Management of failures of total ankle replacement with the agility total ankle arthroplasty . J Bone Joint Surg Am . 2013 ; 95 ( 23 ): 2112 – 2118 . Crossref PubMed Google Scholar
13. DeVries JG , Scott RT , Berlet GC , Hyer CF , Lee TH , DeOrio JK . Agility to INBONE . Clin Podiatr Med Surg . 2013 ; 30 ( 1 ): 81 – 96 . Crossref PubMed Google Scholar
14. Schuberth JM , Christensen JC , Rialson JA . Metal-reinforced cement augmentation for complex talar subsidence in failed total ankle arthroplasty . J Foot Ankle Surg . 2011 ; 50 ( 6 ): 766 – 772 . Crossref PubMed Google Scholar
15. Halverson AL , Goss DA , Berlet GC . Ankle arthrodesis with structural grafts can work for the salvage of failed total ankle arthroplasty . Foot Ankle Spec . 2020 ; 13 ( 2 ): 132 – 137 . Crossref PubMed Google Scholar
16. Kruidenier J , van der Plaat LW , Sierevelt IN , Hoornenborg D , Haverkamp D . Ankle fusion after failed ankle replacement in rheumatic and non-rheumatic patients . Foot Ankle Surg . 2019 ; 25 ( 5 ): 589 – 593 . Crossref PubMed Google Scholar
17. Ali AA , Forrester RA , O’Connor P , Harris NJ . Revision of failed total ankle arthroplasty to a hindfoot fusion: 23 consecutive cases using the Phoenix nail . Bone Joint J . 2018 ; 100-B ( 4 ): 475 – 479 . Crossref PubMed Google Scholar
18. Aubret S , Merlini L , Fessy M , Besse JL . Poor outcomes of fusion with Trabecular Metal implants after failed total ankle replacement: Early results in 11 patients . Orthop Traumatol Surg Res . 2018 ; 104 ( 2 ): 231 – 237 . Crossref PubMed Google Scholar
19. Rahm S , Klammer G , Benninger E , Gerber F , Farshad M , Espinosa N . Inferior results of salvage arthrodesis after failed ankle replacement compared to primary arthrodesis . Foot Ankle Int . 2015 ; 36 ( 4 ): 349 – 359 . Crossref PubMed Google Scholar
20. Paul J , Barg A , Horisberger M , Herrera M , Henninger HB , Valderrabano V . Ankle salvage surgery with autologous circular pillar fibula augmentation and intramedullary hindfoot nail . J Foot Ankle Surg . 2014 ; 53 ( 5 ): 601 – 605 . Crossref PubMed Google Scholar
21. McCoy TH , Goldman V , Fragomen AT , Rozbruch SR . Circular external fixator-assisted ankle arthrodesis following failed total ankle arthroplasty . Foot Ankle Int . 2012 ; 33 ( 11 ): 947 – 955 . Crossref PubMed Google Scholar
22. Berkowitz MJ , Clare MP , Walling AK , Sanders R . Salvage of failed total ankle arthroplasty with fusion using structural allograft and internal fixation . Foot Ankle Int . 2011 ; 32 ( 5 ): S493 - 502 . Crossref PubMed Google Scholar
23. Doets HC , Zürcher AW . Salvage arthrodesis for failed total ankle arthroplasty: clinical outcome and influence of method of fixation on union rate in 18 ankles followed for 3-12 years . Acta Orthop . 2010 ; 81 ( 1 ): 142 – 147 . Google Scholar
24. Henricson A , Rydholm U . Use of a trabecular metal implant in ankle arthrodesis after failed total ankle replacement . Acta Orthop . 2010 ; 81 ( 6 ): 745 – 747 . Crossref PubMed Google Scholar
25. Plaass C , Knupp M , Barg A , Hintermann B . Anterior double plating for rigid fixation of isolated tibiotalar arthrodesis . Foot Ankle Int . 2009 ; 30 ( 7 ): 631 – 639 . Crossref PubMed Google Scholar
26. Culpan P , Le Strat V , Piriou P , Judet T . Arthrodesis after failed total ankle replacement . J Bone Joint Surg Br . 2007 ; 89-B ( 9 ): 1178 – 1183 . Crossref PubMed Google Scholar
27. Schill S . (Ankle arthrodesis with interposition graft as a salvage procedure after failed total ankle replacement) . Orthop Traumatol . 2007 ; 19 ( 5–6 ): 547 – 560 . Crossref . (Article in German). Google Scholar
28. Hopgood P , Kumar R , Wood PLR . Ankle arthrodesis for failed total ankle replacement . J Bone Joint Surg Br . 2006 ; 88-B ( 8 ): 1032 – 1038 . Crossref PubMed Google Scholar
29. Anderson T , Rydholm U , Besjakov J , Montgomery F , Carlsson Å . Tibiotalocalcaneal fusion using retrograde intramedullary nails as A salvage procedure for failed total ankle prostheses in rheumatoid arthritis . Foot Ankle Surg . 2005 ; 11 ( 3 ): 143 – 147 . Crossref Google Scholar
30. Carlsson AS , Montgomery F , Besjakov J . Arthrodesis of the ankle secondary to replacement . Foot Ankle Int . 1998 ; 19 ( 4 ): 240 – 245 . Crossref PubMed Google Scholar
31. Kitaoka HB , Romness DW . Arthrodesis for failed ankle arthroplasty . J Arthroplasty . 1992 ; 7 ( 3 ): 277 – 284 . Crossref PubMed Google Scholar
32. Myerson MS , Shariff R , Zonno AJ . The management of infection following total ankle replacement: demographics and treatment . Foot Ankle Int . 2014 ; 35 ( 9 ): 855 – 862 . Crossref PubMed Google Scholar
33. Kotnis R , Pasapula C , Anwar F , Cooke PH , Sharp RJ . The management of failed ankle replacement . J Bone Joint Surg Br . 2006 ; 88-B ( 8 ): 1039 – 1047 . Crossref PubMed Google Scholar
34. Makwana NK , Morrison P , Jones CB , Kirkup J . Salvage operations after failed total ankle replacement . The Foot . 1995 ; 5 ( 4 ): 180 – 184 . Crossref Google Scholar
35. Groth HE , Fitch HF . Salvage procedures for complications of total ankle arthroplasty . Clin Orthop Relat Res . 1987 ; 224 : 244 – 250 . Crossref PubMed Google Scholar
36. Stauffer RN . Salvage of painful total ankle arthroplasty . Clin Orthop Relat Res . 1982 ; 170 : 184 – 188 . Crossref PubMed Google Scholar
37. Wagener J , Gross CE , Schweizer C , Lang TH , Hintermann B . Custom-made total ankle arthroplasty for the salvage of major talar bone loss . Bone Joint J . 2017 ; 99-B ( 2 ): 231 – 236 . Crossref PubMed Google Scholar
38. Hintermann B , Zwicky L , Knupp M , Hb H , Barg A . HINTEGRA revision arthroplasty for failed total ankle prostheses . J Bone Joint Surg Am . 2013 ; 95-A ( 13 ): 1166 – 1174 . Crossref PubMed Google Scholar
39. Donnenwerth MP , Roukis TS . Tibio-talo-calcaneal arthrodesis with retrograde compression intramedullary nail fixation for salvage of failed total ankle replacement: a systematic review . Clin Podiatr Med Surg . 2013 ; 30 ( 2 ): 199 – 206 . Crossref PubMed Google Scholar
40. Gross C , Erickson BJ , Adams SB , Parekh SG . Ankle arthrodesis after failed total ankle replacement: a systematic review of the literature . Foot Ankle Spec . 2015 ; 8 ( 2 ): 143 – 151 . Crossref PubMed Google Scholar
41. Haddad SL , Coetzee JC , Estok R , Fahrbach K , Banel D , Nalysnyk L . Intermediate and long-term outcomes of total ankle arthroplasty and ankle arthrodesis. A systematic review of the literature . J Bone Joint Surg Am . 2007 ; 89-A ( 9 ): 1899 – 1905 . Crossref PubMed Google Scholar
42. Abicht BP , Roukis TS . The INBONE II Total Ankle System . Clin Podiatr Med Surg . 2013 ; 30 ( 1 ): 47 – 68 . Crossref PubMed Google Scholar
43. Jennison T , King A , Hutton C , Sharpe I . A prospective cohort study comparing functional outcomes of primary and revision ankle replacements . Foot Ankle Int . 2021 ; 42 ( 10 ): 1254 – 1259 . Crossref PubMed Google Scholar
Author contributions
T. Jennison: Writing – original draft, Writing – review & editing, Investigation, Methodology, Formal analysis.
C. Spolton-Dean: Investigation.
H. Rottenburg: Investigation.
O. Ukoumunne: Writing – original draft, Writing – review & editing, Formal analysis.
I. Sharpe: Writing – original draft, Writing – review & editing.
A. Goldberg: Writing – original draft, Writing – review & editing.
Funding statement
The authors received no financial or material support for the research, authorship, and/or publication of this article.
ICMJE COI statement
A. Goldberg reports a grant from NIHR HTA, consulting fees from P28, and speaker fees and travel expenses from Stryker, all unrelated to this study. A. Goldberg is on the National Joint Registry Medical Advisory Committee and Editorial Committee, as well as the FAI/FAO Editorial Board and BOFAS Outcomes Committee. I. Sharpe reports consulting fees, travel expenses, and lecture payments from Stryker, unrelated to this study.
Open access funding
The open access fee for this study was self-funded.
Supplementary material
Preferred Reporting Items for Systematic Reviews and Meta-Analyses checklist.
© 2022 Author(s) et al. This is an open-access article distributed under the terms of the Creative Commons Attribution Non-Commercial No Derivatives (CC BY-NC-ND 4.0) licence, which permits the copying and redistribution of the work only, and provided the original author and source are credited. See https://creativecommons.org/licenses/by-nc-nd/4.0/